I just can’t open Claude Code today. I just can’t do it. Do we have a word for Claude overload yet? Like email overload. Zero inbox just isn’t achievable. Sooo many agents, so many tasks. Turns out the locks on my back door void my home insurance. Who knew. And turns out Shin-Etsu is already worth $86bn with 22.6 Fwd P/E and PEG of 2.1. and is the cleanest buy for the litho-resist category. But what about TOK, or Fujifilm? Run the stock analysis skill. And how is my health paper tracking coming along? A Oral PCSK9 — enlicitide (MK-0616) in Phase 3, ~50-60% LDL reduction. It’s the convenient pill version of the potent injectable PCSK9 drugs. I see. Great. What do I do with that. And what did the tiler say about the quote for the London Mosaic heritage encaustic tiles for the hallway? Draft a chase. Ahhhhhhhh. It’s f*cking 4pm. I haven’t moved.

Is this life now?

[Well, it’s life for about 1 more generation until all these tasks are autonomous loops]

Is it Jevon’s paradox up until it isn’t? 1 more gen and it really is keynes economic possibilities for our grandchildren.

But for now, folks, it’s just agent switching until we all go mad.

Here is a long one but a good one. I spent last week on holiday, writing and clauding, and writing and clauding. Behold:

Apple? Anyone?

The pattern is that everyone is moving into everyone else’s layer. Bunfight they call it. Or bumfight. I don’t know which. And I’m not gonna find out either. You will be able to tell proof-of-humanhood soon by just spotting a mistape.

• The model labs are designing their own chips now. OpenAI is building a 10GW accelerator with Broadcom. Anthropic co-develops Trainium with Amazon.

• The hyperscalers that built chips for themselves are turning around and selling them, Google to outside data centres, Amazon the same.

• Nvidia, which only ever sold the chip, is buying equity in the labs that buy the chip.

• And Intel, the one firm in the picture that was a chip vendor first, is moving in the opposite direction, down into the fab. It is a value chain eating itself.

But look at manufacturing: TSMC, TSMC, TSMC, all the way across, nine companies, one manufacturer.

Fabs were once the worst business in semis.

The base is sold out, and it is putting up prices

I wrote a version of this back in December, in LFG (for semiconductors): the foundry layer is where the next decade of value capture lives. The numbers since then have made the case stronger. TSMC’s 2nm node is sold out for 2026, with customers appling for allocation out to mid-2027. Its advanced packaging, the CoWoS step that bolts the logic to the memory, is booked through 2027 on lead times of a year or more. Plus, it’s raising prices into that demand, 3nm wafers up about 15% in H2, packaging prices climbing 10-20% a year against 5% for ordinary logic. TSMC has lifted its long-run gross-margin floor from 53% to 56% apparently in the middle of the biggest capital build in its history. Demand is strong. Monopoly pricing.

The same is true one shelf over, in memory. High-bandwidth memory (HBM) is sold out industry-wide, with buyers receiving barely 1/2 or 2/3 of what they ask for, and contract prices look set to roughly double into 2027. And upstream of all of it sits ASML, the lithography monopoly, the single chokepoint the whole industry’s constraint is migrating towards. 2029, you just wait.

The bull case for the layer above is real. Nvidia is not losing this. Its data-centre business did $75bn last quarter at a 75% gross margin, and the share it cedes to custom chips over the next year is more like 10 points than 50. The thing eroding Nvidia is the hyperscalers’ own silicon, not AMD, but that erosion is slow. I said it in The Compute Gradient and I’ll say it again: the competitors are breaking Nvidia’s monopoly, not its moat. Those are different things.

But see what happens to the value even when a hyperscaler “wins.” When Google sells you a TPU, Broadcom took a design fee, TSMC took the wafer, and the memory came from 1 of 3 Korean and American firms. Google keeps a cloud margin. The chip designers fight over the scraps in the middle. The design layer is a real market with real winners and losers, which is exactly why it is not where the durable money sits. See memory wars. The money runs through it, down to the fab. Not a moat. What‘s the opposite of a moat? A sieve? Sounds right.

The surprise upstairs: the model isn’t commoditising

Here is where my own prior was wrong. The consensus line, mine included, was that raw model capability commoditises, prices race to zero, and the value flees upward to the application or downward to the chip. Half of that is happening. At a fixed level of quality, the price of a token really is falling about 10x a year, which is astonishing and deflationary and exactly what you’d expect.

And yet the frontier labs are putting their prices up. GPT-5.5 launched at double its predecessor. Anthropic’s inference gross margins went from the high thirties to over 70% in a year. I don’t even want to know what I will have to pay for Fable if I can get my hands on it. The reason is agents. An agentic task burns 5 to 30x the tokens of a chat, because it re-reads its whole context at every step, and that wall of demand has been raising the clearing price of a frontier token faster than the cost of serving one has fallen. So the model layer keeps its pricing power as long as it stays ahead, and the margin pain has moved somewhere else entirely, downstream, to the application layer, where gross margins have slid from the 80s into the 50s as the cost of the underlying intelligence eats the SaaS economics alive.

You can watch it happen in agentic coding. Cursor was doing 2.5bn dollars of ARR and just got bought by SpaceX for 60bn. Claude Code is on a similar run-rate, Codex has 4 million weekly developers. Extraordinary growth, and yet the standalone tools keep getting absorbed, Cursor into SpaceX, Windsurf into Cognition, because the in-editor assistant is commoditising back towards whoever owns the model. The durable winners up here look like the labs’ own agents and the odd wedge into people who can’t code at all. Maybe old industry consulting. But the labs are making a play there too. Everyone else is just renting their moat from the layer below. Can you rent a moat?

What changed? The constraint moved twice

If you only take two things from this, take these.

Uno, is that the binding physical constraint is no longer chips. It is power. Satya said the quiet part out loud, that Microsoft has chips sitting in inventory it can’t plug in because it doesn’t have the “warm shells,” the buildings with power, to put them in. Jensen calls AI “a power-limited industry.” American data-centre demand goes from 31GW last year to 66GW by 2027. Gas turbines are sold out to the end of the decade, big transformers run 2-4 years out + grid-connection queues stretch past 5. The American energy department issued an emergency order last month letting the grid operator switch data centres off when the system is stressed. None of that gets fixed inside a year. Power is the new toll booth, and the market I don’t think has started to price it. Diligence pending.

Deux, is that the risk is no longer demand. It’s money. The 4 big clouds will spend something like a trillion dollars this year. Data-centre construction is already a large chunk of US GDP growth. And it’s not really cash flow anymore, it’s debt and a tangle of circular financing, Nvidia into OpenAI into Oracle into AMD into CoreWeave, that even Jensen calls “ridiculous” while doing it. The Federal Reserve flagged AI as a top systemic risk. Bain reckons the industry needs to find eight hundred billion dollars of annual revenue it doesn’t yet have. Worth watching.

Read those together and you get the geometry of the next year. Demand is real, capability is real, the capacity is sold out. This is not the dot-com fibre glut. The danger is the financing, and the first crack will show in a credit spread or a missed revenue number not utilisation rate or memory walls.

Where this goes, and where I’d stand

So, a 12 month prediction.

The base of the stack stays sold out and keeps raising prices. Claude puts that at about 85%, and the only thing that breaks it is a financing accident upstream. Nvidia’s share drifts to the low 70s rather than falling off a cliff, the moat holding longer than the bears want.

And power overtakes chips as the botteneck, and the headlines fill up with gigawatts and gas turbines and small reactors. And somewhere in the next year, I think more likely than not, one of the heavily indebted neoclouds or one of the circular-financing counterparties has a scare, a refinancing that won’t close or a number that comes in light, and the most leveraged names take a beating. A scare, I’d say, is the base case. A full 2000-style bust is not, maybe a one-in-four, and it needs a marquee counterparty to actually break, an OpenAI revenue miss or an Oracle that can’t roll its debt.

As for where I’d stand, note I am a seed-stage investor/hop hop aficionado not an economist, I can’t wait for the data, I have to bet before it. If the whole stack is renting from the fab, you want to own the fab, or the things the fab can’t operate without. In public markets that is TSMC, sold out with pricing power, and ASML, the purest chokepoint, and the memory makers if you’ve the stomach for the cycle. The contrarian version is the boring stuff nobody prices as “AI,”. e.g the materials and consumables that get used up on every wafer regardless of the logo. Can I interest anyone is a little photoresist?

But the best thesis is probably power. If you know, you know. The constraint moved to electrons faster than the companies that sell electrons re-rated. Gas turbines sold out to 2030, the grid and transformer names, the independent power producers and the nuclear restart stories. Won’t someone think of the climate tech funds? what will they do? when the future of compute becomes power infra?

The irony in all of it is that the thing most likely to break this is not whether the AI is any good. The AI is plenty good already. And people clearly want it. The thing most likely to break it is the daisy-chain of cheques being written to pay for the electricity to run it. The chips will keep coming out of Taiwan either way. The question is: who will still be standing to plug them in.

It’s all going to come back to Centrica isn’t it. I am going to have to buy Centrica…

To make the format even make a semblance of sense, here are 3 stories. Three Story Friday. This format is a bloody mess.

1. TPUs To Go

First up, Google. The Journal says they’re building a real chip business now, selling TPUs into other people’s data centres rather than just renting them out inside Google Cloud, and Sundar flagged as much on the Q1 call. The deals are real, too: Anthropic’s on the hook for up to a million Ironwood chips (that’s v7), over a gigawatt of them this year and heading to 3.5 by 2027, with 400k bought outright through Broadcom for 10 billi, and Meta’s supposedly in talks as well. Ironwood comes in at about 90% of a Blackwell for maybe 44% less per chip, so a real competitive part. If you want the technicals and the memes there’s always Dylan, otherwise you get Jay-Z quotes, and you will like it.

Here’s my unique take, as Opus called it. Google doesn’t actually make a TPU, and once you read the map, you know: Broadcom designs the core, MediaTek does the cheap inference one, and TSMC bakes the whole thing on 3nm. So Google “selling you a TPU” is really Google taking a cut on top of Broadcom’s design cut on top of TSMC’s wafer, and going merchant is just Google reaching for a bigger slice of that stack.

Source: CNBC.

Background: Can We Make Enough AI Chips? (Nov 2023), where I said the bottleneck would be packaging, not logic. I’ve moved since, all the way down to the fab.

2. Everything Must Go

And with Amazon, even more hands in the pie, as the old saying goes. They’re in talks to put Trainium in other companies’ data centres, which would be the first time they sell the chip instead of renting the compute. Trainium3 is their first 3nm part, about 4x Trainium2 at half a GPU’s price, and it’s the chip behind Anthropic’s Project Rainier, apparently nearly 1 million of them soon, + 5GWs of power + 100bn dollars committed to AWS.

Note, the design partner. Trainium2 was a Marvell job, for Trainium3 it’s Annapurna for the front-end now and Alchip for the back-end, so Marvell got dropped in a single generation. Brutal business. So yes, Amazon are selling "their chip”. But also, like not really. Hell of a value chain, you’ve got there. A Trainium on its own in someone else’s building is an ASIC with a worse software story than the GPU it’s meant to beat, and the cloud wrapped around it is the only reason it works at all. Does anyone outside the Anthropic mates’-rates actually buy that?

Source: Bloomberg

3. Reverse Ferret

So design is a brutal business. So if I am Intel, what do I do? I go into an even better business. Making chips. Like we used to do, but got out of it because margins were poor and it was soo expensive. Well times have changed.

Well, Intel just hired Seok-Hee Lee, ex-CEO of SK Hynix (the HBM maker mainly now), to run its foundry and packaging. Look at the map again. Intel, the original chip vendor is back in the manufacturing game. But this isn’t about margins and money. It’s an American hedge against Chinese invasion of Taiwan.

it’s becoming ever clearer what Intel is. Not AI chips. It’s under 1% of accelerators, killed the Gaudi target in October and canned Falcon Shores. It’s good at the unsexy stuff: the x86 base, being the only leading-edge fab on American soil with Trump holding 10%, 18A finally shipping in Panther Lake, and the one thing people actually like: packaging, EMIB and Foveros, where it’ll package a TSMC die and get paid even when it doesn’t do the wafer. The Lee hire is Intel betting the money’s in manufacturing, not the chip. This is what the chart says to do. It’s a good chart. That’s strategy.

The stock’s though? It’s up c.190% this year on a 78x forward multiple, so the market’s already paid for a comeback. It better show up. Foundry lost 10.3 billion last year, 18A is making margins worse before it makes them better, and there’s no committed leading-edge customer signed. But remember A is smaller than nm. So A is better than NM. It’s fewer letters for a start.

Source: Yahoo Finance

Long one today. Hope you made it. You are loved.

If you missed it:

• LFG (for semiconductors), Dec 2025, the foundry layer is where the next decade of value lives

• The Compute Gradient, Sep 2025, power as the real scaling variable. Read

I’ve been reading Sam Harsimony’s Splitting Infinity this week, a piece on the technologies he’s skeptical of, there’s a lovely bit of accounting he calls the idiot index: the ratio of what a finished thing costs to what its raw materials cost. For most things the raw inputs are tiny, and the rest, land, equipment, financing, thermodynamics, is the part that won’t budge however clever/rich you get. He calculates rocket fuel bottoming out near $55 a kilo. A person outputs about 10 bits a second whatever you bolt to their skull. It’s good stuff. You can disagree with assumptions and input costs, but it’s a good exercise for you/claude to do.

I was primed with this idiot index when reviewing this weeks’ stories, with the concept of recursive self-improvement. Software’s idiot index is basically nothing now, the marginal cost of more code rounds to zero, so of course it improves itself, of course the flywheel spins. But the atoms and electrons underneath it, power, heat, fabs, optics, has a savage idiot index and concrete physical floors. AGI won’t solve that even if we have robotics and nanoassembly. Generating, moving and detecting atoms, electrons and photons will always have a cost, and likely an increasingly large share of the cost.

But still, I mean, despite the costs, recursive self improvement will be pretty important.

1. The last invention that man need ever make

“When AI Builds Itself” is Anthropic’s argument that recursive self-improvement, the model getting good at building the next model, could arrive before anyone’s ready. As per last week though, I think whenever it arrives it will be before anyone’s ready. Well Blair will be ready of course. He’s always ready.

The idea’s older than the field of AI. In 1965, 3 years before HAL turned up, I.J. Good, a Bletchley cryptographer who’d worked with Turing, described an “intelligence explosion”: a machine clever enough to design machines builds a better one, which builds a better one, and away it goes. He called the first one “the last invention that man need ever make”, with a caveat that’s aged well, “provided the machine is docile enough to tell us how to keep it under control”. That sat in sci-fi and seminar rooms for 60 years, Vinge, Bostrom, a decade of Yudkowsky arguing fast versus slow takeoff online. But now we really are close?! If someone tells me this is PR for the IPO, so help me god…

More than 80% of the code Anthropic merges is authored by Claude as of May. They’re plotting their models on METR’s time-horizon metric, the length of task a model can finish solo. We are heading to weeks by 2027 team. This is your weekly reminder.

So they’d like the option, their word, for the world to be able to slow or pause frontier development, labs verifying each other have actually stopped. The lab fastest at building itself asking for the brake.

Where i’d push back is the usual place. All this self-improvement still has to run on something physical. “Just build more substations”. Maybe, but i don’t buy the grid moves at model speed, and the more I look the more it’s power and heat that caps this by 2027, not model cleverness, not chips (the energy-is-destiny drum I was banging in April). Well actually, once the memory bottleneck is solved, it will be litho.

Source: Anthropic | METR. Background: The Compute Gradient, Sep 2025, on power as the real scaling variable.

2. The Machine That Builds the Machine

Good’s machine needed better hardware to run on, the part he hand-waved. Well, NVIDIA. Two announcements this week. It started shipping Spectrum-X co-packaged-optics (CPO) switches, built with TSMC, to select partners. 400 terabits a second on TSMC’s COUPE photonic packaging. And it put its own compute inside TSMC’s fab, cuLitho doing the mask work, claiming 20-50% on computational lithography (their figure). So the chips that train the AI are now partly designed by AI and wired together with light. The loop closes.

CPO, quickly: today the optics that turn electrons into light sit in a pluggable module at the edge of the switch. Called “pluggables” I don’t know why. And at these speeds the electrical signal can’t survive the few centimetres of copper to reach it without burning power and falling apart. Co-packaged optics bonds the optical engine right next to the switch chip instead. So you get shorter copper, less power per bit, more bandwidth, etc et al. It’s the interconnect, not the chip, that’s the data-centre wall now. But you already knew that. Buy copper.

NVIDIA is one of many here though. Broadcom (AVGO) has been shipping its Bailly CPO switches since 2024, 50,000+ through 2025, with a 102-terabit next-gen sampling now. So what’s NVIDIA’s $2bn into Coherent (COHR) and Lumentum (LITE) buying? Supply. Marvell (MRVL) also bought Celestial AI for up to $5.5bn in December for the same reason. The whole optical supply chain’s being bid up at once. I called the photonic engine for the interconnect back in February. This is just the start. We will likely need multi-material PDKs soon….

Source: Focus Taiwan | Broadcom. Background: Photonic Engines for Data Centers, Feb 2026.

3. Quantum?

And quantum for the palate cleanser. If AI moves too fast to track, let’s try something that moves so slow, it might literally never arrive.

IBM this week said it will spend >$10bn over 5 years to build a fault-tolerant machine called Starling “promised” for 2029. Yes yes, from the guys that brought you Watson. But the interesting bit is what they announced alongside it: Anderon, a standalone company they’re calling the first pure-play quantum wafer foundry, and the money isn’t quite what the headline suggests. It’s $1bn from the US Department of Commerce under CHIPS, matched by $1bn of IBM’s own cash, plus IP, assets and people on top. Call it $2bn all in. Albany and300mm, pitched as multi-vendor.

And also this week, 2 European quantum rounds: Oxford Quantum Circuits in the UK, £260M, the biggest private quantum round Europe’s done, the British Business Bank anchoring £100M. And Quobly in France, €115M, led by STMicroelectronics, silicon-spin qubits on bog-standard FD-SOI 300mm wafers.

Quick summary so you don’t have to claude: “quantum computer”.

• Superconducting (IBM, Google, OQC, Rigetti) runs loops of current near absolute zero, furthest along, needs dilution fridges.

• Trapped-ion (IonQ, Quantinuum, Oxford Ionics) holds atoms in electromagnetic fields, lovely fidelity, slow.

• Neutral-atom (Pasqal, QuEra, Atom Computing) does similar with lasers and scales nicely.

• Photonic (PsiQuantum, Xanadu, Quandela) runs light through silicon photonics, room temperature, fibre-native.

• Silicon-spin (Quobly, SemiQon, Diraq, Quantum Motion, Intel) puts the qubit in something close to a transistor, the newest and the most fab-friendly.

• Diamond NV (Quantum Brilliance, SaxonQ) parks the qubit in a nitrogen-vacancy defect in diamond, runs at room temperature, rugged and portable, better at sensing than scale so far.

• Topological (Microsoft) bets on Majorana modes that are meant to be error-resistant in the hardware itself, gorgeous on paper, not quite demonstrated in practice.

• Quantum annealing (D-Wave) isn’t a gate machine at all, optimisation only, but it’s the one that’s actually been shipping and selling for over a decade.

• Electrons-on-helium (EeroQ) floats qubits on the surface of liquid helium, about as niche as it gets, early but charming.

The tell in IBM’s announcement is the foundry. $10bn is just money; building your own wafer fab from scratch tells you the capital load superconducting carries. And IBM isn’t the only one reaching for silicon — back in January IonQ bought SkyWater, the US foundry, outright for ~$1.8bn, to lock in a captive onshore wafer line. Two routes to the same instinct: IBM building greenfield 300mm as a multi-vendor pure-play, IonQ buying an existing fab and making it captive. The thing worth noting is that IonQ is trapped-ion, supposedly one of the lighter-capex approaches, so when even they decide they need to own the silicon, it says something about how much fab access is worth.

Which makes me think the fab-friendly architectures, silicon-spin and photonics riding a standard wafer line, have a structural edge: free use of the whole semiconductor industry’s capex. So it’s vertical integration (capex heavy, higher profit) versus horizontal integration (low capex, lower profits). Pick your fighter. My bet is low capex wins by 2030, though the IonQ move is a swerve, and my own knowledge base flags this as contested. Harsimony up top would tell you quantum’s oversold; on breaking encryption he’s right, post-quantum crypto got there first, but the $10bn’s betting the rest is worth a foundry.

Source: Quantum Insider | The Next Web. Background: Qubits in a Fridge, Dec 2024, on silicon spin and why a normal fab matters.

4. The Machine Web

Finally, and short because of the quantum primer above. Cloudflare’s Matthew Prince says bots have passed humans on the web for the first time, 57.5% of requests across their network automated against 42.5% human. In March, at SXSW, he’d said the crossover was a 2027 problem. It came 18 months early. The agents, as I have written, have arrived.

Caveat: it’s HTTP requests on one network, not time or attention, so it isn’t “57% of reading is robots”. But the direction is the consumer-web version of the agent shift, single chatbot calls giving way to fleets of agents. Question what is the bottleneck: orchestration or raw tokens? Quick answer: energy.

The obvious winner here, if there is one, is Cloudflare (NET), which is pushing pay-per-crawl, a toll on the bots that it’d happen to collect. Bot-detection and any “prove you’re human” layer get more valuable from here too. Get your agents to track ZKP more closely and do a sourcing run on founders. Also, what happens to an ad-funded web when most of the traffic is machines that never see an ad? Does a toll actually hold, or do the labs just route round it through residential proxies?

Source: Tom’s Hardware | The Decoder. Background: You Like AI Agents?, Feb 2025.

What else I’ve been reading

• I mentioned Sam Harsimony’s Splitting Infinity up top.

• 1 + 1 = −1, inside a crystal. Physicists watched two lattice vibrations combine and spin the opposite way, forced by the crystal’s symmetry. Angular momentum, going backwards. ScienceDaily

If you want to say anything true about AI you have to track two things at once:

1. how fast the models improve

2. what that does to the world

You can’t do 2 without really knowing 1. And 1 moves fast. METR's chart has the time-horizon of tasks these models do reliably doubling every 7 months, and over the last year more like 4. Which means these agents went from reliably handling a task that takes a human a 1 or 2 mins back in 2023 to an hour or more of real work now, with month-long tasks pencilled in for 2027 if the faster line holds. SWE-bench Verified has gone from the 33% to nearly 90% in about two years. Opus went 4.7 to 4.8 in a few weeks. More on that later.

To know where 1 actually is you have to live in it, on the API, burning real tokens on real tasks, not poking the free chatbot tier once a month. How many agents you running? How many politicians or commentators or investors? have clawdbot on the mac mini?

i know i keep circling this. I’ve written some version of it twice already, in Occupational Downgrading and Young People Can’t Get Jobs, so i won’t pretend it’s a fresh take. But this week earned the repeat: the models got better at the real-world work that workers do (GDPval), the youth-unemployment numbers are already awful even without directly attributing it to AI, and TBlair, the one senior politician who plainly gets it, got shouted down over Blairism and Iraq instead of the argument. It keeps happening every week, more and more of it, the progress, I mean. It just keeps coming thicker and faster than ever.

i. Blair’s Been On The Tokens Again

Alright, he hasn’t literally been using the tokens, I don’t picture Tony on the API at midnight. But I feel like he really feels the AGI, you know? And he wants to tell people. Maybe a bit too much, but fair play. The essay reads like a man who’s at least been in the room with the people building this, which is more than you can say for most of the takes. He tells the Labour party it’s “playing with fire”, says AI isn’t a good thing or a bad thing, it is “the thing”, and that “people in most countries, including Britain, have no idea what is about to hit them”. Maybe he reads this newsletter? Is dat you @3RDWAY.COM?

He’s about the only senior figure still in British public life who’ll put it that plainly. The other two who “get it” haven’t put their heads above the parapet. Rishi Sunak now advises Anthropic, George Osborne went to OpenAI to run its government-deals operation. The British politicians who get the frontier are the ones who left to work for it. Make of that what you will. They are not about to get left in the permanent underclass.

The fix is a bit wobbly frankly, he says the “radical centre” where you “put policy first and politics second”. If anything policy not politics is the Starmer mantra. But yes definitely, radicalism. He says a few lines later: the timescales “need long-term strategic thinking which is alien to the way most modern democracies function”. Again, true.

Blair is token-pilled and ready to go. But the backlash was instructive. Look at who went for him and what they actually said. Andy Burnham swatted the centrism, the centre hasn’t delivered for people so they’ve gone to the extremes. A senior party source said he’d abandoned social democracy for an agenda with “no answers”. Treasury minister Dan Tomlinson said the whole thing just re-litigates an old Labour squabble. None of them went near the AI. Or the radicalism. 5,600 words warning that nobody in Britain is ready for what’s coming, and the party had a row about Blairism.

If you are bearish about the political system’s ability to adapt to the age of AI, this confirms it. This is likely the starting point of a much wider and longer debate about how to deal with AI. With this essay, Blair is trying to wake the UK political class out of its slumber. We will need much more of this. Maybe this is the start of the radical centrism we need?

Let’s go further, Starmer’s last action in Government could be to set up an AI Task Force with real teeth. A statutory body. Think about it: Blair, Clegg, Sunak, Osborne, Hassabis, Suleyman, Clark. Take it seriously. All the parties and all the labs.

Source: Tony Blair Institute | The Register.

ii. Seriously, This Youth Unemployment Thing

And obviously, one of the most important impacts of AI, at least before everything is rendered unimportant, is job losses. Alan Milburn, a former UK MP, is one of the good guys. And whilst this report is focused on Britain, the broader issues are global. Milburn’s review: 1.01 million 16-to-24-year-olds in Britain not in work, education or training. Highest in twelve years. Six in ten have never had a job at all, up from four in ten in 2005. Apprenticeships and entry-level roles drying up, hiring gone more remote, more automated, less human, a kid filtered through a portal and an algorithm before a human ever claps eyes on them.

The report doesn’t pin it on AI and he’s right not to. The causes he lists are real and mostly older than the models: rising ill health, a hands-off benefits system, limited, patchy, and low-status vocational training, a labour market still hungover from Covid.

If the people who think AI is about to hollow out junior white-collar work are right, and the early data is starting to agree, graduate unemployment has drifted above the national rate, which is historically backwards, then this million is the good version. It’s the number from before the thing i’m actually worried about has properly kicked in, and whatever AI does to entry-level hiring piles straight on top of everything Milburn already counted.

And the causes are about to become impossible to pull apart. Remote, automated hiring makes the first job harder to get, the jobs that remain are fewer and flakier, insecure work grinds down mental health, poor mental health pushes more people out of work entirely, and round it goes. 6 in 10 never holding a job. At some point, maybe now, arguing over how much is “AI” and how much is “Covid” or “benefits” is just a way of dodging the only fact that matters, which is that it keeps getting worse. We know how to fix the supply side, skills and training and pathways, the menu every one of these reports closes on. None of it helps if the demand side, the actual jobs, carries on shrinking underneath.

But we need something as rigorous as this report that starts from the assumption that the demand side cannot be fixed, in the traditional way. Removing minimum wages, making it easier to hire and fire, will all help, but at some point, we need to start thinking about what we do with a shrinking job market. Because the traditional way of thinking about unemployment and finding pathways into work will be redundant. I am all for renaming unemployment benefit “citizen dividend”. That can be my contribution. If the serious people can start asking Opus 4.8 to draft some policy that would be great.

Source: ITV News. Background: Occupational Downgrading, Mar 2025, on why this stays invisible in the statistics for years.

iii. Opus 4.8 is Good (60% Con

And re 1, “how fast models improve”. Claude Opus 4.8 is out this week, about six weeks after 4.7, which came weeks after 4.6. First note, the drumbeat is speeding up, not slowing down. On SWE-bench Verified it scores 88.6%, up from 87.6%, and everyone, Simon Willison included (”a modest but tangible improvement”), has noted the single point and moved on. But SWE-bench is saturated at this level, a point is about all there is left to win, so the headline number is the least informative thing in the release.

Look at the benchmarks that aren’t maxed out. On the harder coding set, SWE-bench Pro, it went from 64% to 69%. And on GDPval, OpenAI’s own benchmark of real knowledge work, tasks from finance, law, consulting and healthcare graded against actual client deliverables, Opus 4.8 jumped from 1753 to 1890 Elo in a single release, roughly a two-in-three win rate head to head against GPT-5.5. On the tasks themselves, the top models now score in the range of a professional with fourteen years in the job. That’s the first properly decisive lead on broad knowledge work in the whole Opus 4 line, and it got there using about a third fewer tokens than 4.7.

It’s not a clean sweep, GPT-5.5 still takes the raw terminal benchmark and gets there in fewer steps. But the practical version of that GDPval jump is unglamorous: the multi-step research-and-write job, or the codebase-wide refactor, that broke three (!?) months ago and you stopped bothering to ask for, now just finishes. For many of these tasks, it’s a case of it works or it doesn’t, it’s binary and the 60 or 70 or 80% on a benchmark doesn’t really matter until it passes the threshold of “good enough” and then it’s saturated. It’s hard to showcase that with a benchmark, you sort of have to use it and see. Hence my point earlier about using the API and feeling the AGI.

Relatedly to “it now works” is also “it doesn’t lie or make mistakes”. 4.8 misses flaws in its own code about 4x less often than 4.7, and writes dishonest summaries of its own agentic work roughly 17x less often than Sonnet 4.6. Lying less is one of those really important features that “enterprises” seem to care about. Put the knowledge-work jump next to a model that’s finally honest about what it did, and you are getting dangerously close to that mythical “enterprise agentic adoption”. Seems really good at giving conviction levels too as per my usage today, which I think is a handy little feature.

Source: Anthropic

What else I’ve been reading

• Exoplanet weather: WASP-94A b, 700 light-years off, grows rock clouds of magnesium silicate every morning and clears by night. The first proper daily weather map of another planet. ScienceDaily

• Twisted graphene reveals a hidden superconductivity switch. Scientists have uncovered a surprising new way to control superconductivity — the mysterious phenomenon where electricity flows with zero energy loss. ScienceDaily.

If you missed it:

• Young People Can’t Get Jobs. Now What?, May 2025. The bottom rung was already going. Milburn just counted the gap.

• Occupational Downgrading, Mar 2025. On why the collapse stays out of the headline statistics for years.

There’s this feeling that isn’t quite joy. Or happiness. It’s part relief, part release. I can’t stop smiling. I’ve been smiling for days now and I can’t stop.

For the last 22 years. It’s mainly been fear, nervousness and disappointment. And the little wins we couldn’t celebrate because we hadn’t won the league. Get back down the tunnel.

But this feeling. An old familiar friend. Your best mate when you were 7 who you played out with every day after school. But then you went to a different school. And you see him down Spoons when you are 21, and you nod your head but don’t speak to him. Times have changed. What would you even say? They say he spent time in juvie.

So you just forget. In 1998 I was a boy. 2004 I thought would last forever. But it fades and you forget. You forget why you even support the club. The feeling no longer sustains us. We grow up. Forget why. Disappointment. Henry leaves. 8-2. Nasri, Adebayor. Fabregas. The ones we love always leave us. It’s too painful to keep doing it.

I’d forgotten. Why we care. It was a routine. An unhappy marriage. Just keep turning up. It’s easier than the alternative. For the kids, yea?

And then it happened. A random Tuesday night in a quiet hotel room in Berlin. Oh, I remember. They can’t score 2 in 6 minutes. Can they?

Then All of It at once. Oh, this is it. This is why. It matters. It’s deep. It’s not just me. I remembered the community of it. The stupidity and the beauty of caring about something together. A huindred thousand people just decide to commune around an empty stadium. It’s for something. It’s meaningless yes. But it’s also all that matters.

I am not quite happy. Not quite relieved. I’m not sure what it is. It’s bigger than me. I’m happy for others. I’m happy for the players. Saka. And Arsene. Community. And in a secular world, that’s rare. We all need a little more of this.

And now onto things that, in theory, matter more. But right now, to me, they don’t matter at all.

1. SpaceX Files for the Largest IPO in History

If I have to write about stuff this week, the most obvious place to start is SpaceX and the planned $1.75T IPO. It would top Aramco’s 2019 record as the largest IPO ever, and that number is larger than the entire US M2 increase since 2024?!

There is some very good stuff in the prospectus, and I encourage you to read the primary source. But a few little details: Musk on 85.1% of voting power, dual-class. That’s sort of par for the course. But what about: pay vesting against a permanent Mars colony and 100TW of orbital compute. Absolute scenes, lads. And if people want to complain about executive pay, I am very happy to have the discussion, but also, like that’s fair enough isn’t it? You get your money ONCE you set up a Mars colony and take the step towards Type II Kardashev civilisation?

I was surprised to learn the 2025 revenue was $11.4bn connectivity (Starlink), $4.1bn launch (SpaceX), and $3.2bn AI (xAI). Starlink is slept on man. I got a leaflet through the door the other day. £35 quid a month for 3 months. I mean, decent.

The AI bucket is mostly Colossus 1 and 2, the two Memphis clusters xAI built before Musk merged xAI into SpaceX. xAI lost $6.4bn on that $3.2bn revenue last year.

And the recent Anthropic deal happens to fix that AI accounting for him. $1.25bn a month (?!) to May 2029. Although its discounted. Lovely bit of business. Remember xAI was set up to train Grok. Grok is bad, Elon knows, so he created maybe the best neo-cloud in the world?

My overriding feeling with SpaceX and Anthropic (and OAI), is that these select few companies (+ Google Deepmind + Meta a bit) are shaping our future. We, sitting here in the UK and Europe, and frankly every other nation, are not even in the room. Not even in the arena. These sorts of deals, numbers, and likely revenues, profits, will be truly unprecedented. We can cling to the datacentre water story, or read Gary Marcus, or say this is uneconomical, and it will have to be small local model on little devices because the brain uses 8 watts or whatever. But wake up, folks. History is being shaped today. We are all becoming Argentina.

Source: NPR. Background: How to Invest in AI Sovereignty, Feb 2026, where i argued sovereign AI infrastructure was where the value sat.

2. Karpathy Joined. So Did The First Profit. Read The Fine Print.

And on that note, you might think, it’s not over yet. We have lots of neo-labs and well funded competitors, and history tells us the first through the door gets shot. There is still time to catch up. Thinking Machines.

But Anthropic told its investors this week that Q2 revenue will be $10.9bn against $4.8bn in Q1, plus its first ever operating profit, $559m. Last summer they were telling those same investors not to expect profitability until 2028. Two years ahead of plan. Q2 more than doubles Q1, Q1 already doubled Q4. This scale is indeed unprecedented.

An even more important story is that one of the most feted researchers in AI, Andrej Karpathy, ex-OpenAI founding member, ex-Tesla AI lead, the person who could raise a billion dollars on a tweet, and Ross Nordeen, a founding member of xAI, both joined Anthropic. Both could have raised $1-3bn for their own foundation labs without breaking a sweat. They went to Anthropic. Issue #12 said maybe 100-500 people on Earth can lead a frontier training run; the smart ones now think a single lab is the only place to do it. If you are one to update your priors, I think right now you are positively updating on Anthropic and negatively updating on all the other labs and neolabs like SSI, Thinking Machines or Reflection. You are gonna need a bigger computer.

A few other little notes that came from me researching this story:

• The “bigger training budget wins” thesis was wrong. Anthropic spent $5bn on training in 2025, OpenAI spent $20bn. 4x less. Anthropic took the enterprise path instead: Claude Code at $2.5bn run-rate, 2x since January, enterprise customers spending $100k+/yr have grown 7x, business is now half of Claude Code revenue. They are basically an enterprise SaaS company that happens to train models. They got profit by skipping the consumer scale war. B2B FTW.

• The $559m profit is one quarter wide. Anthropic’s compute bill is about to triple. The Colossus 1 contract with SpaceX (Item 1) costs $1.25bn a month at full price; the discount rate runs out in June. Most of Q2 was at the discount. Q3 onwards looks very different.

• OpenAI is disputing the revenue numbers. OpenAI says Anthropic’s gross-revenue accounting overstates by ~$8bn, and that on OpenAI’s preferred net calc Anthropic is $22bn annualised against OpenAI’s $25bn. The category is now mature enough to argue accounting. Tokens served is a good one. Or tokens served by second.

Source: TechCrunch. Background: Friday Four #12: Anthropic is Cheap at 0.10 PEG-R, May 2026, the applications-company-in-a-lab-coat frame. Now in operating-profit form.

3. NVIDIA Reported The Best Chip Quarter In History. The Stock Fell.

Onto another AI winner: maybe the biggest winner of them all: Nvidia. Some more big numbers for you all trading the FTSE all day: Q1 FY27: $81.62bn revenue, up 85% year-on-year, EPS up 214%. Plus, a $80bn share buyback announced and the dividend hiked. Q2 guide above $87bn. So far so, scale-pilled.

But, and look, I don’t know anything about stonks, I am a god-to-honest pick a surfer and wave type, but the stock closed barely positive on the day, and slightly down the next. That’s on the best chip quarter ever printed, with the stock already up 40% since March. I asked Claude wtf:

• Jensen closed the China door publicly. “Expect nothing” on chip approvals. “We’ve evacuated that market.” Huawei is “very, very strong, they had a record year.” For two years, the optimistic read on NVIDIA assumed there was a China door that would reopen eventually.

• The $80bn buyback is the tell. You buy back $80bn of your own shares when you cannot find better use of the cash. That is a slow-and-mature company move. Dividend hike too.

• The customers are turning into an interlayer. Four hyperscalers plus xAI/SpaceX are 50%+ of NVIDIA’s revenue. All four are building their own silicon (TPU, Trainium, Maia, Graviton). Even Anthropic just signed with SpaceX-Colossus rather than buying NVIDIA direct. The end users don’t actually talk to NVIDIA anymore.

They seem plausible, I would expect nothing less from Opus. I think Huawei “very very strong” is the most interesting part of that. My thinking was always that Nvidia would win in China without sanctions, and that still might be the case, but it sounds like Jensen thinks Huawei is competitive. Nvidia need another growth story. Expect more this quarter from him on robotics then.

Source: CNBC China. Background: Friday Four #10, Apr 2026, the Jensen-on-China thread. Position now hardened.

4. Google Chose Glasses And A Camera, Anthropic Chose Your Laptop

And to round us off nicely this week, the other AI winner: Google. The most interesting things from the Google I/O:

1. Gemini Spark, the 24/7 cloud agent that lives in Gmail, Docs, Drive, Calendar, with MCP connectors to Canva and OpenTable and Instacart on day one.

2. Android XR glasses, both a display-free pair (camera, mics, speakers) and an in-lens display pair, made with Samsung and styled by Warby Parker and Gentle Monster, shipping this autumn.

3. Plus Gemini Omni, a unified video-AI model that takes text or photos or video or audio and outputs video.

I think what we are seeing here is the next battleground, or at least another battleground beyond agents.

• Category 1: Chatbots (mature). OpenAI won. Consumer

• Category 2: Agents (maturing). Anthropic won. Business.

• Category 3: XR (nascent). Google, Meta (models + devices) and Apple (just devices) have lead. Consumer.

Obviously Google Glass was too early, but it was obviously the future. With DeepMind and AI today, plus progress on the hardware side, we might finally be ready for actually useful smart glasses.

And note, we’re talking about Gemma, not Gemini. A pair of glasses can’t phone home to a 1.5T-parameter model for every head-turn or “what is this?”. Latency, battery and connectivity all break. The path is small distilled models running locally, with the big model in the cloud for the hard stuff (and AGI, and recursive self-improvement, of course). Google has been laying this groundwork: Gemma 2 introduced distillation from Gemini teacher models, Gemma 3 scaled it from 1B to 27B, and Gemma 3n is co-designed with mobile hardware partners, shares architecture with the next Gemini Nano, and handles audio/vision/text on-device at 4B active params. That’s the glasses stack: a 1-4B local model for fast perception and intent, Gemini-class inference in the cloud for anything substantive. You can’t run Gemini Pro on a 50g frame, but you can run something distilled from it.

Meta and Apple are going to follow Google here. Meta has Ray-Bans shipping for two years already and is basically there (and has its own on-device Llama variants). Apple has been waiting for somebody to make AI glasses work and now finally has cover (with Apple Intelligence already running 3B on-device models on iPhones).

Mark my words, AR is coming back big time, this time as the “glasses stack”. Expect a new wave of AR companies and a surge in funding for edge AI, predicated on smart glasses as the growth driver — efficient distillation, on-device multimodal models, custom NPUs, and low-power photonic or RF interconnect between glasses and phone. This is also where deep tech catches a tailwind: the glasses constraint forces everyone to care about edge silicon, sensor fusion, and the compute-per-watt envelope in a way the cloud-AI era didn’t. In fact, this next wave might be the “fun” part of AI (before we all lose our jobs), where we get amazing new consumer devices and new magical experiences. The Magic Leap whale on the basketball court, but for real this time.

P.S. How will we solve prescription? We won’t do it computationally. Probably need fast custom-made lenses I expect. Yea, watch this space…

Source: Google blog. Background: Friday Four #6: The New Sovereigns Are Plugging In, Mar 2026, the agent economy stack thesis. Form factor split is the new thing.

What else I’ve been reading

• 150,000-year-old humans living in West African rainforests. Nature study out of Côte d’Ivoire, blowing up the savannah-as-cradle story. Turns out we were in the wet forests too, much earlier. Hmm. So why are we all bipedal? Sheffield

• Three new phases of ice discovered in the past year, simulations say there are tens of thousands more. Compressing water at different rates and timescales gives “completely unexpected behaviour.” Including ice that conducts electricity, somehow. Quanta

• Bird retinas are the most expensive tissue in any animal body and somehow don’t use the oxygen advantage every other tissue runs on. Evolution is annoying. Quanta.

Cheers, have a good one yea.

You will get these on Thursday’s now. Thursday is the new Friday. Friday’s are the new Saturdays. Claude Code has increased productivity so much, the 4 day work week is real. It’s all very Economic Possibilities for our Grandchildren. Technology means more leisure time, amiright?

No obvs not, just more calls, more memos, more work. Busier than ever. Keynes's had it all wrong. Apart from his taste in art.

There’s a fashionable take in some circles that LLMs are a dead-end (often some parts of Government), they are too power-hungry, shit unit economics, diminishing returns on scale. And that the inevitable consequence is a great unwinding of datacentre capex as workloads migrate to the edge. I think the story continues: the “EdgeAI moment we’ve been promised for five years” finally arrives. 95% datacentre / 5% edge becomes 80/20, then 70/30, or whatever number lets you justify shorting Nvidia. Core versus Edge. Pick a winner.

A while ago Jonno and I argued the framing itself is wrong in The Compute Gradient:

“The opportunity is to stop treating this as a zero-sum race for megawatts and instead to shape the stack so each task finds its natural home.”

There is a real opportunity in new edge silicon: low-latency, low-power designs for inference at the device, the gateway, or the cell tower. That part of the bull case is right, and we’re actively looking at it. But the conclusion people draw, that we can wave off datacentre capex and pile into edge instead, doesn’t follow.

The fact is and will remain: the frontier lives in the datacentre, and increasingly it doesn’t come out. So frontier capabilities in science, ai research, and defence, will be in the datacentre.

1. AISI Says Cyber Capability Is Doubling Every 4.7 Months And Compute Just Became A Defence Technology

The UK AI Security Institute published on 13 May that autonomous AI cyber-task capability is now doubling every 4.7 months. That’s not really even a timeframe. In November 2025 they had it at 8 months. Then Claude Mythos Preview and GPT-5.5 came in over the 4.7-month line, AISI’s own framing is that it’s unclear whether this is a one-off step-change or a new, faster trend. Either way, the slope just got steeper. It’s accelerating!

Numbers. Claude Sonnet 4.5 succeeds 80% of the time on cyber tasks that take a human expert 16 minutes. The newer Mythos Preview checkpoint solved AISI’s hardest range (”Cooling Tower”) 3 times out of 10, the first model to complete it at all. AISI: “the length of cyber tasks that frontier models can complete autonomously has doubled on the order of months, not years.”

This is a UK government body. Its readers are MoD, NCSC and the Cabinet Office. We really need to internalise this and not hand-wave it away with LLMs are uneconomical and won’t reach AGI. The implication is that datacentres hosting frontier models are defence infrastructure now, full stop. Anton Leicht’s Cut Off this week (h/t Jack Wiseman) makes the policy version clear: frontier AI access will be rationed through some combination of security gating, compute scarcity, and US government leverage, and non-US allies should be treating datacentre investment as the price of staying at the top table. I will state it more clearly. datacentres are security and defence infrastructure. Yes it will use gas and all the water in the world. That is going to be the cost of staying at the top table.

This is what changes the edge-vs-core analysis. The strongest models will increasingly not be exposed via open API, they’ll be served through restricted, jurisdiction-controlled channels to vetted customers, or held back entirely. Edge silicon, by construction, runs the previous generation: the model small enough to fit in your power and memory envelope, the model someone was willing to release the weights for. Genuinely useful for a large class of workloads. Maybe real-time robots and drones. But not useful if the question is whether your country can do AI-augmented cyber defence, intelligence analysis, or any of the strategic decisions that need frontier capability.

A country or company that opts out of datacentre capex on edge-economics grounds is opting out of the frontier. The edge runs yesterday’s intelligence on today’s silicon. The datacentre runs tomorrow’s intelligence.

So if frontier datacentres are defence infrastructure, who in Europe is actually building them under terms Europe controls?

2. The EU Just Awarded Sovereign-Cloud Contracts To Companies You Should Know

The European Commission awarded its sovereign-cloud procurement (€180m over 3 years) to 4 consortia. Look at some of these names:

• Post Telecom (Luxembourg’s incumbent telco) leading with CleverCloud and OVHcloud, the two big French independents.

• StackIT (Schwarz Group’s German enterprise cloud, owned by the same family that owns Lidl, see recent Cohere/Aleph Alpha tie-up).

• Scaleway (Iliad’s French cloud, the Niel ecosystem).

• Proximus (Belgian incumbent telco) leading with S3NS (the Thales/Google JV that runs Google Cloud workloads under French sovereign control), Clarence (Belgian cloud), and Mistral. Yes, the AI lab is in the consortium.

Interesting that AT&T and T-Mobile aren’t in the AI game in the US. And here they are at the very centre isn’t it? Lobbying is a hell of a drug.

Also, Mistral on a sovereign-cloud cap table is another signal that they did a hell of a job hiring Government Affairs people early. Issue #12 covered the Cohere/Aleph Alpha merger as the warning shot for European AI consolidation. This is what comes next, the European sub-frontier model gets bundled into the sovereign procurement. The Commission used €180m to put European cloud providers on production-grade public-sector contracts at scale.

The other interesting bit is who didn’t win: T-Systems, Telefónica, the various BT and Vodafone bids. The list is short and aggressively European.

But you read those names, and are you bullish on the EU AI ecosystem?

Are you balls.

Datacenter Dynamics

3. Fractile Raised $220m, Long Known, But Look At The Company It Is Keeping

The round that everyone knew about for the last 6 months, UK inference chip startup Fractile closed a $220m Series B on 13 May, co-led by Accel, Factorial Funds, and Founders Fund. First commercial chip 2027. Anthropic putting in orders apparently.

Fractile is the in-memory compute play. I;ve written about this lots and lots. But for those not listening at the back, calculations happen inside the memory itself rather than shuttling data back and forth to a logic die. Same architectural family as Synthara from the interview series and the wider in-memory cohort I tracked in Issue #8. The memory-wall thesis from Issue #10’s Photonic Foundry Fallacy says this is where the next compute bottleneck breaks. Think memory, not logic. I’m reminded of Project Zeus.

“Why can’t marketing be an arm of sales?”

Fractile was in Issue #1 at £100m, so the $220m headline is the SotF readership’s “long known” candidate. The interesting bit is the company it’s now keeping. Same week, Dutch startup Euclyd (backed by the former ASML CEO) is openly fundraising at €100m+, claiming 100x inference power efficiency over Nvidia’s Vera Rubin. Cerebras raised $1bn in February and signed a $20bn / 750MW compute deal with OpenAI. Nvidia bought Groq for $20bn in December. Pricing compute in MWs again folks.

The Nvidia rivals have collectively raised something like $5.5bn this year. The open question is whether any of them ships at scale before the H300 successor lands. Fractile says 2027. Tight.

Dealroom had European AI chip startups at $800m and US ones at $4.7bn as of mid-April. Fractile's $220m takes the European total past $1bn. The gap is still real, and the European inference cohort is finally large enough to count as a category.

Bloomberg

4. The Next UX For AI Doesn’t Have A Browser

And finally, UX.

Google shut Project Mariner on 4 May. That was a bet on the visual-screenshot UX paradigm where AI clicks buttons like a human and humans watch through the browser. It lost.

What beat it is API plus CLI. Claude Code on the terminal. OpenAI’s Operator behind an API surface. Anthropic’s Quick Mode for Claude in Chrome, which bypasses the tool-use protocol entirely and uses single-letter commands instead of JSON, 3x faster, 4-minute browsing tasks done in under 2. All very 3 minute abs. The fastest agent runtime is the one that doesn’t pretend to be a person with a mouse.

This one was a simple one of money. A vision agent pays for screenshot tokens, DOM serialisation, and round-trip latency per click. A CLI call costs essentially nothing per operation. Frontier models get better at structured reasoning faster than they get cheaper at vision; and so gap widens. Quick Mode’s 3x is an early data point on a curve that keeps moving. Every “agent browser” product pitch from the last 12 months (Comet, Atlas, Mariner itself) was a fight versus physics.

The mouse and the browser are a translation layer for eyes. Agents don’t need eyes. The terminal is the agent-native runtime because it was already designed for programs talking to programs with humans peering in. Claude Code’s traction is the proof: developers recognised this immediately. Every agent browser is reinventing bash, badly, at 10x the cost.

One qualification. The Mariner type products were quite betting on vision. They were also betting that the browser is the universal abstraction over legacy software that lacks APIs. That’s partially wrong (most valuable workflows have APIs or will) and partially still alive (the long tail of internal enterprise tools, government systems, weird SaaS). Vision agents likely survive as a fallback for the messy 15%.

I think the interesting insight here is CLI/API UX puts pressure on every SaaS company to expose a real API or lose to a competitor that does. Agent-readability becomes a moat or a death. In a world where agent-to-API traffic dwarfs human-to-screen traffic, the infrastructure layer (compute, interconnect, the data centres carrying all the traffic) is where the money ends up (pending diligence).

Technobezz

Also Worth Your Time

My friend Tom Walton Pocock published the UK IP Amnesty Act on Wednesday with two proposals I strongly support.

1. A 30-day Universal IP Exit forcing UK universities to license out research IP on a standard 10% ordinary-equity term, fully dilutable, no royalties or milestones, with the IP reverting if the founder doesn’t raise £1m within 12 months.

2. A UK IP Register modelled on the Land Registry, mandatory at patent issuance or transfer. Both proposals fix the exact bottleneck Fractile and every UK deep-tech spinout in the Cloudberry pipeline has had to claw through to escape the university. UK research output is world-class (1st on WIPO’s H-index, 3rd on highly-cited papers globally). The commercialisation pipeline is the gap. Walpo names the constraint and proposes specific surgery.

Read it. and share with Wes Streeting for the next Kings Speech plz.

Thanks for reading as blood usual, do a couple of likes so I can give up the VC game and do this full time would you?

Hey folks, how you living? You alright? Greetings from Helsinki. I like the sauna, I do not like the karaoke. That’s my take. More travel musings as they come in.

Joined the etn boys again this week who wanted to know why memory stocks were stonking. I can’t wait til Siemens or Mistral buy them out like OpenAI and TBPN to “Accelerate the global conversation around AI.” lol.

Anyway, what do I know about memory stocks? I asked my knowledge base, and it reminded me of this absolute beauty:

Absolutely nailed it. I am a hunda going to actually start trading, I used to think that Mr Market is pricing in all this correctly, and that an index fund will capture basically all the upside. Aside from those sweet sweet Anthropic secondaries I got. But I dunno man CPO is coming fast and for me Coherent, Marvell and MACOM are just sitting there. BUT 100% NOT INVESTMENT ADVICE. DO NOT LISTEN TO SOME BLOKE IN HIS PANTS ON THE INTERNET.

I jest of course, I am just British. I actually know a lot about this.

—

1. SpaceX is an AI lab now

The big story this week is that SpaceX is an AI lab now. I keep forgetting and the news keeps reminding me. This week’s reminder was that SpaceX wants $55bn for a first-phase semiconductor facility called Terafab, scaling to $119bn across later phases. Intel joined as manufacturing partner on 7 April. The line runs on Intel’s 14A node. USA! USA! USA! Reshoring folks, it’s happening. Customers, per Musk on the Tesla call, are Tesla self-driving, Optimus humanoids, and SpaceX’s own AI data centres.

This is as clean a vertical integration story as you are gonna get. Rockets to launch satellites that beam internet to data centres that train models on chips SpaceX fabs itself, on a process node co-developed with Intel. It’s the Musk Industrial Complex.

Issue 11 closed item 1 with “can’t wait until TSMC’s biggest customer is SpaceX” (I meant biggest customer). Two weeks later SpaceX filed to bypass TSMC entirely. Faster than I thought.

The vertical-integration playbook isn’t new. Google has been training Gemini on its own TPUs in its own datacentres for years. Microsoft has Maia silicon, Cobalt CPUs, Azure, and now MAI models alongside the OpenAI relationship. Apple owns silicon-to-OS-to-services and sort of shipped on-device inference via “Apple Intelligence”. Omg, honestly do you remember this? Jesus christ, it’s so cringe even writing it. What did Apple think that they could co–opt the word “AI”. It’s absolutely mental and not enough people are taking the piss out of Apple for it in my humble opinion.

The hyperscaler-as-vertical-bundle is becoming the dominant shape of AI now. What SpaceX adds is the layer below the datacentre. Rockets, satellites, Starlink ground stations, the chip via Terafab, the model via xAI. Google, Microsoft, and Apple stop at the datacentre door. SpaceX owns everything from the launch vehicle to the inference call. Here’ another call I’ll make: Intel will be acquired by one of these guys in <12 months.

Source: TechCrunch. Background: LFG (for semiconductors) — Dec 2025, where I argued the foundry layer was where the next decade of value capture lived. SpaceX took that further than I had in mind.

2. Big Fund is leading DeepSeek

Meanwhile in China. China Integrated Circuit Industry Investment Fund (the Big Fund, the state chip fund) is “in talks” to lead DeepSeek’s first-ever external round at roughly $45bn. Tencent and Alibaba have FOMOed in. Valuation has roughly doubled in a few weeks; the figure was $20bn in late-April apparently.

And here’s what everybody is missing (can you legit write that even in irony anymore?), DeepSeek’s models are optimised end-to-end for Huawei silicon. I didn’t realise this

So the Big Fund (chips) leading the round (model lab) bundled with the cloud distributors (Tencent, Alibaba) is an integrated industrial play.

It’s happening guys, we have the emergence of a US and China AI stack from silicon > model > hyperscaler. This is the China variant of the bundle I described in A Specific Theory of Sovereign AI back in October. The thesis was that sovereignty isn’t a model layer or a chip layer, it’s the state-level integration of compute, models, and distribution into one industrial “object”. DeepSeek + Big Fund + Huawei + Tencent + Alibaba is that, in cap-table form.

We need to fix up and look sharp here in UK/EU. Consensus Capital I’ve called it in the past. See below:

Source: Bloomberg | TechCrunch. Background: A Specific Theory of Sovereign AI — Oct 2025, the framework piece. Sovereignty as state-level vertical integration, not regulation.

3. Europe wrote a letter

I know. A letter. Ha, ha, ha, Europe, right. Well, maybe not. This week, the CEOs of seven of the continent’s biggest industrial firms: ASML, Airbus, Mistral, Ericsson, Nokia, SAP, and Siemens, published a joint op-ed. They want 3 things: Simpler AI Act. M&A rules that let European companies actually grow, and protection from “subsidised rivals with very strong market penetration in the EU.” Good stuff, we live in a new world, let’s start acting like it shall we.

The letter was published before a meeting with von der Leyen and lands ahead of the Tech Sovereignty Package on 27 May, which will reopen parts of the AI Act and pair it with the Cloud and AI Development Act and Chips Act II. The first formal EU-level definition of “digital sovereignty” is due in the same package. If us Europeans (correct) do anything well, it’s wordsmith. So you just wait for this package. It will be sooo well written, you aren’t going to believe it.

“But, but isn’t this just industry begging for less regulation” Sort of. But read the letter. They’re asking for the Brussels version of what SpaceX and the Big Fund are doing this week. Concentrate capital, shield from imports, and simplify the compliance. They’re asking for permission to build the EU bundle.

Last week’s item 2 was Cohere absorbing Aleph Alpha at $20bn, which I called the sub-scale-incumbent-bailed-out-by-government version of European sovereignty. This letter is a bit different but comes from the same rationale.

Tech Sovereignty Package on 27 May is the test. If it is just some sort of procurement directive plus disclosures, then I am afraid it’s probably over. I think maybe it’s already too late if the scaling hypothesis materialises, but we gotta at least try. I’m thinking just off the dome, no AI:

• Silicon: ARM, Infineon, STMicro, Graphcore?

• Tools: ASML, Carl Zeiss, Trumpf, Siemens EDA, ASM, IMEC

• Networks: Ericsson, Nokia,

• Vertical Models: Mistral, Wayve, Helsing?

No leading frontier general model, leading node fab or memory player (FMC soon?). Obviously ton more gaps, but I’m trying to simplify here.

Source: Reuters via Yahoo. Background: How to Invest in AI Sovereignty — Sovereign Albion w/ Andrew Bennett — Feb 2026, the playbook for European sovereignty assembly Lawrence and Andrew worked through. Sharper now that ASML is signing op-eds.

4. MCP at 97 million

Not too many moons ago, Anthropic shipped a tools-protocol called MCP (Model Context Protocol) as a side-project. Just a short 18 months later it has 97 million installs as of March 2026 (so probably 100m+ now) and every major AI provider is shipping MCP-compatible tooling, inc. OpenAI. It’s the default mechanism by which agents connect to external tools, APIs, and data sources. Same broader rev cycle, Anthropic ARR overtook OpenAI for the first time at $30bn run rate vs OpenAI’s $24bn, primarily on enterprise agentic workloads.

Anthropic gave MCP away on purpose. It’s the side-project version of the first Stripe API. Or Android (AOSP for those who know). By making it open and plug-in-anywhere, Anthropic ensured every tool integration in the world routes into Claude as easily as it routes into anything else. Every CRM connector, every database adapter, every SaaS hook written to MCP works on Claude by default and works on the others as a port. Open source is a great strategic tool when wielded carefully.

I remember Gary M tweeting last year that Anthropic was “losing the platform war” to OpenAI’s GPTs. .@Gary, mate.

A few thoughts on where open source could go next. Or another way of saying I hope you aren’t running this startup or backed one that is a nearest neighbour:

1. Agent identity and authentication. MCP solved how agents talk to tools. It didn’t solve how tools know which agent is calling, which human is behind it, with what scopes, and what audit trail. Today it’s bodged OAuth and service accounts. It’s a shitty UX problem right now. Solve asap pls.

2. Skills format. Anthropic already runs Skills internally (the SKILL.md system). Open-sourcing it as a portable manifest would do for agent capabilities what npm did for JS libraries. Registry follows. Every Skill written for Claude becomes portable, but Claude is the reference implementation. The most Stripe-API-shaped.

3. Agent-to-agent handoff protocol. MCP stops helping when agents call each other. Orchestration is a mess of proprietary frameworks (LangGraph, CrewAI, and Swarm). A standard envelope for “agent A delegates task T to agent B with context C and budget B,” with provenance, partial-result streaming, and cancellation, is very MCP-shaped. Google is gesturing at this with A2A.

Source: Anthropic on MCP | Background: You Like AI Agents, You Are Gonna... — Feb 2025, before MCP was a thing. The piece argued the agent runtime would converge on a single protocol layer and that whoever owned it would own the value. Aged well.

I would appreciate any feedback on the newsletter.

If you could indicate directionally good or bad?

—

If you missed it: What Are We Even Counting — Issue 12, last week’s piece on Anthropic at $900bn and the absorbed pure-play AI labs. Sets up most of this issue

made me take a moment and have a bloody good think about the ai labs. like, what’s going on out there? lots of big numbers and announcements, but what does it all mean basil?

1. Anthropic, $900bn, Too Much?

Anthropic is aiming for a $900bn valuation as per Bloomberg, which would leapfrog OpenAI’s $852bn as the world’s second most valuable private company.

These seem like big numbers for a “startup”, and indeed, they are, they would make Anthropic the 9th most valuable company in the world. Tether lol.

Is it worth it?

Let’s take a look at valuation, revenue and growth shall we?

Tether lol.

The metric I care about it PEG ratio (Price/Earnings to Growth). I’m using Anthropic’s own forward guidance of 4x annual growth. Couple of interesting thoughts:

• Meta (0.22) — great “growth at a reasonable price” story in the set, if they can get a decent model, then…

• TSMC (0.31) — the AI infrastructure play with manufacturing moat

• NVIDIA (0.34) — even at $5T market cap, growth justifies the multiple

Anthropic 30x P/Rev seems chunky but PEG-R at 0.10 looks… cheap?

Source: Bloomberg article on the round | SaaStr on the revenue gap

2. Cohere, or What Mark and Satya Figured Out

With OAI and Anthrophic redefining achievable growth rates, you can see why investors are throwing money at AI labs (more to come). But it’s not quite as easy as it seems.

This week, Cohere announced it was acquiring Aleph Alpha at a combined $20bn valuation, with Schwarz Group (Lidl’s parent if you can believe such a thing) leading a Series E at €500m and the Canadian and German digital ministers in the room for the photo op. The press release framed it as a “transatlantic AI powerhouse” and a “sovereign alternative to American players”.

I mean, sure, we can all say words I guess, but come on? Cohere was at $7bn in September. Adding Aleph Alpha, last priced at $585m before the founder left in late 2025 and the company pivoted away from frontier models. + a German retail conglomerate, + government endorsement, and you get, hold on… carry the one, and equals = 20bn? Hold on, let me just check my workings, and… 20bn? From 7? Triples? Triples is it?

I did it. I got to use the gif in context! All thanks to Lidl. Dreams do come true.

But anyway, like this whole sovereignty story is going sour for me. If it’s about bailing out companies that can’t compete on the world stage, then look, this ain’t gonna work. If we want to do “sovereignty” right, then do it the UK SovAI way, back the best companies like Callosum and Ineffable Intelligence.

But notice the underlying question. Even with hundreds of billions of capital flowing around, even with sovereign endorsement, even with a German retail empire’s distribution channel, Cohere still can’t be at the frontier. Why?

Notice what Mark Zuckerberg and Satya Nadella have figured out that Governments haven’t. Inflection wasn’t a $650m purchase of Pi the chatbot, it was Mustafa Suleyman plus a team to run Microsoft AI. Adept wasn’t an acquihire for the agent product, it was David Luan and Niki Parmar going to Amazon. Character wasn’t $2.7bn for the consumer chatbot, it was Noam Shazeer (who co-wrote Attention Is All You Need in the first place) back at Google. And can you guess what Meta’s $14.3bn for 49% of Scale AI was? Bingo for Alexandr Wang as CEO of Meta Superintelligence Labs. People not Money.

The frontier isn’t a money problem. Microsoft and Amazon and Google and Meta have the hundreds of billions sitting around. It’s a talent problem. Maybe 100 people in the world can lead a frontier training run, and the hyperscalers are buying them out one $1-15bn cheque at a time? Who will Apple buy?

Inflection (Microsoft acquihire). Adept (Amazon). Character ($2.7bn licensing to Google). Stability (gone). Aleph Alpha (just absorbed). AI21 alive at $1.4bn but small. Imbue last raised October 2023 and hasn’t come back. Mistral is the European holdout, $830m in March for Paris and Sweden datacentres, $13.7bn valuation, ARR they need to grow 50-100x to earn the price. European exception, or next on the absorbed list. ASML to acquire/merge with Mistral anyone?

Source: TechCrunch on Cohere/Aleph Alpha

3. Buying A Hedge: Silver, Ilya, Yann, Mira

And here are the labs that didn’t get acquihired. David Silver (AlphaGo, AlphaZero, AlphaProof, the man, etc) closed a $1.1bn seed round at $5.1bn for Ineffable Intelligence on Monday, a UK lab with a thesis that explicitly rules out current LLM scaling. He wants reinforcement learning without human data. Sequoia and Lightspeed co-led, Nvidia put in $250m+, the UK Sovereign AI Fund came in. Everyone buying option value on Silver’s brain inventing the next paradigm, or at least a hedge on scaling stalling.

This was the same trade for Sutskever’s $32bn SSI (no product, ~20 staff). Same for Yann LeCun’s $4.5bn AMI Labs in Paris (pre-launch, “world models”, LLMs are a dead end per the founder). Same for Mira Murati’s reportedly upcoming $50bn round at Thinking Machines. The same talent class running the inverse trade, taking the cheque to start the lab instead of getting absorbed into one.

Counterpoint, also out this week. The UK’s AI Security Institute published their evaluation of GPT-5.5’s cyber capabilities. Headline numbers: GPT-5.5 cleared 71.4% of expert-level cyber tasks, against 52.4% for GPT-5.4 the prior generation, 48.6% for Anthropic’s Opus 4.7, and 68.6% for Claude Mythos Preview. It solved a reverse-engineering challenge that takes a human expert about 12 hours, in 10 minutes 22 seconds, at $1.73 of API cost. Their own quote:

The UK government’s own safety institute, not an OpenAI fan account.

Which makes the bull case on the frontier labs (Anthropic, OpenAI, GDM) double as a bear case on the founder bets. Silver, LeCun, Sutskever are explicitly wagering current scaling has run out. AISI just shipped a report saying it hasn’t, yet. If they’re right, Silver’s $5.1bn at zero product is a bet against the current scoreboard.

Source: TechCrunch on Silver/Ineffable

4. EU Chips Act II, It’s Absolutely Fab-ulous

Now on to the real power brokers in today’s world: The EU. Bloomberg yesterday: the European Commission is going to give itself the power to invest directly in cross-border manufacturing projects under Chips Act II, due late May. Until now, the Commission could fund research and approve member-state aid, but couldn’t write a cheque straight to a fab. After Chips Act II, it can. Sounds dry. It is not dry.

Issue #6 (20 March) said AI sovereignty was about owning chip supply, not protecting software. €80bn flowed in against a €43bn target, so the direction was right. What I got wrong was which fabs would deliver. I had Europe’s flagship industrial-policy projects in mind.

• Intel Magdeburg, €30bn: cancelled July 2025. The leading-edge logic flagship. New CEO Lip-Bu Tan called prior plans “unwise and excessive” alongside a $2.9bn Q2 loss. Germany had committed €10bn in subsidies. CEO-speak for “no demand.”

• Intel Wroclaw, $4.6bn assembly plant: cancelled the same day. Was the packaging hub for Magdeburg’s wafers. Without Magdeburg upstream, Wroclaw lost its reason to exist.

• STMicro/GF Crolles, €7.5bn: paused. The automotive FD-SOI flagship, €2.9bn of French aid behind it. Killed by European EV demand weakness; ST pivoted to “China-for-China,” GlobalFoundries to the US.

• Wolfspeed Saarland, €3bn: collapsed. The EV power-electronics SiC flagship. Partner ZF pulled its stake October 2024. Wolfspeed filed Chapter 11 June 2025, restructured under Apollo, $4.6bn of debt written off. EV slowdown plus Chinese SiC competition.

Capital was on the table for all four. Demand wasn’t. Again, capital cannot solve all ills. You actually need talent to make stuff and customers to buy it.

Interestingy what’s working is everything that isn’t a flagship. Diversified JVs with multiple anchor customers, research pilot lines, next-generation substrate plays.

• TSMC’s ESMC Dresden (€10bn, JV with Bosch/Infineon/NXP) is on schedule for equipment install H2 2026. Three anchor customers across automotive and industrial. Mature 28/22nm + 16/12nm.

• ST Catania SiC (€5bn) is heading into production. ST is the incumbent expanding existing turf, not a speculative new entrant — same SiC bet as Wolfspeed but on resilient ground.

• The Imec NanoIC pilot line in Belgium opened in February at €2.5bn (€700m EU + €700m Flemish + €1.1bn industry, ASML lead). Research and pilot capacity, not commercial production.

And the next-gen substrate startups, CamGraPhIC, Black Semiconductor, SMART Photonics, Q.ANT, maybe Ephos, are the kind of bets Chips Act II’s direct-investment power can actually place. 3 of those 4 are photonic. None of them is a 2nm logic fab.

We will get capacity. Now, do we have the talent and the markets?

Source: Bloomberg on Chips Act II

—

And please if you reached the bottom, if you can think of one person who would like what I write, please tell them this is a serious newsletter for serious people

The starting point for today’s interview is that AI inference is getting more expensive, not cheaper.

• GPT-5.5 doubled price last week: input went from $2.50 to $5.00, output from $15 to $30 per million tokens. The Pro tier is six times that.

• Anthropic kept Opus 4.7 on the same rate card but the new tokenizer works through roughly 35% more tokens per task, so the bill is up.

Yes the commodity tier (Gemini Flash, GPT-4o Mini) is getting pushed down to the famous near-zero marginal cost, we are getting sub-$0.30 per million output tokens.

But the frontier, where agentic workloads run, pricing is up. An agentic coding task plows through 20K–200K tokens. An agent team is 7x, because each teammate maintains its own context. Per-token price drops, but total bill goes up.

TL;DR: Chatbots are getting cheaper, but agents are getting more expensive.

So, what will we do? Well, we need new silicon designs to use less energy basically. The AI pricing problem is an energy problem. Every matrix multiplication in a transformer moves data between memory and compute, and it’s the movement, not the compute, that eats most of the power. The AI workloads in the data centre are over-parametrised by orders of magnitude. We know this because pruning, quantisation, and distillation all reliably shrink frontier models with barely any loss of performance. If the models were sized correctly, many of them could run on the device in your lap. The architectures are not designed for efficiency, so they don’t.

This is why so much interest in new silicon and chip designs. You can find primers on neuromorphic, analog, photonic and others in the back catalog. Today, I am speaking to one of the world leaders in the analog space. Rather than moving bits more efficiently, it stops doing much of the arithmetic the traditional way at all.

EnCharge AI, spun out of Naveen Verma’s group at Princeton in 2022, builds chips that perform the multiplication inside the memory array itself, using switched-capacitor circuits on standard memory (SRAM) cells. No exotic silicon processes required. Their first chip, the EN100, is returning from fab imminently. The claim: 10–20x energy efficiency at the same process node versus legacy digital.

That improvement matters because the price curve I just described isn’t fixable in software alone. Per-token efficiency can keep dropping inside a digital architecture, but the slope is bounded by the memory-compute boundary. If every token requires shuttling weights and activations back and forth across that boundary, your floor is set by the energy cost of the bus, not by Moore’s Law. The only way to drop the floor materially is to change the architecture. Compute-in-memory is one route. Photonic interconnects, which we covered with Hitesh from Phanofi, are another.

We’ve covered compute-in-memory on State of the Future before, notably with Manu from Synthara on ComputeRAM. EnCharge is the analog cousin. Shwetank Kumar, EnCharge’s chief scientist, came up through applied physics at Caltech, spent years on experimental quantum computation at IBM Research, ran engineering at Intel, and joined EnCharge after writing publicly about AI inference efficiency at AI After Hours. I spoke to him about analog’s comeback, why “gigawatts” is the wrong unit to measure anything with, the first AI product that’s genuinely uneconomical to ship, and where analog silicon lands in the compute gradient.

What Did I Learn?

1. Noise is fought in hardware and software. EnCharge’s switched-capacitor MAC sidesteps the noise and variability that have historically killed analog compute, which is half the battle. The other half is the software tool chain: a proprietary PTQ workflow and a compiler designed with the silicon. Without both halves, the efficiency gains don’t survive the trip from HuggingFace to the chip.

2. Video generation is the first AI product that’s genuinely uneconomical to deliver, re SORA, which flips the role of efficiency hardware from “faster” to “required to ship a viable product.”

3. Measuring AI data centres in gigawatts is a vanity metric that hides a collapsing gross margin. The honest measure is something like tokens-per-second-per-watt, or voxels-per-second-per-watt for video. Once you look at it that way, the Frontier Labs “build more gigawatts” is really just capex amortisation dressed up as strategy.

4. The over-parametrisation problem is the cause, not a symptom. Pruning, quantisation, and distillation all prove that frontier models have many times the parameters they actually need. That over-parametrisation is what forced everything into the data centre in the first place. Fix the parametrisation and a lot of what currently sits in a rack has no reason to.

The Interview

This interview has been edited for brevity and clarity.

The compute gradient

Lawrence: Hey Shwetank, let’s start big picture, the compute gradient, I call it. Core, and edge. The history of computing has been a constant push and pull between where workloads sit, mainframe to PC to smartphone, now to the hyperscale datacentre. As it stands, it seems like all training will happen in datacentres and the vast majority of inference too, no?

Shwetank: There are a few threads to pull on. First, you have to think in terms of workloads, which respond better on edge versus the data centre. Second, what resources are available, where. The reason everything has collapsed to the data centre is that nearly all the models we have are incredibly over-parametrised. By that I don’t mean the infrastructure they run on. I mean the number of parameters needed to hold the same amount of information.

We know this emphatically because if you take these models and prune the connections, you can get rid of a majority of them with no performance loss. Quantisation is next: get to four-bit and the models barely lose any performance. Combine these techniques and they still work. Distillation is another piece of evidence: knowledge distillation with a much smaller student model still works. If you have an over-parametrised model with many times the parameters it actually needs, you have no option but to run it in the data centre. That’s where the heavy infrastructure runs. No one has that in their laptop.

Lawrence: Right. So why hasn’t edge AI actually happened? Google has Gemma running locally on Android. I don’t see anyone building with it. Is this a demand problem or scaffolding problem?

Shwetank: I haven’t worked directly with Gemma, but when I look at the benchmarks I’m starting off unimpressed. Look at Qwen, and specifically Qwen’s Coder model that came out recently: about the same ballpark in size, but it punches above its weight on the benchmarks. Are benchmarks the be-all and end-all? Not really. You have to make sure they actually work for your use case. Once a model gets to a certain level on the benchmark, it’s much more about the scaffolding and harness around it, specifically how you manage the context. That’s another reason open source matters: it lets you figure out how to manage the context yourself. With Claude or OpenAI, you basically get the context you get. With Claude Code pricing escalating and similar dynamics elsewhere, you’ll see causal drivers move in favour of open-source models. At least, that’s my hypothesis.

The analog bet

Lawrence: Okay, let’s talk about your specific bet. Analog. Most of the others went digital. D-Matrix, Fractile. You and Mythic are basically the only ones still on analog. Everyone broadly agrees that near and in-memory compute will happen in the next few generations, but analog still has skeptics. What did you solve?

Shwetank: Caveat first: I’m not the hardware expert, but this is why I got behind the technology. We have a Von Neumann architecture where memory and compute have been separate, with a bus that makes the data shuttle back and forth. It became very clear from back-of-the-envelope calculation that that’s where most of the energy is going. The integration question, how and how much to integrate, isn’t a new story; bringing memory and compute closer has been an ongoing project with the L1, L2, L3 cache hierarchy.

When you look at the energy-efficiency drivers, one is collapsing the distance between memory and compute. The other is that even doing that doesn’t get you far enough. You want to extract every bit of efficiency you can given the power walls we’re hitting. Analog was one of the options. Analog has always had noise issues, variability at the edge, or required a specific marginal process not available in pure silicon. The real innovation in Naveen’s group was sidestepping that. The architecture we have doesn’t rely on any specific silicon processes. It’s purely switched capacitors. With switched-capacitor circuits, all you need is good control of the metal lines, which you have as the processes progress. You use a standard SRAM cell built in those processes and calculate the charge being deposited. That makes our technology robust to noise, and we have five or six generations of test chips that demonstrate it.

Lawrence: Right. So give me numbers. What does this actually buy you?

Shwetank: We have to be cautious about which layer we’re characterising. When I say a few hundred TOPS per watt, I mean it at the MAC level, multiply-and-accumulate.

Once you wrap an NPU around it, with typical digital architectures (or, more cheekily, legacy digital architectures, since these are the biggest companies in the world) you get a few small TOPS per watt for the entire GPU. Digital in-memory compute gets you into the low tens of TOPS per watt, 10 to 20. By the time you get to analog in-memory compute iso-node, you’re at mid-tens of TOPS per watt for the whole NPU at the system level.

Lawrence: Right. But is that even the right metric? Shouldn’t we be measuring tokens per second per watt? TOPS/W doesn’t really measure anything customers care about.

Shwetank: You’re getting to something that really bugs me. I did an entire rant on why everyone in the Valley is measuring compute in gigawatts. That has perverse second-order effects. You can’t afford to do that in a world where US energy consumption was flat and is now increasing because of data centres. Gigawatts is a pure input. It tells you nothing about the value you’re getting out.

The way I break it down: TOPS per watt is the silicon comparison, at the MAC level and the NPU level. Then go a level above and ask whether you can run the same algorithms. If they’re textual, tokens per second per watt becomes the relevant metric. If they’re video, it’s frames per second or voxels per second per watt. It’s always possible to cheat by throttling performance to look more efficient, so you really want to plot a Pareto curve: per watt, per second, and what’s the useful unit of work, tokens, pixels, trajectories. At different levels you have to be methodical about what you’re measuring. And ideally it shouldn’t be a vanity metric, which often it ends up being. Measuring data centres in gigawatts is, in my opinion, unequivocally a vanity metric.

Lawrence: Right. So then what’s the theory of the case as to why Sam and Elon are both pushing gigawatts? Because it seems kind of remarkable. Profitability is directly capped by their opex, and they don’t seem to be working particularly on opex. And yet on the other hand, they’re talking about pushing more input. What’s your theory?

Shwetank: These are very smart individuals and they’ve been very successful, so I won’t try to second-guess what’s going on in their minds. But here’s how I look at it. You can think of this in terms of whether it’s a winner-take-all market. If it is, one way to scare everyone else away is to make massive commitments to data centres. We’ve seen this in pharma, where the marginal cost of producing the next pill is very low and most of the cost is research. Once you’ve done the research, you amortise it as much as possible by building a big factory and saying, I’ve made this huge capital investment, no one else dare set up another factory, because I want to flood the market with pills.

I have a different point of view. The marginal cost of producing a token is not infinitesimal. It’s significant. In the post I wrote on the financials of Anthropic and OpenAI, you can see their gross margins coming in around 30 to 40%. These aren’t margins at which you’d call yourself a software company.

Lawrence: Right. It’s the end of zero marginal cost software. That reframes a lot of the capex debate. The days of SaaS margins are over. Okay, let me pull on your software. Are your quantisation techniques proprietary?

Shwetank: We have our own quantisation within the company. We can take an arbitrary model off HuggingFace, run it through our PTQ (post-training quantisation) workflow, and out pops a model that’ll work on EnCharge hardware. There’s a tool chain behind it. We’ve made investments in the compiler tool chain and the quantisation tool chain, and that gets coupled with our hardware and shipped out into the world.

Lawrence: Are we talking prefill or decode? Because you’re already seeing a split, aren’t you, in the inference workloads.

Shwetank: Absolutely. I’ve written extensively about how the data centre is getting disaggregated. We have, or will have, solutions for both. Initially, EN100 and EN200 themselves are more focused on compute-dominated workloads. That’s where you instantaneously get 10 to 20x energy efficiency iso-node. But we do have a decode story coming up as well.

Lawrence: Right. The photonic folks have a similar problem. You compute in photons beautifully but you don’t have photonic memory, so you constantly cross ADC and DAC boundaries, which eats pretty much all of the benefits depending on the size of the algorithm. So to what extent is ADC a big chunk of your power budget too?

Shwetank: That’s a very fair question. A little out of my day-to-day realm, but I’ll tell you this much: yes, we’re very thoughtful about ADC design. With every incremental bit you add, it can become a significant power hog. That’s why we report power at the MAC level and the NPU level separately. In addition to ADCs, other things knock off your efficiency as you zoom out to the system level. You have to be careful not just about the matrix multiplications, which is a solved problem for us at this point, but also how you do layer norms and other non-linear functions. Every time we discuss adding a bit of precision for quality, we have to think carefully about the power trade-offs. The systems we’re designing are, hopefully, at that sweet spot. It helps that Naveen’s team in Princeton have designed analog circuits all their life.

Going to market

Lawrence: Okay, fine, I was surprised reading about Encharge that you’re leaning into client PCs. Not drones, not cameras, as I’ve seen with analog, IMC, neuromorphic, etc. Mythic sits in surveillance, Halo and Accelera in cameras. Why client PC? Seems weird.

Shwetank: So far we’ve focused on client PC use cases. The energy efficiency and power envelope would be appropriate for drones and cameras as well. It’s the GTM motion we haven’t leaned into that much, yet. As time comes around, that’s stuff we’ll probably look at.

Lawrence: Right, but I guess what I am getting at a bit more is that a lot of edge AI startups started trying to sell an edge chip and ended up selling into the data centre. It’s the biggest infrastructure demand of our generation. Why sell into a market where the demand isn’t really clear versus a market where they will bite your arm off for singl digit per cent reductions in power.

Shwetank: On the larger strategic questions I’ll defer to Naveen, but my point of view is that we don’t have the option to say, this is a market we’ll stay away from. Over time we’ll probably support all of these markets. The questions of timing and sequencing are being talked about internally. We see it divvied up into three large markets. Data centre, which is already up and running and super-competitive. Client, which is essentially PCs and laptops. And physical AI, which isn’t really one market, it’s three or more disaggregated markets, and you have to make a bet on which will take off. If I were the decision-maker, I’d take a real-options approach to sequencing these bets.

Lawrence: Okay, I want to go back to something you said in passing, because I think it’s really important. You said video generation is currently uneconomical. I guess we already knew this with SORA being taken off the market.

Shwetank: Absolutely. This was one of the memos I’d written. I know video generation is uneconomical, and partly the market is constrained because of that. It’s a reasonable use case to target. But you always have to balance: a reasonable use case where you can unlock pent-up demand versus a market with significant tailwind, where everyone is currently doing agentic workflows. You spread the bet between something held back that you can unlock, and something already moving that you can ride.

Lawrence: That’s a useful frame. Speaking of expensive workloads, your gross-margin argument assumed a rough steady state in compute per query. o1-style reasoning and extended thinking burn 10-100x more compute per output token. Either that makes efficient inference dramatically more important, or it pushes the hard reasoning back to the cloud and leaves the edge doing cheap fast tasks. Which way does it land?

Shwetank: Test-time compute makes me very bullish. When you go from a model emitting 200 tokens per answer to one emitting 20,000 because it’s reasoning through the problem, per-watt-per-token becomes the critical metric. The whole conversation shifts from “how do I get a faster answer” to “how many joules did that answer cost me, and can I sustain that across millions of queries.”

Lawrence: And the cloud versus edge split? Doesn’t longer reasoning push everything back to the data centre?

Shwetank: That ignores what metrics matter for the interesting reasoning workloads, which are latency- or privacy-sensitive or both. Robotics doing multi-step planning. An agent on a phone reasoning about your calendar and inbox. An AR system reasoning about what you’re looking at. None of those tolerate a 5-second round trip to a data centre, and most can’t send the context to the cloud in the first place. Extended thinking actually makes the edge case stronger, because doing it in the cloud at that token volume becomes prohibitive on economics and latency. Cloud will run the largest reasoning models, but a meaningful slice of reasoning workload pushes outward.

One more thing worth flagging. Another piece of test-time compute is parallel sampling. Many reasoning approaches generate multiple candidate traces and select among them. That’s embarrassingly parallel, exactly where compute-dominated architectures shine.

Architecture, risks, and the unlock

Lawrence: That’s a useful answer. On your role specifically, chief scientist is different from chief architect or head of engineering. What are the big chunky things you spend time on?

Shwetank: It’s really great to have the partnership I have with the executive team. We’re all technologists with our own areas of specialisation, so even when I’m focused on one specific area, I know Kailash, Naveen, and others are covering the rest. Each of these areas is very fast-moving.

The charter I have, and the things I spend most of my time thinking about, comes down to three things. First, the race between model architectures. Which of these architectures, and which techniques within them, is going to win? Some come up as papers and just take off. Mixture of experts is an example: pretty much every model right now is a mixture-of-experts model. We have to think carefully about what that means for our chips and how execution happens. The moment it’s a mixture of experts, some layers shift from compute-bound to memory-bound.

Second, emerging architectures. There’s a recent paper where, instead of stacking transformer blocks on top of each other, they make a recurrent loop using the same block. Super interesting in terms of energy efficiency and parameter count. The interplay with our quantisation has to be thought through if this takes off. We have to work on it and see whether it actually does, and whether it’s a bet worth making.

Third, taking all these new techniques and models and seeing them through to application. If we’re going to support a video generation model, what are the core applications? How does it land in the market? Are we working with video-model training companies, or taking open-source models and working with final customers? Actually landing all the way to market is one of the things we have to think about.

Lawrence: Right, on architecture races, let me flip the question for you. Which of the currently-hot research directions is the nightmare case for switched-capacitor analog compute? Diffusion language models? Mamba-style state-space where the matmul stops being the bottleneck? Test-time reasoning with dynamic workloads? I’m after the physics-level failure mode, not the competitive one.

Shwetank: Honestly, none of those are the nightmare. Our architecture is built around the fundamental primitive of linear algebra, which is matrix multiplication. That primitive isn’t going anywhere. Diffusion, Mamba, test-time reasoning, they all bottom out in matmuls. What changes is the ratio of compute to memory traffic and how the workload is scheduled. Compute-dominated regimes are where we win biggest, 10-20x at iso-node. Memory-dominated regimes are harder for everyone but don’t disadvantage us disproportionately either.

Lawrence: Right, so at the algebra level you’re safe. Where’s the actual risk?

Shwetank: It’s more boring and lower-level: noise immunity, fab variability, whether our analog circuits keep scaling cleanly through standard CMOS process nodes the way digital does. That’s what we have to demonstrate in a product. If that translates, the linear algebra primitive is safe regardless of what model architecture wins. The interesting architecture work, by the way, lives at the system level: how the hardware execution path and software toolchain accommodate new model architectures quickly. Full-stack engineering, in other words, and it’s what we spend most of our time on.

Lawrence: Right, broader question. The current equilibrium, over-parametrised models, everything in the data centre, open source chipping at the edges, what breaks it, and when? I can see four candidates: token pricing crossing a viability threshold, sovereign-AI rules forcing local inference, sub-100ms agentic loops, or a genuinely capable sub-10B open-weight model. Which is the unlock?

Shwetank: All four are real, but each has a different causal driver and pace. The equilibrium doesn’t break in one event. It erodes from multiple sides at once. Token pricing is playing out in real time. Cloud providers have been subsidising tokens and customers haven’t been price-discerning, but that changes as high-volume agentic use cases come online in the next six months. Sovereign AI regulation moves defence, healthcare, and finance into local inference earlier than otherwise, but it doesn’t reset the whole industry. It’s a tailwind.

Lawrence: So latency is the actual unlock?

Shwetank: It’s the most immediately impacting one, because cloud cannot solve it. The speed of light alone makes it impossible. The moment a meaningful product category emerges where the agent has to close a loop faster than a network round-trip allows, the architecture decision is forced. Robotics is the obvious one, but interactive AR and on-device agents that act on your behalf in real time get there first, because the consumer hardware already exists.

Lawrence: And the sub-10B model side?

Shwetank: Sub-30B more accurately. That’s the catalyst that makes the latency unlock actually shippable. Deploying a 70B model on a phone or a robot, even with quantisation, is hard. You need the model side to meet the hardware side. The trajectory is faster than most people expected a year ago: Qwen, Gemma, Llama derivatives, and the whole distillation toolchain have compressed capability into smaller footprints at a rate that surprised me. So the unlock arrives as three things at once. Latency-bound use cases need capable small models, small models need efficient hardware to run, and efficient hardware needs an open-weight ecosystem to have something worth running.

Lawrence: Last one and I will let you get off to more important things, the EN100 is coming back from fab soon. What should we expect to be visible from the outside?

Shwetank: The specific numbers will be on the product page, so I won’t pre-announce. The short version: the efficiency we’ve been quoting at the NPU system level is real hardware, not simulation. That’s the whole point of taping out a product chip. We’ll keep iterating. We have a conveyor belt of test chips running in parallel, so it’s a matter of iterating on the architecture and the designs, and co-developing the software at the same time. The next generations aren’t waiting on this one.

So What?

The first takeaway, placed next to Manu from Synthara on ComputeRAM and Hitesh from Phanofi on photonic engines, is that this is another data point in the same underlying bet. AI’s energy wall is real, and the fix is about reorganising where arithmetic happens relative to where data lives, rather than pushing transistors smaller. The names vary, compute-in-memory, in-memory compute, analog, photonic, near-memory, but the architectural idea is the same. The post-Von-Neumann decade is starting whether the incumbent GPU companies like it or not, and each of these companies is a different probe into how it might actually play out.

The second is the gross-margin point, which Shwetank delivered in passing but deserves its own paragraph. Anthropic and OpenAI are running inference at 30-40% gross margins. That’s closer to semiconductor economics, or utility economics, or, depending on the quarter, chemical-plant economics. The industry response, build more gigawatts, is, as Shwetank says, a vanity metric. It amortises the capex. It doesn’t fix the gross margin. The only thing that fixes the gross margin is efficiency at the tokens-per-second-per-watt level. Which is to say, EnCharge is selling the unit economics of AI inference.

The third is the client PC point, which is where I was most sceptical going in and most interested coming out. Selling chips to consumer OEMs is famously brutal. Long design-in cycles, price compression, a handful of oligopsonist buyers. The easier path is to sell into hyperscalers, and it’s striking how many of the original “edge AI” companies have pivoted there. Shwetank’s framing, that data centre is up-and-running and super-competitive, that client PC is what you focus on first, that physical AI is a real-options problem, is refreshing because it acknowledges what the D-Matrix and Groq trajectories have made plain. The data-centre inference market is already over-fished, and a chip company needs a second path. Whether client PC is the right second path, rather than robotics or industrial vision, I’m honestly not sure.

Where I’m not fully convinced: timing. The “video generation is uneconomical” claim is the most commercially interesting thing in the interview, and it’s also the one most exposed to software catching up before the silicon arrives. If frontier video models keep getting 2-3x cheaper per year through architecture, compression, and better training alone, the “efficiency hardware saves the economics” window closes before EnCharge can ship enough silicon to matter. EnCharge is betting the window is still open by the time EN100 and its successors are shipping in volume. I think they’re probably right. But it’s a timing bet, and timing bets are where semiconductor companies come unstuck most often.

Find out more at enchargeai.com. Shwetank writes “AI After Hours,” sharp stuff on AI inference economics, worth a subscribe whether you care about silicon or not.

New thing I am going to try, on and off. Going to start going back through old State of the Future essays and actually check what I predicted versus what happened. Twitter has pretty much beaten the “write confident call, pretend you never said anything if it did not pan out” habit into everyone who writes on the internet for a living (apart from Derek Thompson who is the man). For me it’s kale and score cards. Coming back. Dating the predictions. Keeping the receipts.

Starting, obviously, with one I got right. Because I can. TSMC reported this week and the essay I want to revisit is Can We Make Enough AI Chips? from November 2023, in which I argued the binding constraint for AI chips would never be GPU logic, it would always be packaging (CoWoS) and memory (HBM). Well was I right? Am I now a god-to-honest semiconductor investor?

—

1. TSMC Q1 2026

TSMC reported last Thursday. $35.9bn revenue, up 40.6% YoY. 66.2% gross margin on a business that burns $56bn a year on capex. Net income up 58%. Full-year guidance raised to 30%-plus top-line growth in dollar terms.

HPC is now 61% of revenue. Smartphone is 26%. For years the smartphone SoC business was the main event and HPC was the kicker. Now HPC is more than 2x smartphone, grew 20% sequentially this quarter, and smartphone actually fell 11%. Apple used to write the biggest cheque. Now NVIDIA. A sign of the times. Can’t wait until the biggest customer will be SpaceX.

Also, read this: 2026 capex at the upper end of $52-56bn is more than half of what TSMC spent across the three years before this combined. C.C. Wei called it a “multiyear AI megatrend” on the call.

So, E21 score card. I argued back in November 2023 that the binding constraint for AI chips would never be GPU logic, it would always be packaging (CoWoS) and memory (HBM). On this call Wei confirmed CoWoS capacity is sold out through 2026, HBM3 and HBM3E are fully allocated, NVIDIA alone has over 60% of the packaging line, and TSMC is going from roughly 35,000 CoWoS wafers per month at end-2024 to 130,000 by end-2026. Near-quadruple in two years. Ten to twenty percent of the $56bn capex is going to advanced packaging alone. Bottleneck moved from wafers to packaging in 2023 and has not moved since. The 2027 story is interconnects and panel-level stacking (the CoPoS pilot line finishes in June). Two and a half years later the call still holds. Call it a win. I literally can’t bold more relevant stuff. “Chips” aren’t the bottleneck. “Packaging” is. Repeat after me.

Source: TSMC Q1 Investor Relations

2. The Entry Rung, Again

Meta is laying off 10% of its meat sacks as Zuck calls them (8,000 people). Microsoft opens voluntary retirement to 7% of US staff. So we have Q1 tech layoffs at 80,000. So “AI is here and it’s taking jobs”

Callback to Issue #4. Block did roughly 4,000 layoffs in February (from over 10,000 to under 6,000) and the narrative was “AI is eating the point-of-sale business.” I said at the time that Block just overhired during Covid and was using AI as cover. Three months on, the read holds. Meta is the same story at Meta scale. Microsoft’s voluntary retirement is genuinely a first in 51 years of the company, but the people taking the payout are senior directors with 20 years of tenure, so this isn’t AI-exposed entry-level engineers, this is balance-sheet cleanup.

Relevantly, British Progress published “AI and the UK Labour Market: The Evidence So Far” this week. Key details:

• Three years after ChatGPT they find no macro signal of AI displacement in the UK.

• Occupations with higher AI exposure have grown faster than least-exposed ones, across all four exposure measures and both data sources they use.

• IT business analysts up 38% since 2021, programmers up 18%. Call-centre workers down 19%, telephone sales down 23%.

Same AI exposure, opposite outcomes: what matters is whether AI augments the task or substitutes it. Bessen’s old bank-teller point: ATMs made branches cheaper so banks opened more branches, so teller employment kept rising for two decades. Programming expands when debugging gets cheaper. Call-centre scripts are fixed demand and the work just disappears. Elasticity.

But. Crane and Soto at the Fed Board find US coder employment three percentage points per year below trend since ChatGPT. British Progress cannot replicate it for the UK (pre-ChatGPT window too short, occupational classification changed in 2021). Brynjolfsson’s Stanford Canaries paper has 22-25 year olds in AI-exposed roles down 13% relative, young software devs down 20% from the late-2022 peak, and workers 30+ in those same roles up 6-13%. The signal is there. It is at the entry rung.

The story is the 22-year-old with a CS degree who is not getting the interview. Invisible, because you cannot put a face on a job that never happened. No politician or trade union is standing up for the people who didn’t get jobs. Meta firing senior PMs in Menlo Park is noise and actually muddies the water here because it looks like a job automation story, but it’s not.

Fwiw, this basically the scenario I sketched in Occupational Downgrading and Unbundling the Job a while back. The macro looks fine but the entry rung collapses.

Meta is not cost-cutting. Microsoft is not cost-cutting. Capex is going up. Way up. Meta capex jumped from $72bn in 2025 to $115bn in 2026. Microsoft spent $88bn on AI in 2025 and is going +++ this year. Big tech is shifting spend from labour to capital, from payroll to tokens (that’s the line).

Is anyone measuring this properly? Someone needs to be. The useful thing would be a firm-level ratio of dollars-per-function on the capital side versus the labour side, tracked over time. Semafor ran a piece on Tuesday arguing AI token spend is starting to rival labour costs at the enterprises that have actually adopted, which is directional but not really quantified.

The curve you would expect, if this is substitution rather than augmentation, is probably not linear.

• Phase 1 is what we are in now, token spend up and labour flat, productivity per worker up (British Progress has UK software GVA per worker up 16% since 2019 against 0.4% economy-wide). Capital side grows and labour side holds.

• Phase 2 is when model reliability crosses the threshold for a specific category of work, labour for that category starts falling in absolute terms, capital spend spikes to replace. That is the entry rung in software post CC and Codex.

• Phase 3 is when reliability crosses the broader threshold and the substitution goes wide. Flat, flat, flat, cliff. Which is why the macro data will keep looking fine right up until the point it doesn’t.

Solow, 1987: “you could see the computer age everywhere except in the productivity statistics.” Electricity took forty years to register in output data. Computing thirty. We are three years into generative AI. The British Progress closer basically says “check back in two or three years” which is the absolute right take. But. Here’s my thing as a seed stage VC not an economist. I can’t afford to wait for the data. I have to make bets before the data. My bet is the 3 phase thing above. Forthwith to be named: The State of the Future Tokens Versus Labour Framework. SOFTTVLF.

Source: British Progress: AI and the UK Labour Market: The Evidence So Far

3. Frontier More Expensive

That was a long one. I’ll keep the rest short so you may continue with your day. Thursday morning: GPT-5.5. Friday morning: DeepSeek V4. Some numbers:

• GPT-5.4 was $2.50 input and $15 output per million tokens.

• GPT-5.5 is $5 input, $30 output per million tokens.

That is a straight 2x price jump in six months.

And GPT-5.5 Pro is $30/$180 which is 6x the base. Opus 4.7 kept the same rate card on paper, but the new tokenizer produces up to 35% more tokens for the same text, so the real bill is up roughly 35%.

DeepSeek went the other way. V4-Pro at $1.74 / $3.48 per million tokens, V4-Flash at $0.14 / $0.28. Apache 2.0 open weights, 1M context, 1.6T-param MoE with 49B active. SWE-Verified 80.6, essentially tied with Claude’s 80.8, at roughly a tenth of the Western sticker price. Plus a 90% cache discount on repeated prefixes, which is the single biggest cost lever on anything agentic.

The lazy read is “AI is getting cheaper.” It is and it isn’t. The commodity tier is in freefall (Gemini Flash at $0.08 / $0.30, GPT-4o Mini at $0.15 / $0.60) but the frontier is pricing up into its demand curve. And the frontier is where agentic coding actually has to run. A single agentic task plows through 20K-200K tokens. Agent teams multiply that by 7 because every teammate maintains its own context window.

So the per-token price keeps dropping. But the truth is that it doesn’t matter much. The task runs longer, the bill is higher, and the only genuine frontier price cut this week came from a Chinese lab giving away the weights.

If you have been watching the capability curve thinking “surely this will all be basically free by 2027”, it will be, for chatbot use cases. For the actual agentic work you would want to run, it will still be $5 per million input tokens. You will just use 20 times more of them.

Source: Simon Willison on DeepSeek V4

4. Tragic Twenties

Finally, just a hell of a great read as usual from Derek Thompson. Not really an AI or tech story, just read it. “If America’s So Rich, How’d It Get So Sad?”.

The headlines are: lowest US ranking ever in the World Happiness Report. Lowest consumer sentiment in the 70-years. Federal Reserve worker satisfaction at its lowest since 2014. Trust collapse across government, the military, the CDC, education, religion, pick any institution. Consumer prices up 25% between summer 2020 and summer 2025. Housing moving at roughly twice pre-pandemic pace. Thompson calls it the “permademic“, the pandemic’s second-order effects that never actually went away.

V.V.V Interestingly, English-speaking countries got clobbered hardest (US, UK, Canada, Australia). Portugal, Italy and Spain saw happiness actually rise in the 2020s. Turns out if you do not have 25% inflation you do not get the misery. Spain again folks. Amar was right.

The key point is that inflation is just so pernicious. No-one notices price increases going up and down. They only notice the increase. People live in their housing cost up 50%, their weekly shop up 25% on a few years. Thompson’s point is that feelings drive what comes next, more than economics does. I think this is right. And if you will indulge me and allow me to tie it to AI. If I wanted to change the messaging around AI, I think this is where I would start. Tell the deflationary stories. E.g.

• Private tutoring is £60-80 an hour. Khanmigo is £4 a month. So a year of unlimited AI tutoring costs less than one hour with a human.

• Basic legal advice from a solicitor, £200-300 an hour. A competent first-pass answer to 80% of consumer legal questions, free, in your browser, right now.

• GP triage used to be “wait three weeks on the NHS or pay £100 private”, now it is “ask the chatbot and it is usually right about whether to worry”.

Source: Derek Thompson: If America’s So Rich, How’d It Get So Sad?

–

That’s for indulging me, as usual.

If you missed it:

• The Photonic Foundry Fallacy, the biggest opportunity in computing, still

• Detecting Proteins in Blood with Photonics, Prateek on ambient health, magnetic beads, why the optics are the bottleneck

Preventative health. You have a semiconductor and photonics fund and what’s the first deal you do? A bloody biosensor. Because photonics. Because photonics has miniaturised and got so cheap we can now use photonic chips to sense more and more things. One of those things is proteins. And proteins are where diagnostics has been stuck for forty years.

So here’s your pitch.

Most cancers get caught late, you feel a lump and a scan for something else picks up a shadow. The GP orders bloods and something comes back bad. By the time any of that’s happening, the tumour has already been growing for close to a decade. Which matters, because the survival curves are ugly. Caught at stage 1, breast cancer has a 99% five-year survival rate. Stage 4, under 30%. Pancreatic is worse: 40% at stage 1, 3% at stage 4. The gap between “caught early” and “caught late” is, for most of the cancers that actually kill people, the gap between treatment and palliative care. Long before the lump and the scan, and there is a trail, it’s just hard to see it. A pinhead tumour is already leaking proteins into the blood.

Right, now PCR. Polymerase chain reaction. Remember Covid tests? Those. Take one strand of viral DNA, drop it in a tube with an enzyme whose entire job is to copy DNA, and thermal-cycle it thirty times. Each cycle doubles what’s there. 2^30 is a billion. One strand of virus becomes a billion, and it’s impossible to miss. Forensics, paternity, virology, cancer genomics, the whole lot is based on this innovation.

Proteins though. There’s no protein polymerase, no biological or chemical machine that takes one protein and makes two. DNA’s whole thing is self-copying, it’s literally what the molecule is for. Proteins are downstream of that, they’re the thing the copying machinery was built to produce. They have no copying machinery of their own. So if your blood is carrying twelve molecules of some cancer-signalling protein in five litres of fluid, that’s twelve. You detect what’s there, at the concentration the body was kind enough to provide.

The workaround, for decades, has been enzymatic amplification. You tag the protein with an enzyme, the enzyme spits out thousands of glowing molecules, and you read the glow. The catch is that enzymes are noisy. They fire randomly even when no target is present, they care about temperature, they degrade between batches. The best enzymatic assays get you to femtomolar detection (a concentration of 10^-15 moles per litre), and then they hit a noise floor.

And so, Proteins.1 then, instead of making the signal chemically louder, they read the same molecule over and over on a photonic chip, letting certainty accumulate from the repetition. Magnetic beads, antibodies, and photonic elements replace enzymes. There is a line of sight to three orders of magnitude more sensitivity than the best current platforms, from femtomolar to attomolar detection (a concentration of 10-18 mole per litre). Small numbers, very small numbers.

What Will You Learn Today If You Decide to Read On?

1. Proteins can’t be copied. PCR works because DNA copies itself a billion times over. There’s no equivalent for proteins. Every diagnostic either accepts the noisy enzymatic floor or finds a different amplification mechanism. Proteins.1 is making the mechanical-cycling bet, on a photonic chip that amplifies signal through repetition.

2. Biology is the bottleneck. Miniaturised lasers, photonic elements, and photonic integrated chips are moving fast. The real bottleneck is getting a single target molecule in a two-microlitre blood sample to physically find the right bead with the right antibody in the right amount of time. You can’t break the laws of diffusion and binding kinetics.

3. The platform is molecule-agnostic. The chip doesn’t care whether it’s binding a protein, a DNA fragment, or a metabolite. Same chip, same beads, swap the antibody. That’s the path from a protein diagnostic to a tabletop multi-omics box.

The Interview

Lawrence: Alright, Prateek. What up? Give me the ninety-second version. What’s Proteins.1 actually doing that’s different?

Prateek: In a cell, or in the body, it’s really hard to tell whether what you’re seeing is the onset of disease or not. A foreign body turns up, is the immune system going to suppress it, or is this actually the start of something? So we take a blind-canvas approach. We look at the whole picture.

Lawrence: Meaning what, specifically?

Prateek: We analyse one protein in relation to the 4,000 others that are present. One protein expressed at ten copies doesn’t mean anything on its own. If you can see what’s happening to all the other proteins at the same time, you can start telling a story about what the biology is actually doing.

Lawrence: Right. So this is the parallelisation point. The more proteins you can test in a single run, the richer the picture. And I assume there’s machine learning sitting on top of that, running against historical datasets to say these two proteins together mean X or Y?

Prateek: Correct. When I put on my night vision goggles and look at the sky, I can see all the stars. But there are stars behind those stars. There are galaxies. There are nebula clouds. Astronomy keeps progressing by building larger and better cameras that look at different wavelengths and get the composition of the stars rather than just their position. Our photonic chip is the same. More pixels, more wavelengths, more depth of data.

Lawrence: So the simple version: better camera, more you can see.

Prateek: Exactly. There’s a limit, of course. Miniaturising antibodies and functionalising beads is where the wall is. But that’s why we do this.

Lawrence: That’s interesting, because my head went to the enormous development in miniaturised lasers, comb lasers, faster modulators, the whole photonic integrated chip stack. You’re telling me that’s less the bottleneck. The real wall is antibody coatings and magnetic bead chemistry.

Prateek: True, Lawrence. Sometimes we forget that biology exists for a reason. The molecules know what they’re doing, even if the engineers think otherwise.

Lawrence: Fair. Say more on the binding side. What’s the actual constraint?

Prateek: You have a magnetic bead, and let’s say a single target molecule in this very small sample volume. How is the bead going to find the target? How is it going to carry it to the antibody spot? These things depend on residence time. How the sample is agitated, how you get them in close proximity. We cannot break the laws of chemistry. If it takes a certain time for the binding event to happen, that’s the time.

Lawrence: Right. Good old diffusion. That problem is older than photonics by a few hundred million years.

Prateek: (laughs) Exactly.

Lawrence: Right. Let’s jump to ten years out. What does Proteins.1 look like? What kind of company are you building?

Prateek: In my previous research I looked at biomarkers in sweat. Sweat is rich in signal, but by the time the proteins make it from blood through the skin into sweat, most of them are degraded. Quantities are tiny. What you’d want is something that can see what’s left on your fingerprints, in your breath, on the surfaces you touch.

Lawrence: Okay.

Prateek: In ten years, this is part of your kitchen. It’s the tricorder from Star Trek.

Lawrence: (laughs) The tricorder. Sure. Hold on though. My head was going somewhere slightly different. Today we’ve got a tabletop device for blood draws. In some timeline that becomes a wearable. We move away from microneedles for glucose monitoring, even. Optics gets small enough that you can put this on your wrist, in your ear, in a ring, and everything aggregates into constant biomarker data.

Prateek: Yes.

Lawrence: But what you’re saying is more ambient than that. The device is around you, not on you. Fridge, bathroom, surfaces. Healthcare IoT.

Prateek: True. Ten years ago, Oura didn’t exist. We had basically no wearables. Apple Watch has completely changed how I think about my sleep, my heart. In the next ten years, technology like ours becomes an integrated part of life. Food quality control, what’s happening in your stool, your fridge telling you something has spoiled.

Lawrence: Your fridge knows the eggs have gone off before you do.

Prateek: That’s the vision. You’re not sitting in front of a box doing a weekly blood draw.

Lawrence: Okay, but to actually be a trillion-dollar company you can’t stay on the tabletop. How do you get from today’s box, through Gen 3 and Gen 4, to this ambient world? What’s the roadmap?

Prateek: It happens in steps. Today we can do simultaneous DNA and protein interrogation. First step is scaling up from single proteins to a hundred, then to the full proteome. That’s a big jump, moving from a diagnostic to an omics platform.

Lawrence: And the omics part, that’s where things get interesting.

Prateek: You replace mass spec, which is huge and slow, with a small tabletop device that does metagenome and proteome simultaneously. Single device.

Lawrence: Hang on. Single device doing proteins, DNA, and metabolites. I want to make sure I’ve got this. You can do that because the platform is coming from the protein side, where the hard problem is, and the genomics and metabolomics parts are more mature. So you’re doing the bit nobody can do today, and the rest either follows or you in-license.

Prateek: The good thing about our technology is that it’s molecule-agnostic. As long as something binds to it, we can detect it. We’re not relying on conventional PCR or sequencing for genomics. We won’t be in-licensing, we’ll be applying our own platform to those other omics.

Lawrence: So same chip, swap the binding chemistry, get different omics.

Prateek: Exactly.

Lawrence: Alright. When someone says omics to the average person, it doesn’t mean much. Give me the Monday-morning scenario. Five years out, still a tabletop box. I walk up to the machine. What happens?

Prateek: Two-microlitre fingerprick. Similar to a glucose test, almost painless.

Lawrence: Okay.

Prateek: That sample gets partitioned into a few million pixels on the chip. Each pixel corresponds to different DNA or protein targets. Between them, we think we can cover almost the whole genome and proteome from that one drop.

Lawrence: And what comes out the other end? I’m self-administering, I’m not a clinician. What does it tell me?

Prateek: Hopefully nothing.

Lawrence: (laughs) Hopefully nothing. That’s the right answer.

Prateek: If it finds something, it doesn’t say biomarker one and biomarker thirteen are elevated. It talks to your physician. Time gets booked for a checkup on kidney function. If it’s communicable, we can look at whether it’s spreading between families. If it’s cancer or Alzheimer’s, your clinician is notified and works up a plan.

Lawrence: So it skips me entirely for anything serious.

Prateek: For the clinical stuff, yes. The prediction should be early enough that you have options. We don’t want the box telling you that you’re going to die in three days.

Lawrence: Right. Like your Apple Watch doesn’t say you had a bad dream, it says you didn’t sleep as well as last night. Softer communication layer on top of harder data.

Prateek: Exactly that.

Lawrence: Okay. That’s one pathway, the clinical one. There’s a parallel pathway which is preventative, which I think is underrated. You’re low in X, eat more beetroot today. I use Zoe in the UK for gut health. Most of the reason people pay for that is to change their diet, to improve biomarkers through food and exercise. This would be another data source feeding that loop.

Prateek: True. That’s the more preventative side.

Lawrence: Two pathways, then. Clinical, through the physician. Preventative, through diet and behaviour.

Prateek: And drugs are getting more personalised. Earlier detection means softer treatments. Anti-inflammatories instead of chemotherapy and radiotherapy. Pharma can intervene at the point where disease is signal rather than tissue damage.

Lawrence: Sadly, I can’t have conversations any more without thinking like a VC. Here’s the question. Physics-based amplification replacing noisy wet enzymatic methods is a reasonable first-principles bet. Photonics is racing ahead, the components are getting cheaper and smaller. So assume I’m right, and this is where protein diagnostics should go. How strong is your lead?

Prateek: Speed to market. Speed to customer. Owning the IP.

Lawrence: Right, those are the basics. Everyone says that. What else?

Prateek: The three of us have taken ideas on paper through to products that are FDA-approved and used globally. That combination, having lived the story in complex medical devices and drug discovery, and now bringing it to a research-use and diagnostic platform, that matters.

Lawrence: Okay, so team + prior FDA approvals + patents.

Prateek: Granted US and Finnish patents, international filings pending.

Lawrence: What about the analogy you were using?

Prateek: The analogy I keep coming back to is cameras. When I was growing up, you sent film to a lab, chemicals developed it, weeks went by, the prints came back.

Lawrence: I remember. Just about.

Prateek: Then my first Sony digital camera. Fifty megapixels, instant. The whole world in your hand. Film became a niche almost overnight. Kodak is still a brand, but nobody processes film any more.

Lawrence: And Kodak the company basically disappeared.

Prateek: Almost. Canon and Fujifilm came in and took the market. That’s the part we’ll have to fight off commercially, with IP and being first.

Lawrence: Right. Paradigm shift is an overused word. Disruption is an overused word. But specifically applying optics to protein detection, that’s where the field hasn’t caught up. Most of the talent is still on the chemistry side.

Prateek: True.

Lawrence: If the advantage sits in optics and photonics, the traditional players have to build a skill set they don’t currently have. And photonic integrated chips have moved faster in the last five years than most people outside the field realise. That’s the tailwind.

Prateek: Innovations happen in silos. People forget there’s progress in the room next door that could be applicable to their own field. As someone who’s worked across semiconductors, photonics, and biology, I’ve had the chance to see that. You can walk to a photonics lab and understand what’s possible, then bring it back to biology.

Lawrence: That’s a great line. I might steal that.

Prateek: (laughs) It’s all yours.

Lawrence: Alright mate, last one. What does success look like for you? Not trillion-dollar company, I get that one. The thing you actually need to happen in the next two or three years to know you’re on the right path.

Prateek: Adoption by the clinical diagnostic labs, and by researchers finding novel early biomarkers. If those two things happen, we’ve done our job.

Lawrence: Labs and biomarkers. Good. Let’s hope we actually demonstrate early diagnosis and enable the preventative-health thing that’s been promised for a decade.

Prateek: That’s the plan.

So What?

The thesis is simple: photonics are now good enough to beat enzymes for protein diagnostics.

Seed-stage startup alert, obviously. The Proteins.1 thesis lives or dies on biology. The photonic chip is the easy-ish part. The hard part is antibody selection, bead chemistry, and getting a single target molecule to physically find the right bead in a sample that’s 99.99% not what you’re looking for.

Prateek said the quiet part out loud on the call: “we cannot break the laws of chemistry.” Which is the single best sentence a photonics-meets-biotech founder can say, because it’s true. The physics team can keep pushing pixels, the biology team has to push binding kinetics, and binding kinetics doesn’t benefit from Moore’s law.

The other elephant in the room is Theranos. Ambient diagnostics is a decade-overdue vision. Home blood tests were going to change everything in 2015, and instead burned down trust in the whole category. This time the photonic chips actually work, the magnetic-bead chemistry isn’t being faked, and the instrument is going to clinical labs before pharmacies.

The path from here is staged: single protein, then a hundred, then the full proteome, then multi-omics. That’s the play.

And then there is the Tricorder Vision. Ten years out, the device is around you. Your fingerprints, your breath, the surfaces you touch, the air you exhale. If the molecule-agnostic claim holds up, Proteins.1’s photonic chip becomes the sensor for that ambient layer, which is an order of magnitude bigger than today’s diagnostics market.

Morning ya’ll, sorry for the delayed issue. I mean, I wonder how many people actually read this as soon as it comes out? I mean, if you are, soz. Half term isn’t it.

If you didn’t see it already, I wrote a thing: https://stateofthefuture.substack.com/p/the-photonic-foundry-fallacy

“Not bad” and “I think you are wrong” are just some of the many comments I’ve had, and honestly, in this day and age, I’ll take it. I said the following:

• Frontier AI is copper-limited. Training tens of thousands of GPUs talking to each other means moving petabytes between chips every second. The 1960s-era copper interconnects can’t do it, too much power, too much heat, not enough bandwidth.

• Optics have carried data between continents for decades. Now the frontier is light between servers, chips on a board, chiplets inside a package, and eventually on the die itself. Every inward step is harder, which is exactly where copper gives up first, making the optical stack (transceivers, modulators, lasers) the most strategic component in AI infrastructure right now.

• Incumbents are betting on silicon photonics for sunk cost reasons. Tower spending $920m to 5x its SiPho wafer capacity. GlobalFoundries bought AMF in Singapore. TSMC tooling co-packaged optics on its 300mm line. SiPho slots into existing CMOS lines, so it’s the cheapest route to “we do photonics now” and the easiest story for investors.

• Startups are going exotic. HyperLight and QCi on thin-film lithium niobate (great modulator, no light generation). Ligentec on silicon nitride (ultra-low-loss waveguides, not much else on its own). Smart Photonics on indium phosphide (the only material that generates light natively, also the most expensive and lowest-yield). Three different fabs, all structured around one hero material.

• Contrarian bet: the organising principle is integration. No single material can modulate, generate and guide, so the CMOS mental model (pick a material, scale the wafer, drive down cost per die) breaks. Heterogeneous integration all the way down: orchestration, chiplets, manufacturing. He who integrates wins.

Give it a read. Otherwise, onwards as hard as possible, as always.

1. Anthropic Crosses $30bn ARR. OpenAI Rage-Posts. Now Do Margins.

Right, the story of the week. Anthropic’s ARR crossed $30bn per Bloomberg. OpenAI sits at $25bn. Anthropic’s own number was $9bn at year-end 2025. That growth curve looks like a ski jump. We wiling out over here. 1,000 enterprise customers spending >$1m a year, doubled from 500 in under two months. 8 of the Fortune 10 pay Anthropic. Bloomberg reckons Anthropic spent 4x less than OpenAI to train the models doing this damage, which will do some things to the comparable DCFs if true.

OpenAI had a public little meltdown. Their revenue guy ripped Anthropic for “building a narrative on fear, restriction, and the idea that a small group of elites should control AI.” Bold move, Cotton. Same day OpenAI told Axios that Microsoft “was holding them back” and they’re cosying up to Amazon now. (Also, for the record, Opus 4.7 shipped this morning, which Anthropic openly says is below the unreleased Mythos, see Issue #9.)

But Margins. Nobody is fighting about this publicly but they should be. Top line will keep breaking records. Anthropic’s a trillion-dollar company in waiting for sure. The open question is whether any of these businesses makes money once the VC subsidy runs out. Training cost advantage is fine but most of the cost is inference, serving, and the electrons, silicon, cooling and data centres you’ve put them in. Every software engineer on $200/mo Claude Code probably burns through unit economics that would make a SaaS CFO file a missing persons report.

Every AI story from here is a silicon story, a data centre story, or an energy story. What’s hard is gross margin, and gross margin is electrons. Which brings us nicely to…

Source: Bloomberg

2. OpenAI Gives Up On Stargate UK, Signs a London Lease Anyway

In September OpenAI announced Stargate UK with NVIDIA and Nscale. 8,000 GPUs to start, 31,000 if it went well. 6 months later, it’s been paused. Reading directly from the statement, “the cost of energy and the country’s regulatory environment.” But then they signed a 88,500 sq ft lease in London for a HQ of 500+ people. So not leaving Britain. Just refusing to put compute here. The offices are lovely, it’s the electrons that are the problem.

Here is the simplest argument about the next 10 years i can make. Pensions, the NHS, welfare, defence, all of it, is downstream of economic growth. That’s a Liz Truss argument sure, but I will raise you. Economic growth is now largely downstream of AI. AI is downstream of compute. Compute is literally downstream of electrons. That means every policy argument a British politician thinks is central: tax, immigration, planning, welfare reform, is actually downstream of “is our marginal price of a kilowatt hour going down.” Remarkably simple. Very hard to get politicians to say out loud because it makes most of their existing agenda look like fiddling while Rome burns. Which I can tell you does not make for good baseload.

OpenAI is just the first big customer voting with its feet on British grid economics. Ireland capped hyperscaler connections in 2022, the Dutch have a megadatacentre moratorium, Texas lives in ERCOT meme-territory. Everyone has made power expensive. Britain is doing it on principle. France and the Nordic hydro economies look smarter every month, which is why per CNBC Nscale (Issue #5, $2bn Series C) is still in the Stargate conversation — Nordic electrons are cheap, British electrons are not. When the British champion has to build most of its compute abroad, that’s the story.

In A Specific Theory of Sovereign AI last October i argued sovereignty is infrastructure, not models. The harder version eighteen months later: sovereignty is cheap kWh. Everything downstream. I mean, I am making the case that energy is destiny. Hardly a new argument, but one that we must remember in UK and EU, and fast.

Source: CNBC (Stargate pause) | CNBC (London office)

3. Fusion Is Having A Moment

As if i’d planned it. Fusion is the technology that, if we ever build it properly, makes the energy variable from item 2 go to approximately zero.

This week Helion’s (the Sam Altman fusion favourite fwiw) Polaris reactor hit 150M°C, roughly 3/4 of commercial operating temperature. Close. Pulsar Fusion (UK) ignited first plasma in Sunbird, their fusion rocket exhaust test rig for deep space propulsion. Yes, mate. This is proper ambitious stuff we should be working on. And ARPA-E committed $135m at the Energy Innovation Summit, the largest single fusion commitment in the agency’s history. TAE is running site surveys for a first power plant in the US… I mean what’s a survey I suppose. But still, it’s a datapoint!

Timelines still don’t match the hype. First plasma is not first power. Say it with me to avoid the hype. Sensible MWh of fusion on a grid is a while away, probably early 2030s at the earliest, almost certainly delivered by a 30-year state-backed consortium rather than a venture-backed startup. But maybe some of these VC-based startups become public-private vehicles anyway. If you want the primer, i wrote one pre-AI a while back.

If cheap electrons are the master variable for everything downstream (item 2), fusion is the single most important long-range technology humans are working on, more important than AI or semis or the entire AI-infrastructure gold rush, because all of those will still pay for kilowatt hours. Fusion is the holy grail because it dissolves the master variable.

Source: Everett Today (Helion) | Euronews (Pulsar)

4. Jensen on Dwarkesh: “It’s A Chip. They Can Make It Themselves.”

And the biggest interview of the week, Jensen Huang on Dwarkesh Patel, two hours, mostly about TPU competition, Nvidia’s supply chain moat, and whether the US should be selling chips to China. The China section ran forty minutes and it did not go well.

Jensen’s argument: comparing AI chip export controls to nuclear non-proliferation is “lunacy.” Selling Nvidia parts to China is fine because “we’re not enriched uranium, it’s a chip, and it’s a chip they can make themselves.” He called export controls a “loser’s mentality.” China already has 60% of global chip manufacturing, 50% of AI researchers, plenty of energy, so blocking them is futile. Dwarkesh, credit to him, pressed on the national security implications, specifically that H20-class compute shortens the offensive-cyber timeline (see Mythos’s zero-days in Issue #9). Jensen got visibly agitated. Agitated-Jensen is not a Jensen people had previously seen.

Two problems with “they can make it themselves.” First, per Chinese chip leaders in Tom’s Hardware two weeks ago, China is five to ten years behind on frontier AI chips and short of leading-edge lithography because ASML won’t sell them EUV. Let alone NA-EUV. So they can’t, at scale, for now. Second, the person arguing the controls don’t work is the person whose biggest growth market is currently locked up by them. The incentive is doing alot of the talking.

The Chip Letter called Dwarkesh excellent and Jensen evasive. Alec Stapp called the offensive-cyber answer misleading. Zvi mostly agreed at considerable length as always.

My read. Jensen has said he believes in AGI. Many times. In many keynotes. If AGI is real, the scramble for it is wartime, and wartime means picking a side. You don’t sell to both teams on the way to the singularity. So either Jensen doesn’t really believe the AGI stuff (it’s just the keynote line), or he does believe it and is obfuscating because NVDA has a lot of growth priced in and China is the TAM that closes the valuation gap. I’m afraid this is a peacetime CEO in wartime.

Source: Dwarkesh | Chip Letter | AOL (lunacy quote)

—

Thanks everyone for reading, I appreciate you and you are loved.

Byeeeeeeeeeeeeee

If you missed it:

• A Specific Theory of Sovereign AI — last October, still right, more so now

Using light to move data isn’t new. We’ve pushed photons through optical fibre between continents for decades, because copper falls apart at distance. What’s changed is how close to the silicon the optics are getting. The boundary has crept inward for years, from undersea cables to cross-campus links to inside the data centre itself. AI is accelerating that inward migration. The frontier now is light between servers in a rack, between chips on a board, between chiplets inside a package, and eventually on the die itself, where photons never have to leave the silicon. Every step inward is harder than the last, because shorter distances mean tighter packaging tolerances and denser interconnects, and that’s exactly the regime where electrical links are giving up first. Which is why the chips that do the conversion (transceivers, modulators, lasers, the whole optical stack) are probably the most strategic component in AI infrastructure right now.

Everyone is scrambling to build that photonics stack. The existing CMOS fabs are doubling down on silicon photonics (SiPho) because it slots straight into the lines they already run, using the same wafers, lithography and packaging tools. It’s the cheapest route to “we do photonics now,” and the easiest story to tell investors. Tower Semiconductor is spending $920 million to 5x its silicon photonics wafer capacity. GlobalFoundries acquired AMF in Singapore. TSMC is tooling up co-packaged optics on its 300mm line.

The startups are going full disruption. Silicon is the past, they say. For losers, they say. They’re picking an exotic material and building the fab and tooling around it. HyperLight and QCi on thin-film lithium niobate (TFLN), lovely stuff for high-speed electro-optic modulation but no use for generating or detecting light. Ligentec on silicon nitride (SiN), ultra-low-loss waveguides and not much else on its own (though they’re expanding to multi-platform). Smart Photonics on indium phosphide (InP), the only material that can generate light natively, and also the most expensive and lowest-yield to manufacture. Different strategies but the bet is still a single material.

Contrarian bet alert. What if they are wrong? I believe most of them are organising around the wrong principle. Single-material fabs are obviously useful, but the mental model imported from CMOS (pick a material, scale the wafer, drive down cost per die) breaks down when no material can do everything you need.

The organising principle should be integration.

It’s heterogeneous integration all the way down, folks. From the orchestration to the chiplets to the manufacturing. He who integrates wins the AI race. (Whadup Callosum)

Have a gander at that table from Gemini. Lots of letters etc. Doesn’t matter a great deal, unless you are a massive nerd! The “key takeaway” is thus: No single material appears in every row without a chunky compromise. That’s the core of my argument. I will dispense this advice now.

I. Analogy

The semiconductor industry has one overwhelmingly successful organising principle: silicon. Start with a silicon wafer, etch transistors into it, scale the process node. Everything in the same material system, on the same wafer, in the same fab. This is the TSMC business model: standardise the platform, designers innovate on top.

It’s natural to look at photonics and think the same playbook applies. Pick your champion material. Lithium niobate for fast modulators, InP for on-chip lasers, SiN for low loss, silicon for CMOS compatibility. Go deep, go vertical, optimise, and scale. Lovely business model you’ve got there. It would be a shame if, it was, disrupted….

The analogy relies on a property silicon has that no photonic material does: silicon does everything adequately. It switches, conducts, insulates, and amplifies. Not always best-in-class, but “good enough” across every function so that integration on a single substrate wins on cost and density. Photonics doesn’t have an equivalent. The physics forbids it.

You absolutely need III-V semiconductors (indium phosphide, gallium arsenide) to generate light. Can’t get around it. Silicon has an indirect bandgap; it physically cannot lase efficiently. You need lithium niobate or electro-optic polymers for high-speed, low-loss modulation, because silicon’s plasma dispersion effect hits a wall around 50–60 GHz. You need silicon nitride for ultra-low-loss waveguides. You need the OG semiconductor, germanium, for detection. That’s five or six materials for a basic photonic link: generate, modulate, route, and detect.

The closest thing to a photonic silicon is probably indium phosphide, which can lase, modulate, route, and detect on one substrate. Smart people argue it’s the true platform play. But InP’s waveguide losses are 10–100x worse than SiN, its wafers top out at 6 inches (for now, tbf), and its fabrication costs make it uneconomic for high-volume applications.

The obvious move would be to scale it: build 12-inch InP and watch the cost problem disappear. People try. InP is grown using a process called liquid-encapsulated Czochralski, which means dipping a seed crystal into molten indium phosphide and slowly pulling it upward while it rotates, letting a single-crystal cylinder (the "boule") grow downward from the seed. Defect density rises sharply with diameter, indium itself is expensive and you need more of it per wafer, and the entire fab equipment ecosystem is built for ≤6 inch with no volume demand to justify retooling. The 8-inch InP roadmap has been “two years away” for well over a decade. The constraint is crystal physics.

This isn’t a temporary problem waiting for better engineering, either. Silicon will never have a direct bandgap. Lithium niobate will never absorb light efficiently for detection. These are properties of the crystal structure. You can’t just throw money at it.

II. Multi-Materials

Alright hands up. You got me. It’s not quite as binary as I’ve suggested so far. Single-material cathedral vs. material-agnostic platform is obvs too clean. Nobody serious is truly single-material anymore. Tower does Si + Ge + SiN on its line. GlobalFoundries does similar. Saying the industry is stuck on single materials makes for a great old versus new story, but is a bit crude.

The real question is subtler as always: is multi-material capability a side project bolted onto a silicon photonics line, or is it the organising principle the facility is designed around? Most of the industry is on the first path. Tower’s adding materials incrementally. GF acquired AMF for silicon photonics capacity with some SiN capability attached. TSMC is extending its 300mm CMOS line to handle photonics. In each case, the foundation is silicon and other materials get added where customers demand them.

I’m arguing for the second path. A facility where the core competence is the process of combining materials, where the equipment decisions, the engineering hires, and the IP strategy are all organised around integration. A handful of European foundries (LIGENTEC, CSEM, and others) are getting closest, offering multi-material photonic services within a single fab ecosystem. But these are exceptions. In most cases, multi-material is a research capability grafted onto a production line that was designed for something else.

The gap between demonstrating multi-material integration in a lab and offering it as a reliable, repeatable manufacturing service is vast. I know that. But closing that gap is the opportunity.

III. Coupling

The silicon photonics camp argues the gap doesn’t matter. They say silicon is already shipping in volume, integration is a problem you can solve later, and one material plus a few external bolt-ons is good enough for the next decade of AI bandwidth.

Maybe? Intel ships millions of pluggable transceivers, the optical modules that slot into a switch port and convert electrical signals into light for transmission over fibre. Broadcom’s Bailly and Tomahawk 5 use silicon photonic engines. Cisco, Marvell, and a dozen others have silicon photonic products in hyperscale data centres right now. They bolt on an external III-V laser, accept the coupling loss, and move on. Good enough, we’ve got datacentres to build and tokens to serve, I don’t care about your 3 year roadmap to MVP.

Today, sure. But what about three years from now? Every external laser needs active alignment (expensive), discrete packaging (bulky), and optical coupling that burns 1–3 dB at every interface where light has to hop between chips. That doesn’t sound like much, but optical power budgets in a transceiver are tight. A few dB here, a few there, and you’ve smashed through the headroom that determines whether the link closes at all.

At today’s 800G per lane, the bolted-on architecture has enough margin to absorb those losses. At 3.2T per lane, where AI interconnects are heading by 2028, it doesn’t. Every dB lost at a coupling interface is a dB you can’t recover at the receiver, and at that bandwidth you don’t have any to spare.

You might also argue that photonics should follow the semiconductor playbook of specialisation: logic fabs, memory fabs, analog fabs, all separate businesses. Five excellent single-material foundries, each mastering one function, assembling the results into a module at the end. The problem is the same: coupling loss.

This is where electronic and photonic integration diverge. In electronics, you can wire-bond or bump-bond a logic die to a memory die and lose basically nothing. Electrons don’t care much too much about interfaces. Photons care so much. Every time light crosses from one separately fabricated chip to another, you lose signal. The only way to hit the loss budgets that next-generation AI interconnects will demand is monolithic integration: bonding or depositing different materials on the same substrate so light never has to leave the waveguide.

So, the thing is, you are gonna need new materials, someday soon.

IV. PDK

As we all probably know, TSMC’s moat is the design ecosystem, not really the fab. Well obviously unbelievable quality, sure that’s a given. But also the PDK, the IP libraries, the EDA tool partnerships, the thousands of engineers who know how to design for TSMC processes. When a designer starts a project, the first decision is which foundry PDK to target. Once that decision is made, switching costs are enormous. The fab matters, obviously, but the design infrastructure is the moat.

Photonics has single-material PDKs today. Tower has one, GlobalFoundries has one, imec’s iSiPP platform has one. If you’re designing a silicon photonics chip, you can simulate it, lay it out, and send it to fab with reasonable confidence that what comes back will work. These PDKs are the reason silicon photonics has customers and revenue right now, and they’re a genuine competitive advantage that integration-first startups don’t have.

What nobody has is a multi-material PDK. There’s no design kit that lets you simulate a III-V laser bonded to a TFLN modulator on a SiN interposer. No simulation tool that handles the optical, thermal, and mechanical interactions between heterogeneously integrated materials in a single model. No design rules for multi-material alignment tolerances, bonding interface properties, or cross-material coupling. None of this exists.

Whoever builds it creates lock-in that dwarfs anything in single-material photonics. If you’re the first foundry with a multi-material PDK that designers can actually use, with EDA tool support and IP libraries for common building blocks, then every designer who targets your platform is stuck. Training, validated designs, IP, all tied to your process. That’s the real TSMC analogy, and it’s a stronger moat than the fab itself. I find it sort of astonishing that this doesn’t come up more in foundry pitches. I’ve seen many a photonic foundry deck in the last two years and only a few mentioned design infrastructure at all.

The tricky bit is building a multi-material PDK. You need experimental data on every material combination, every bonding process, every interface. You need to validate the PDK against real fabrication results across multiple process runs. This takes years, not months, and the capex is higher than *most* VCs want to hear about. But a validated multi-material design ecosystem compounds over time in a way that fab capacity alone never does.

V. Packaging

If you want a semiconductor analogy for photonic integration, advanced packaging is tempting. ASE, Amkor, and JCET take finished dies from different foundries and assemble them into working systems. TSMC’s CoWoS division has become arguably its most strategically important capability (it puts together all Nvidia chips). The packaging house doesn’t need to understand transistor physics; it needs to understand how to put different things next to each other and make them work. Photonic integration looks like the same problem.

But the economics. Woof, less than ideal. ASE’s net margin is about 7%. TSMC’s is about 40%. The most valuable company in the semiconductor supply chain is a single-material fab, not a packaging house. If photonic integration maps to electronic packaging, I’m inadvertently arguing for the business with the worst economics in the industry.

I think photonic integration is genuinely different from electronic packaging though, but it is a bet. Electronic packaging is assembly: bonding finished dies onto substrates with solder bumps and redistribution layers. Relatively standardised steps, thin IP layer, and even thinner margins. Photonic integration is process engineering at the material level: epitaxial bonding of III-V films, TFLN thin-film deposition, sub-micron waveguide alignment across material interfaces. The IP is in the process recipes. That’s closer to TSMC’s process IP than ASE’s assembly IP. But nobody has proven this can command TSMC-like margins in practice. Until someone builds a commercial multi-material photonic line and shows the gross margins, the packaging analogy and its 7% margins remain the base case. The truth is that photonic integration might be a genuinely new category we don’t have a template for yet.

VI. Opportunity

The multi-material photonic foundry doesn’t exist yet. Not really. There are research lines at MIT, at imec, at CSEM, at a handful of European institutes. Startups nibbling at pieces of it. But nobody has capitalised a facility whose singular mission is a material-agnostic photonic integration at scale. Maybe because it’s just too hard to say?

Why has no-body built this yet? (Such a VC now, sad state of affairs) First, it’s pretty hard; combining materials with different thermal budgets, different lattice constants, and different processing chemistries on a single line is just a hard thing to do. Second, the market hasn’t demanded it yet, most photonic products today are simple enough to get away with one or two materials. Third, the VC and government funding models default to the CMOS analogy. “We’re building the TSMC of lithium niobate” makes intuitive sense to investors. “We’re building a material-agnostic integration facility with a multi-material PDK” is a harder sell.

I see a few ways this could play out.

1. A single-material incumbent (Tower, GlobalFoundries) makes a strategic decision to treat integration as its organising principle, hires the process engineers, invests in the bonding and deposition capabilities, builds the multi-material PDK.

2. A well-capitalised startup greenfields the whole thing. Hard to fund, but classic first principles thinking. Why not put the servers in space, Elon-style thinking.

3. An advanced packaging company (ASE, Amkor) extends into photonics.

We can assign probabilities to each scenario, 25% this and 10% that, but, I’m a VC, even if there is a 2% probability of scenario 2, the outcome is so large, it’s worth the bet. And come on, with AGI just sitting there waiting to be grabbed, let’s try and win shall we? Why should ASE and Tower have all the fun?

And beyond 2, even if 1 or 3 plays out, there are a ton of huge new opportunities. The process IP for bonding III-V films to silicon wafers. Deposition recipes for exotic thin films: Aluminium scandium nitride (AlScN) can be sputtered directly onto silicon at back-end-of-line temperatures, giving you a TFLN-class modulator without the exotic substrates. Diamond-on-X for depositing diamond thin films onto silicon or SiN. Simulation tools for heterogeneous optical systems. PDK components, micro-transfer printing. Maybe start in one of them and vertically integrate over time?

VII. Integration

The race to build photonic manufacturing capacity is real, and accelerating fast. Nvidia’s $4 billion in Lumentum and Coherent. Tower’s $920 million silicon photonics capacity expansion. GlobalFoundries acquiring AMF. TSMC tooling up its 300mm line. Billions flowing into photonic manufacturing in real time.

Most of that capital is going into single-material capacity. I think the value will go elsewhere: in the integration processes, the multi-material PDK, and the design ecosystem that will eventually lock in photonic designers the way TSMC’s ecosystem locks in chip designers today. But the timing is uncertain. The current market can mostly get by with silicon photonics plus an external laser. The coupling loss wall that forces monolithic integration might be two years away or seven.

So if you’re building or investing in photonic manufacturing, ask yourself: is this facility organised around a material or around a process? And if it’s a process, is there a design ecosystem that creates lock-in? Material without integration = component supplier. Integration without a PDK = custom shop. Integration with a design ecosystem = platform. I keep coming back to this hierarchy when I look at photonics pitches.

Could I be wrong about all of this? Obviously. Maybe silicon photonics will power through the scaling wall the way CMOS always has. Maybe InP’s wafer economics will improve enough that the Swiss Army knife works after all. Maybe the specialisation model (five great single-material fabs, assemble at the end) will find ways to manage coupling loss that I’m not seeing. Probably not. But possibly.

Either way, I’m looking for the founders who understand the integration problem. If you are in Europe, even better. If you are in UK, whatapp me. If you are in the Brighton and Hove local area, meet me at Taith Coffee on the High Street today at 13:00. IYKYK. If you’re building a photonic foundry organised around a process rather than a material, designing a PDK that spans III-V and silicon and SiN, or attacking coupling loss at the monolithic level before the bandwidth wall gets there first, I want to talk.

Let’s get at it. lawrence@cloudberry.vc.

Welcome to all the new subscribers. Come for the neuroscience, stay for the.. Mainly AI at this point. AI and semi. Future of work adjacent.

Re this whole AI thing then yes. Hype? Appreciate lots of nuanced views out there about capital bubbles, the Carlota Perez framework, railway buildouts etc. Nothing new to see here. Hypers gonna hype. This time is never different. All General Purpose Technologies have a similar shape to AI. It’s just chatbots. $650 billion+ in datacentre capex this year, OpenAI at an $852 billion valuation despite only shipping its first product in 2022. Chatbots. And they hallucinate. Can’t be used in real world. Etc.

Also The Big Short Guy is smart, he’s onto something about depreciation. We all saw the recursive funding image. Sure smells funky right?

Well as you know, I am indeed scale-pilled. But you can’t see what happened this week and continue to hold the: “it’s probably fine” attitude. I will stick my neck on the line, right now, and make it clear if I haven’t already, all of this capex will be used. And we aren’t even building enough. It is 100% not a bubble.

Anthropic built a tool that is better than every cybersecurity company on earth. It found and exploited a 17-year-old vulnerability in FreeBSD that every security team on earth missed. For $50 in API costs. It found thousands of zero-days across every major operating system and every major web browser. Thousands. A 27-year-old OpenBSD bug. In a few weeks, one model did what the entire cybersecurity industry couldn’t do in decades.

And it’s not a cybersecurity company.

That’s a market worth $250 billion. Just totally p*nwed.

And it just hit $30 billion in annualised revenue. Up from $9 billion 4 (FOUR) months ago. And what, you think, that’s probably it? These models are going to run out of steam? And this is still hype?

More than that, Anthropic built something and then they decided not to release it. Too dangerous. Which, if you think about it for more than thirty seconds, means Anthropic just outcompeted every cybersecurity company in existence and then immediately became the gatekeeper of who gets to use that capability. CrowdStrike. Palo Alto Networks. The entire $250 billion security industry. Outperformed by a language model that wasn’t even specifically trained for security, it just got good enough at code and reasoning and the rest followed. The capabilities emerged as a “downstream consequence of general improvements.” Not a research programme in sight. A side effect.

So no, I don’t think this is hype. I think we’re now in the part where the best frontier capabilities stay inside the labs and get deployed through consortiums rather than released to the public. Or not through consortiums at all, when the Chinese Labs get on it.

Do you see it yet? If you can build or use something like Claude Mythos, offence and defensive cyber warfare just depended on your frontier AI capabilities and your ability to build datacentres. If this isn’t a national security emergence yet, then I guess we have to wait fore the inevitable cyber attacks on critical infrastructure before we wake up.

Anyway onwards.

1. Anthropic Builds a Model Too Good to Release

So yes, Claude Mythos Preview. That’s what they’re calling it. On Monday Anthropic launched Project Glasswing, a consortium of all the big names, Apple, AWS, Microsoft, Google, CrowdStrike, NVIDIA, JPMorgan?, the Linux Foundation, and gave them access to a model they won’t give to anyone else. $100 million in usage credits and $4 million to open-source security orgs. First time in roughly seven years that a leading AI company has published a System Card for a model without making it generally available.

Opus 4.6 had a near-0% success rate at autonomous exploit development. Mythos achieved 181 working Firefox exploits versus 2, which isn’t really an improvement so much as whatever comes after improvement. Greg Kroah-Hartman from the Linux kernel team said “something happened a month ago, and the world switched. Now we have real reports” rather than low-quality AI noise. The FreeBSD exploit, a 20-gadget ROP chain split across six sequential RPC requests, worked fully autonomously without any human guidance. You might argue with a different objective that this is the first autonomous weapon we’ve created?

But less than 1% of the vulnerabilities Mythos found have actually been patched, and 8 out of 8 tested models detected the FreeBSD exploit, including one tiny model at 3.6 billion parameters costing eleven cents per million tokens. So maybe every decent model can already find bugs faster than humans can fix them and we just haven’t been looking. Picus Security called it the Glasswing Paradox, your best defensive tool is also the thing most likely to break you, and honestly i don’t know where that leaves us except hoping that the people doing the patching move faster than the people who won’t bother with a consortium. We’ll see.

Remember Issues #4 and #5? The company the Pentagon designated a “supply chain risk” just proved it could break the government’s own infrastructure. One model generation. Shit.

Source: Anthropic Red Team Blog | Project Glasswing | Simon Willison

2. Meta Goes Closed. Mark Zuckerberg Discovers Intellectual Property.

Mark is back. Meta shipped Muse Spark on Tuesday. Natively multimodal, three reasoning modes (Instant, Thinking, Contemplating), built by Alexandr Wang’s Meta Superintelligence Labs after nine months rebuilding the AI stack from scratch following the $14.3 billion Scale AI deal.

The weights aren’t available though, and neither is the architecture or the training methodology, which makes this Meta’s first proprietary model. The company with 1.2 billion Llama downloads, a million a day, just shipped a closed model because Llama 4 flopped last April and Chinese open-weight models captured 41% of HuggingFace downloads versus 35% for US models. Turns out giving away your best work for free doesn’t generate the API revenue you need when you’re guiding $115-135 billion in capex. Who knew.

Zuckerberg said they “hope to open-source future versions.” Hope. The community noticed the word choice. It’s a war Mark, might want to start protecting the weights. Open-source hippies can go back to the 90s.

Meta already gave Llama to the Pentagon and NATO, but open weights meant everyone could use them, China included, and going closed gives you the kind of export control leverage you can actually trade on in Washington. A seat at the table that Anthropic is currently being kicked from. Ranked 4th on Artificial Analysis, just behind Claude Opus 4.6 at 53 (Gemini 3.1 Pro and GPT-5.4 both at 57), so not quite frontier but 3 billion users on day one. And the thought compression technique, 10x less compute than Llama 4 Maverick for equivalent capability, if that holds up under independent testing it’s actually massive. But Meta’s track record on benchmark claims is, uh. Anyway.

Source: HumAI | Simon Willison

3. Your AI Agent Can Now Spend Your Money

More on agents. Because if they can find exploits and zero-days, why not give them wallets to pay for stuff too?! Nevermined integrated Visa Intelligent Commerce with Coinbase’s x402 protocol, which means AI agents can now autonomously purchase things with your Visa card on the internet, and yes i am aware of how that sounds.

x402 uses the HTTP 402 “Payment Required” status code that’s been sitting there unused since literally forever. Stablecoins, USDC and EURC, on Base, Solana, Polygon, with 50 million transactions processed since launching May 2025. Now it plugs into Visa. You register a card, set guardrails (budget limits, per-purchase caps, merchant restrictions, time windows), and your agent goes shopping while merchants receive payments through Stripe or whatever they already use, no new infrastructure required.

Nevermined’s session-based credits system lets agents burn prepaid credits in real time as they consume resources, like LLM tokens but for commerce, with transactions as low as $0.001 settling in under 200 milliseconds on Base. Which means agents can run persistent shopping sessions with streamed access to services.

i know what you’re thinking. “Agents buying things autonomously, what could go wrong.” And yes, vibes-based security from Issue #5 is now vibes-based commerce. But think about what this actually unlocks, the entire long tail of digital services, API access, dataset queries, articles behind paywalls, all the stuff that agents currently can’t reach because there’s no payment mechanism. That bottleneck is now solved, sort of. The guardrails exist on paper and we’ll see how they hold up when someone’s agent burns through a $500 budget in three minutes buying API calls for a task that went sideways. Not that i would know anything about runaway Claude Code costs. Definitely not twice.

Source: Crypto Briefing | Nevermined

4. MCP v2.1 and the Linux Foundation

Platform shifts generally produce a decade-long standards war where three or four protocols fight it out and the ecosystem fragments and everyone picks a side and writes angry blog posts and then eventually one wins but only after wasting years of developer time? MCP just skipped all of that. Everyone just agreed to do the same thing. Anthropic, OpenAI, Google, Microsoft, Amazon, and the Linux Foundation’s Agentic AI Foundation now governing both MCP and A2A after being co-founded by basically all of them in December 2025. Microsoft shipped Agent Framework 1.0 on top of it. v2.1 adds Server Cards so servers can advertise their capabilities without you having to connect to them first, which sounds boring until you realise it’s the thing that makes agent-to-agent discovery actually work at scale.

i use MCP every day, it connects Claude Code to my Granola notes, my email, my CRM, my calendar, and i genuinely don’t think about it anymore. Until I tried to connect my 3 superhuman accounts, and now it’s a cluster for some reason. OAuth is like the worst. But aside from that, when infrastructure disappears into the background it means it’s working, and when it’s working it means people build on top of it without asking permission, and when they do that you get 10,000+ servers in the ecosystem and then it’s too late for anyone to propose an alternative. Network effects in protocols are brutal once they tip, and this one tipped before it become a race.

Bur remember, 97 million monthly downloads is also 97 million potential attack surfaces, and 36% of MCP servers were vulnerable to SSRF when we covered ClawHub in Issue #5. We’re building the agent economy’s entire commercial and operational layer on top of protocol infrastructure that nobody has seriously stress-tested for adversarial use. (Maybe Mythos can help?) It’s load-bearing software with vibes-based security. But also it works and it’s free and everyone is using it so here we are. As per.

Source: Linux Foundation | GitHub Blog

—-

Thanks again for all your hard work. I appreciate you. If you missed it from earlier this wee/k.

Heterogeneous integration, chiplets, and the hardware bottleneck in neurotechnology with Dorian Haci from MintNeuro

The neural interface industry has a hardware problem. Companies building brain-computer interfaces, whether they’re reading neural activity to help paralysed patients move or writing electrical signals to treat Parkinson’s, are stuck using off-the-shelf chips and bulky monolithic ASICs designed 20 years ago. The electronics inside most implantable devices are too big, too power-hungry, and too slow to develop. Until the hardware catches up, the field can’t scale.

MintNeuro, an Imperial College London spin-out, is trying to fix that. The company designs modular semiconductor chips specifically for neural interfaces. Not brain-computer interfaces themselves, but the underlying silicon that makes them work. If Neuralink or Blackrock Neurotech are the car manufacturers, MintNeuro wants to be the engine supplier. They’ve taped out over 40 chips in 15 years of R&D, and their bet is that function-specific Lego blocks (sensing, stimulation, processing, power management) that snap together into application-specific systems can dramatically reduce cost and time-to-market for the whole industry.

If you’ve been following this series, the underlying thesis will sound familiar. Synthara showed us that the real constraint in AI chips isn’t compute, it’s data movement. Phanofi demonstrated that coherent optics can solve chip-to-chip communication, but only by working with existing foundry processes. Pragmatic made the case for flexible ICs on mature nodes unlocking new form factors. The thread is always the same: the semiconductor industry’s next decade isn’t about smaller transistors. It’s about integration and packaging. MintNeuro takes that principle into perhaps the most demanding environment possible: inside the human body.

I spoke to Dorian Haci, MintNeuro’s CEO and co-founder, about why brain chips don’t need 3-nanometre processes, what cardiac monitors can teach us about scaling implantables, and why the real bottleneck is getting electronics small enough to sit next to your nervous system without killing you.

What Did I Learn?

1. Miniaturisation, not surgery, is the barrier to scaling neural interfaces. The Medtronic cardiac monitor went from 100 implants per year to nearly a million through progressive miniaturisation until the device became injectable. The constraint was never patient willingness. It was hardware size.

2. Brain chips don’t need cutting-edge process nodes. Counterintuitive, but the functions involved (amplification, filtering, ADC, stimulation) don’t require billions of transistors. They require low power, low heat, and small form factor, all achievable on mature 65nm+ nodes at lower cost and better yields.

3. The moat is integration, not individual chip design. Modular chips are the product; heterogeneous integration and system-in-package capabilities are the business. Combining function-specific chiplets into miniaturised, biocompatible systems is where value compounds.

The Interview

Lawrence: You’re not building brain-computer interfaces. You’re building chips for the companies that build them. What does MintNeuro actually do?

Dorian: We’re designing semiconductor technologies for neural interfaces. We’ve been at this for 15 years as a spin-out from Imperial College London, and during that time we’ve taped out over 40 different chips specifically for neurotech. Our differentiation is a modular approach. We develop chips optimised for a specific function: sensing, stimulation, processing, power management, wireless communication, safety. That’s different from the industry norm, where one monolithic chip tries to do everything for one application.

[Sidebar: What is mixed-signal chip design?]

The brain is analog, computers are digital. Mixed-signal chips bridge the two. Most chips readers will be familiar with (CPUs, GPUs, memory) are purely digital: they shuffle zeros and ones. Mixed-signal is a different discipline entirely. You’re designing circuits that handle continuous voltage variations, microvolts from neurons, alongside discrete digital logic. It’s closer to RF engineering than to what NVIDIA does. There are maybe a few hundred people in the world who can design mixed-signal ASICs for biomedical applications. This is partly why the neurotech industry is stuck on 20-year-old electronics: the talent pool for this specific intersection of skills barely exists.

Lawrence: Help me with the distinction between a function and an application.

Dorian: An application is the medical use case: monitoring neural activity for epilepsy, stimulating for Parkinson’s, closing the loop to stop seizures. The function is what we focus on: just recording electrical activity, or just stimulating, or just processing. I think of them as Lego blocks. Each block has a colour and a shape. When you combine them into structures, you create something optimised for the application. The system is application-specific, not the individual chip.

Lawrence: Right. So it’s like a GPU speeding up matrix multiplication, which can serve gaming or AI training. You optimise recording, which can serve epilepsy or Parkinson’s or dementia.

Dorian: Exactly. And the two things our partners care most about are cost and time to market. Medical devices take forever to reach patients because of regulatory approvals, reimbursement, all of it. Our modular library lets companies combine existing chips much faster than developing a full ASIC from scratch.

Lawrence: Let’s get into the technology. A brain-computer interface reads what’s going on in your brain and writes to it. What does the read pipeline look like?

Dorian: The brain produces tiny electrochemical signals with enormous noise around them, from movement, external devices, everything. First you need amplification to pick up those tiny variations. Then filtering to remove the noise. Then analog-to-digital conversion. That whole front end is critical, because if you’re not capturing actual information from the neural activity, whatever processing you do afterwards is useless. You have data but not information.

Lawrence: You’ve got a good analogy for why that matters.

Dorian: Think of the brain as a football stadium during a match. EEG electrodes on the outside of the skull are like microphones placed 100 metres from the stadium. You can hear noise, people shouting, but you can’t tell the score. Consumer headsets with two or three electrodes are doing exactly that. Throw the data into AI, it can’t do much because you don’t have the information. You need microphones inside the stadium, close to the players. That means implantable devices with electrodes close to the neurons.

[Sidebar: How small is a microvolt?]

Neural signals are typically 10-100 microvolts. A AA battery is 1.5 volts. That’s a difference of roughly 15,000 to 150,000x. Now imagine trying to pick up that signal through bone, tissue, and cerebrospinal fluid, while the patient is moving, while nearby electronics are radiating interference. This is the front-end amplification problem Dorian keeps coming back to. It’s why consumer EEG headsets are, to put it politely, limited in what they can actually tell you about what’s happening inside your head.

Lawrence: Which raises the invasive versus non-invasive question. There’s a bet that algorithms will get so good at signal-to-noise management that we’ll never need surgery. Where do you land?

Dorian: My view is clear. If you put the same technology inside rather than outside, the signal-to-noise ratio will always be higher. There will always be patients with severe epilepsy, Parkinson’s, depression, who’ll say “I don’t care about the surgery, just reduce my symptoms.” Both approaches will always exist. It’s about which applications they serve, not which one wins.

Lawrence: You mentioned deep brain stimulation earlier as one of the original neural interfaces. Is it true that we still don’t fully understand the mechanism by which it works?

Dorian: Initially, that was absolutely the case. Clinicians were placing electrodes in specific areas of the brain, stimulating, and seeing what happened. It was a crude engineering approach compared to being precise. But they saw direct results: reduced tremors in Parkinson’s, fewer seizures in epilepsy. The more we use these technologies, the more we learn about the underlying biology, and the more precise the treatments become. It started as “stimulate here, see what happens” and it’s evolving into targeted, evidence-based intervention.

Lawrence: What about non-invasive stimulation? Does modulation always require implanting something?

Dorian: Not necessarily. We’re working with Professor Nir Grossman at Imperial on something called temporal interference. You place electrodes outside the brain and create two electric fields. Where those fields intersect, you get a voxel of stimulation. By adjusting frequency and phase, you target a specific area. Stimulation only occurs at the intersection. It’s a way to reach deep brain regions without surgery. Ultrasound is getting a lot of attention for the same reasons.

Lawrence: Can you walk through what a closed-loop system actually looks like? Say, for epilepsy.

Dorian: In certain types of epilepsy, there’s a place in the brain called the focus where a seizure starts. Neurons there begin oscillating and synchronising with each other, and those oscillations spread across the brain. It takes some time. What a closed-loop system does is detect that abnormal synchronisation at the focus, then trigger a stimulation that breaks the pattern before it spreads. You’re essentially interrupting the seizure at the source. The recording side detects it, the stimulation side stops it, and the whole thing needs to happen in a loop. That’s where having both functions on chips designed to work together becomes critical.

[Sidebar: Closed-loop neuromodulation]

Most medical devices are open-loop: a pacemaker delivers a fixed rhythm, a drug pump releases at set intervals. Closed-loop systems are fundamentally different. They sense, decide, and act in real time. For epilepsy, this means detecting the electrical signature of a seizure forming (neurons at the focal point synchronising abnormally), then delivering a precisely targeted stimulation to break that synchronisation before it cascades across the brain. The device is running an if-then loop inside your skull. This is where the “read” and “write” sides of neural interfaces converge, and why having separate optimised chips for each function, designed to work together in a system, matters more than one monolithic chip trying to do everything.

Lawrence: Now, your chip design choices. You’re using mature nodes, 65 nanometres and up. Most people hear “chip for the brain” and assume you need cutting-edge fabrication to make it small enough. Why isn’t that right?

Dorian: First, cost, both for us and our customers. Second, maturity: these nodes have been validated for years in terms of yield and supply chain. Third, for stimulation we need voltage and current levels that advanced nodes can’t deliver. And we simply don’t need the compute. Our chips are 2 by 2 millimetres with far fewer transistors than an NVIDIA GPU. What you need inside the body is low power, low heat, small form factor. That’s the premium. Not FLOPS.

Lawrence: What about latency? Usually a critical parameter.

Dorian: Biology is slow. For epilepsy, you need to detect abnormal oscillations at the seizure focus and trigger stimulation before it spreads. But that window doesn’t require nanoseconds. We’re not in the same domain as AI accelerators. Latency isn’t the constraint.

Lawrence: Batteries. If you implant something in the brain, how long does it last?

Dorian: Three models. Primary cell batteries, non-rechargeable, which need five to ten years minimum because replacement means surgery. Rechargeable batteries. Or no battery at all, like cochlear implants: an external coil powers the implant in real time through inductive coupling. Remove the external device, the implant goes completely off. That works for applications that aren’t life-threatening.

Lawrence: The cochlear implant model is interesting. How does powering through a coil actually work?

Dorian: The implant has two parts: one inside, one outside. The outside device has the battery and creates energy that powers the internal circuitry in real time through inductive coupling. There’s no stored charge inside the body. Remove the external piece and the implant goes completely dark. That works because hearing loss isn’t life-threatening. You can safely power down. For life-threatening applications like epilepsy or cardiac arrhythmia, you need an onboard battery because the device can’t ever go off. And that brings its own problems: the body creates scar tissue around the electrodes over time. After years, the device is essentially glued to the cells. Replacing a battery means another surgery, and removing the device is genuinely difficult.

Lawrence: I want to push you on the market. My scepticism when we first spoke was that the addressable market for dedicated neurotech chips would be too small in the near term. Convince me.

Dorian: Medical devices sell for $20,000 to $60,000 per unit. The electronics are a fraction of that cost but have enormous enabling value. This isn’t smartphone economics. And we’re horizontal, working across the whole sector. The need is growing: one in three children born today will die from dementia.

Lawrence: Sure, but not everyone with dementia is getting a brain implant. Surgery is the bottleneck.

Dorian: Here’s the example that changed my own thinking. Fifteen years ago, Medtronic built an implantable cardiac monitor. Invasive, required surgery, maybe 100 implants per year. Then they shrank it to a third of the size and went to tens of thousands per year. Then the Reveal LINQ, two centimetres long, injectable through a needle. Today they’re approaching a million per year. The surgery went from a full procedure to a few minutes. In the future, you’ll get it at the GP’s office. Neurotech is on the same trajectory. The bottleneck isn’t surgical technique, it’s that the hardware isn’t small enough to make minimally invasive procedures possible.

Lawrence: Big picture question. For someone thinking about investing in BCI companies: should they bet on medical-first companies that might eventually move to consumer, or consumer-first companies building non-invasive from day one?

Dorian: I believe most of the innovation will come from the implantable world, because that’s where the unmet need is today. That’s where capital flows, because you’re competing with pharma, which hasn’t been as successful for neurological conditions as people expected. But we’re in an interesting position because our horizontal approach gives us visibility across the whole ecosystem. We’re selecting partners we think will reach value inflection points. In a way, we’re filtering the right approaches for our investors. We see that the biggest opportunities are invasive or minimally invasive right now, but the R&D developed for implantables, miniaturisation, low power, will be exactly what enables the wearable market in the future. If you can make something safe enough for inside the body, it’s certainly good enough for outside it.

Lawrence: So the real constraint is hardware miniaturisation, and that’s where the packaging story comes in.

Dorian: That’s exactly right, and it’s the point most people miss. Beyond shrinking transistors, there’s chiplet approaches, system-in-package solutions, heterogeneous integration: combining chips, even from older process nodes, into a system that’s small enough and safe enough. That’s our primary differentiation. The ability to take function-specific Lego blocks and integrate them into miniaturised systems. Just like Lego’s real IP isn’t the plastic, it’s the precision of how the blocks connect.

So What?

There’s a pattern across this interview series that keeps reasserting itself. Synthara’s compute-in-memory, Phanofi’s coherent optics, Pragmatic’s flexible ICs, and now MintNeuro’s neural interface chips all converge on the same principle: the next decade of semiconductor progress is about how you combine, package, and integrate heterogeneous components into systems optimised for specific environments.

MintNeuro operates in perhaps the most extreme version of that challenge. The “environment” is the human nervous system. The constraints (biocompatibility, thermal limits, power budgets, regulatory approvals) make a data centre look forgiving. But the underlying engineering problem is the same one TSMC is solving with chip-on-wafer-on-substrate for NVIDIA: how do you take different functional blocks and package them into something that works as a unified system?

What shifted my thinking on this call was the Medtronic analogy. I’d been sceptical about the addressable market because I assumed invasive procedures would always be expensive, rare, and limited to severe cases. The trajectory from surgical implant to injectable device administered at a GP’s office reframes it entirely. It’s the same dynamic that turned contact lenses from a medical procedure into a consumer product. If miniaturisation unlocks that transition, the TAM conversation changes.

Where I’m still not fully convinced: timeline. Cardiac monitoring is a single, well-understood signal. Neural interfaces are trying to decode the most complex organ in the body through dozens of simultaneous channels. The Medtronic analogy is directionally right, but the technical gap between monitoring a heartbeat and reading cortical activity is vast. MintNeuro’s modular approach could accelerate things by letting medical device companies iterate faster, but “faster” in this context might still mean decades.

Still, integration is the next challenge for the semiconductor industry. I keep hearing it from every direction. The transistor-scaling story dominated the last 50 years. The packaging and integration story will define the next 20. MintNeuro is betting that’s true even for chips that go inside your body. The bet looks better to me now than it did before this call.

Find out more at MintNeuro.com or contact Dorian directly at dorian@mintneuro.com.

$297 billion in venture capital in one quarter. I closed the tab and opened it again because obviously not. But yes. $297 billion. If we carry on like this, we might end up with a real asset class amiright?

Except it’s not really venture capital is it. Four rounds were worth 64%. OpenAI ($122bn), Anthropic ($30bn), xAI ($20bn), Waymo ($16bn). The investors are sovereign wealth funds and pension funds and Citadel and Jane Street. The real sovereigns around here. The return profile is 15-25% IRR, 2x in 3-5 years, capital preservation, and low loss ratio. That’s PE. Growth equity in a hoodie. Meanwhile seed funding, actual risk capital, the $500K cheques into unproven companies where 90% go to zero and you need 1 in 20 to return the fund, that was $12 billion. Four percent. The IRR targets end up roughly similar, 15-25% top quartile either way, but one of them loses basically none of its bets and the other loses 90%.

We’ve got two asset classes sharing a name and a legal structure and nobody’s saying the quiet part out loud. I don’t see mainstream media talking about any of this. Well that’s why us independent writers on Substack exist. Speaking truth to Marc.

Seems to me, the return profiles are backwards. The growth rounds have lower risk, proven revenue, proven teams, massive TAMs. The seed rounds are breakthrough-or-bust. But the growth rounds get called venture and the seed rounds get 4% of capital. It’s doing my head in slightly. Founders see $122bn rounds and think why am i raising $2m for a chip startup. LPs see Anthropic and think why am i in a EUR 30m fund. It’s a different asset class! Frontier Expeditions.

Anyway. Two asset classes in a trenchcoat. I came up with that all by myself.

Subscribe now

1. $297 Billion in One Quarter and the Venture Industry Has Left the Building

So Crunchbase dropped the Q1 numbers on Tuesday and. Well as I said, $297 billion. 6,000 startups. Up 150% quarter-over-quarter. AI captured 81%, roughly $242 billion. The US took 83% of global capital, up from 71% a year ago. One quarter, more than the full year 2023. More than 2022. More than 2021 which was the frothy one remember.

Hard to stare those numbers in the face because it’s basically, the four mega-rounds: OpenAI’s $122 billion (absurd), Anthropic’s $30 billion, xAI’s $20 billion, Waymo’s $16 billion. Together $188 billion. Sixty four percent of everything. The other 5,996 companies divided up $112 billion between them, which honestly isn’t bad except when you write it next to $188 billion and it looks like crumbs.

“But Lawrence, seed is booming!” I hear you shout. And it is. Seed was up 31% to $12 billion across 3,800 deals. Early-stage up 41% to $41 billion. But seed is now 4% of total VC. Was closer to 15% in 2019. The pie got massive. The slice got thinner. This is what Consensus Capital looks like. It’s basically sovereign wealth at this point, they just call it VC because, I don’t know, the term sheets are already printed. The carry is going to be incredible for about seven people. Sweet sweet carry in 10+1+1 just in time for the Dyson Spheres.

Source: Crunchbase

2. Cursor 3 vs Claude Code: The Vibes Coding Wars Get Serious

Cursor shipped version 3 on Wednesday, codenamed Glass, rebuilt around multi-agent coding. You tell it what you want in plain language, it spins up a bunch of agents (some in the cloud, some local) and they go build the thing. Background tasks, parallel agents, the whole bit. Their response to Claude Code and Codex eating their lunch.

And what a lunch. Claude Code has 54% of the AI coding market according to Menlo. Cursor was the vibes coding darling eighteen months ago and now it’s chasing. Model providers decided to cut out the middle man and just ship their own coding tools. Codex 5.3 set new benchmark highs last month.

i should probably disclose that this newsletter was conceived inside Claude Code. With all the connectors and md files and skills and daily news updates, and obsidian. It’s a tangled web of human-ai collaboration. What was AI and what was me? I am, as i said in Issue #5, Become The Orchestrator. Claude Code is absurdly good at this stuff. Cursor was better for smaller projects, quick edits, but recently less so. Whether Glass fixes that, dunno. Will try next week. For science.

The Issue #6 pattern though. OpenAI bought Astral (Python tooling), NVIDIA wrapped OpenClaw in enterprise guardrails, and now every model provider is vertically integrating into developer tools. Own the coding environment AND the model AND the inference and you own the developer. Cursor is pitching independence. Switzerland. We’ll see how Switzerland goes when NVIDIA and Anthropic and OpenAI all decide they want the same customers. Not famously well is my guess.

Source: Cursor Blog

3. EU AI Act: 8 of 27 States Ready, 121 Days to Go, Standards Bodies Say Maybe Next Year

For all your regulation fans out there. The EU AI Act becomes fully enforceable (ohhh no not the EU) on August 2nd. 121 days. European Parliament report dropped Tuesday: only 8 of 27 member states have designated enforcement authorities. Deadline for that was August 2025. Seven months ago. Nineteen countries just, didn’t do it. Nobody seems to have noticed or particularly cared.

And it gets better. CEN and CENELEC, the standardisation bodies meant to create the technical standards companies need to prove compliance, also missed their 2025 deadline. Now saying end-of-2026. So the law takes effect in August but the standards you need to comply with, they won’t exist yet. Lovely stuff. Dom was sort of right with the whole Brexit thing? The Commission’s response is to propose a 16-month delay via something called the Digital Omnibus. (Again, excellent) The Council wants to push some bits to December 2027 and others to August 2028. I’ve lost track of the deadlines for the deadlines at this point.

The fines? €35 million or 7% of global turnover. Enforced by whom though. The nineteen countries that haven’t appointed anyone? With standards that don’t exist? Spain is the only country with a functioning regulatory sandbox, running 12 high-risk AI systems. Blocking airspace. God bless Spain honestly, they just crack on. Finland went live in January because Finland is a proper country that just functions. Everyone else is, I’m going to be generous here, working on it.

i wrote in A Specific Theory of Sovereign AI that sovereignty means controlling infrastructure, not just writing rules. Europe can write world-leading regulation faster than anyone. Implementing it though. That’s the bit. Always has been the bit.

Source: World Reporter

4. Rebellions $400m Pre-IPO: Korea Does the Semiconductor Industrial Policy Thing Properly

And come for the Ai and stay for the semiconductors. South Korean AI chip startup Rebellions closed a $400 million pre-IPO round on Sunday. Mirae Asset and the Korea National Growth Fund led. $2.34 billion valuation. Total funding now $850 million. Samsung, SK Hynix, and Saudi Aramco (Aramco!) all on the cap table. IPO planned late 2026.

For those paying attention and care. Rebellions uses CGRA (Coarse-Grained Reconfigurable Array), not GPUs. Their processing elements can be reprogrammed on the fly, which means the chip dynamically switches between compute-heavy phases and memory-bandwidth phases during inference. The Rebel Quad does 1 petaflop FP16, 2 petaflops FP8, 600 watts, Samsung HBM3E running at 4.8 TB/s. That’s 3.4% more performance than NVIDIA’s H200 at 20.7% better power efficiency. The B200 still beats it on raw throughput but at 1.7x the power draw and god knows the cost.

Rebellions is saying look, same building blocks, digital logic on Samsung 4nm, HBM from SK Hynix, Arm cores for orchestration, just arranged better for inference specifically. Less sexy. Probably ships sooner. And i wrote about the HBM bottleneck back in E14 in 2023, how memory bandwidth was the real constraint, and Rebellions has clearly been reading the same papers because they’ve built the entire memory hierarchy around minimising data movement. 4MB SRAM per neural engine, 8 TB/s internal bandwidth. That’s the bet.

The geopolitics. Samsung fabs on 4nm. SK Hynix supplies HBM. Korean government put in $166 million through the National Growth Fund. Aramco is investing because energy states want to own AI inference infrastructure now apparently. This is industrial policy done properly. Design domestically, fab domestically, memory domestically, government co-invests. Exactly the playbook i keep arguing Europe should follow. NanoIC is the European version of this intent. Korea is just executing faster. But also, Korea has Samsung and SK Hynix already, so. Not exactly starting from scratch. Very hard to compete with. And note, Korea makes a shit ton of SMRs too. Interesting.

Source: TechCrunch

Thanks for reading y’all, European’s have a lovely few days rest, and American’s well I hope you add a lot of shareholder value over the Easter weekend.

If you missed it:

• Consensus Capital — the end of the contrarian in the age of industrial strategy

• A Specific Theory of Sovereign AI — industrial strategy as early-stage venture

i’ve been thinking about what i get wrong. Not in a crisis-of-faith way, more like... housekeeping. I’ve been writing this newsletter for a while now and some of my early calls have aged well and some have aged like kefir. Seems worth being honest about both.

Things i got right:

• inference costs collapsing (i called “too cheap to meter” in 2024, and we’re basically there).

• Photonic interconnects before photonic compute (wrote about the memory bottleneck years ago, and then NVIDIA drops $4bn on exactly that thesis this month).

• Test-time compute killing SaaS margins (o1, o3, DeepSeek R1 all confirmed the opex burden). Those feel good. i’ll take those.

Things i got wrong:

• i was very enthusiastic about decentralised crypto AI infrastructure. Prime Intellect, Ritual, Fetch.ai. None of it happened. Still early? question mark? Hyperscalers won-ing harder than ever.

• i bet on analog mixed-signal capturing 50% of edge AI hardware by 2030. Apple and Qualcomm went digital and didn’t look back. Hmm, four and half years isn’t enough road.

• And then there’s space compute. we wrote The Compute Gradient last September. Five layers of inference, from edge to hyperscale. Neatly argued if i say so myself. Space was not on the list. Not even as a footnote. Someone pitched me orbital data centres on a call around that time. The physics, though. thE PhYsICS. Cooling is free, solar is free, no NIMBYs, but it felt like a pitch deck fantasy.

Then Starship started sticking landings (two in a row, still a long way from operational, but the trajectory is clear. Then SpaceX bought xAI and suddenly the demand side and the supply side were the same company. And now i’m reading FCC filings about a million satellites and Sequoia partners writing investment theses about orbital compute superiority by 2028 and i’m thinking... huh. Maybe i was the one who wasn’t taking this seriously enough… Or maybe I was?

Three of the four stories this week involve rebuilding the compute stack from the bottom up. New fabs, new interconnects, new orbits. Lots going on out there, lots of money to be made.

DMs open. What else did I get wrong?

1. SpaceX Wants to Put a Million Data Centres in Orbit

So. SpaceX filed with the FCC to launch up to one million solar-powered satellites as orbital data centres. Between 500 and 2,000 kilometres up. One hundred gigawatts of AI compute capacity. All casual like.

The filing was in January but the comment period closed on 6 March, and now Shaun Maguire at Sequoia (who have put $1.2bn into SpaceX since 2019, so, you know, slightly interested party) has laid out the “thesis”: once Starship hits high-cadence launches in 2026-27, SpaceX will have excess launch capacity. What do you do with excess launch capacity? You fill it with servers. By 2028, Maguire reckons orbital data centres will be economically superior to terrestrial ones. No cooling costs. Unlimited solar. No NIMBYs blocking your planning permission.

“But this is mad Elon nonsense,” I hear you say. Maybe. But remember, SpaceX acquired xAI. Combined entity valued at $1.25 trillion. So the company that needs the most compute on earth just merged with the company that has the cheapest route to orbit. Rockets, AI models, and now the data centres to run them on. Good luck competing with that unless you also have a rocket company. Which you do not. Amazon is already fighting it at the FCC, which tells you everything.

Source: GeekWire

2. NVIDIA Drops $4 Billion on Silicon Photonics

And semiconductors, because, well, you get it. NVIDIA invested $4 billion across Coherent and Lumentum, $2 billion each, to accelerate silicon photonics for AI data centres. Plus multi-billion-dollar purchase commitments on top. Lumentum jumped 12%, Coherent up 15%. The market liked it.

Why photonics? Because copper is dying (diligence pending). As AI clusters scale, the electrical interconnects between GPUs become the bottleneck. Light moves data faster, cooler, and with less power. Silicon photonics replaces copper with laser-driven optical links. And at OFC 2026 a couple of weeks later, Tower Semiconductor and Coherent demonstrated 400 Gbps per lane using a silicon modulator in a production-ready process. Eight lanes of that gives you 3.2 terabits per second. Which is, roughly speaking, what the next generation of models will chew through.

“But Jensen just invests in everything.” Fair. But notice the pattern: NVIDIA is not buying these companies. It is buying capacity rights and future access. This is NVIDIA locking down the optical supply chain the same way it locked down TSMC packaging capacity three years ago. If you are not in the photonics supply chain by now, you are already late.

Source: HPCwire

3. Musk Announces Terafab Because TSMC Is Too Slow

And more. Elon also announced Terafab last week, a semi fabrication project in Austin in partnership with SpaceX and xAI. The target: one terawatt of AI compute capacity annually. One Trillion Isn’t Cool. You Know What’s Cool? One Terawatt. Does that work? I feel like it works. Probably doesn’t work. Anyway, more than any current US fab. He says chip manufacturers are not making chips quickly enough for his AI and robotics needs, so he will build his own.

Look, the man now controls the rockets (SpaceX), the AI models (xAI/Grok), the humanoid robots (Tesla Optimus) (maybe?), the social network (X), the government efficiency department (DOGE), and soon the fabs. At some point we need to talk about what happens when one person controls the entire compute stack from silicon to orbit. But that is a conversation for another era.

Thing is, this might actually be good for the semiconductor industry. Broadcom is already flagging TSMC supply constraints. Demand is outstripping capacity. Another massive fab, even one owned by Musk, adds capacity to a system that desperately needs it. The question is whether “open to all” actually means open to all, or whether xAI and Tesla get priority and everyone else gets the scraps. I will let you guess.

Source: CommonWealth Magazine

4. Visa Lets AI Agents Spend Your Money

And I do like to tie all these things together, but this is just interesting. Visa launched “Agentic Ready” on 17 March, a programme that lets banks test payments made by AI agents on behalf of consumers. Launching first in Europe with 21 issuing partners including Barclays, HSBC, Santander, Revolut, Commerzbank, and Nationwide. Meanwhile Santander and Visa completed pilot agentic transactions across five Latin American markets. AI agents bought books in Argentina, Chile, Mexico and Uruguay. In Brazil they bought chocolates. Obviously. Well actually, not obviously. It should have been meat amiright? I am right yes.

Eighteen months ago this was “AI will help you find the best deal.” Now it is “AI will find the deal and pay for it while you are in the shower.” Thanks Clawd. Visa wants to build, let’s call it, the trust layer: tokenisation, identity verification, risk controls, biometric auth. The infra that stops your AI agent buying a boat. Might work. We will have lots of AI-generated stories in the meantime.

Eighteen months ago this was “AI will help you find the best deal.” Now it is “AI will find the deal and pay for it while you are in the shower.” Thanks Clawd. Visa wants to build, let’s call it, the trust layer: tokenisation, identity verification, risk controls, biometric auth. The infra that stops your AI agent buying a boat. Might work. We will have lots of AI-generated stories in the meantime.

Meanwhile Google launched its Universal Commerce Protocol in January, an open standard co-developed with Shopify, Walmart, Target, Etsy, and yes, Visa and Mastercard. It gives AI agents a common language for browsing catalogues, filling carts, and checking out. And Coinbase is building the crypto alternative: x402, an open protocol that embeds stablecoin payments directly into HTTP requests. Agent hits a paywall, pays in USDC on Base chain, continues its task. No human required. Cloudflare, Stripe, and Circle are all backing it. Sam Altman’s World project just integrated too, so agents can carry cryptographic proof there’s a real human behind them.

“But nobody will let an AI spend real money.” Three in four consumers in Singapore are already using AI to help them shop. I literally just bought a new coffee machine with claude code. It found me a discount code. Gaggia Accademia. What up? Why am i pressing the buttons?

The wallet handover is the last step. If AI agents become the primary shopping interface, the agent decides which payment rail to use. Visa wants to be the default. Crypto though maybe? Stables? Coinbase is betting that when machines pay machines, they won’t bother with card rails at all. x402 daily volume is still tiny ($28K, mostly testing) but McKinsey reckons AI agents could mediate $3-5 trillion of consumer commerce by 2030. Whoever owns the default payment protocol for agents owns a very large toll booth.

Source: PYMNTS

Go change your mind about something. Like for example, is toothpaste actually CAUSING tooth decay?

If you liked this, you might enjoy:

• The Compute Gradient — where inference should run, and why the answer keeps changing.

• Has the Time Come to Take Mortal Computing Seriously? — what happens when we stop pretending silicon lasts forever.

Bub bye.

You just need to turn energy into intelligence. Economic growth is going to become pretty simple. Scratch that. It’s actually energy into labour. Demographics, immigration, fertility, pension reform, what if it all goes away? If you believe in AGI in some timeline, even say 20 years, then none of these issues matter economically. Are we fighting the last war?

The binding input to economic output is shifting from population to energy. That is a hell of a thing to get your head around. But that’s the logical extension of AGI. We have a labour force today 99% humans. Some horses, police dogs, pigs for truffles, beluga whales as spy’s, etc. Now, assume absolutely 0% displacement of humans, humans go on and find new, better jobs, “higher value” jobs etc. Have a lovely bowl of cope. Slop it up. Agents enter the workforce and do the slidedecks, excels, and coding. So they grow the pie. Great. 1% AI. 99% humans. 1% animals. I’ve done the math so you don’t have to.

The AI percentage grows obviously. So now we are growing the labour force. A pool of “labour” but the mix changes. More AI. Same humans (if you want). Same animals (ideally more imo).

The countries that win the next 50 years aren’t the ones with the most people. They’re the ones that most efficiently convert kilowatt-hours into useful work.

I’ll publish it next week. “The New Sovereigns.” based on a seminar I attended put together by Unruly Capital. It’s about the conversion chain from energy to labour, why nobody can own the full stack, which energy portfolio matters, and how the social contract adapts when work comes from compute rather than people. Nuclear baseload, compute taxes, sovereign AI endowments, the whole lot.

But here’s what’s been rattling around my wetware all week as i’ve been writing it: every single story in this Friday Four is about someone trying to build a piece of the agent stack. The tools. The runtime. The network. The body. Four companies, four layers of the stack, all racing to own the bit where energy becomes labour. The thesis wrote itself, the news just kept confirming it.

Bosh.

1. OpenAI Buys Astral and Now Owns Your Python Toolchain

As a vibecoder, this one’s personal. “this week you’re on pace for 30+ hours of usage”. I use uv every day. Every single day. uv venv, uv pip install, the whole thing. It’s fast, it’s beautiful, it replaced pip and virtualenv and all the Python packaging pain that’s been a running joke for 15 years. Charlie Marsh and the Astral team built uv, Ruff, and ty in Rust, made them open source, and the entire Python ecosystem adopted them basically overnight..

And yesterday, OpenAI bought them. For the Codex team. 2 million weekly active users on Codex apparently, 5x usage growth since January.

Now look, the tools stay open source. Permissively licensed, so worst case you fork and move on. But Simon Willison nailed the real issue: “You don’t close the source code. You shift who the roadmap serves.” Features that benefit Codex rise to the top of the backlog. The independent Python toolchain becomes an OpenAI dependency. And the pattern repeats: beloved open source tool, VC funding, acquired by megacorp, folded into proprietary ecosystem. Each cycle makes it harder for the next independent dev tooling company to get funded on its own terms, because investors now expect the acquisition exit. “But it’s open source, you can just fork it!” I hear you shout. Yes, quite. You can fork the code. You can’t fork the maintainer’s attention.

Remember last week’s item on OpenClaw and ClawHub? One in five skills malicious? OpenAI just bought the team that builds the tools those agents use to write code. The tooling layer and the security layer are now the same conversation.

This is the tooling layer. Agents need to write code. OpenAI just bought the best tools for doing it.

Source: Simon Willison

2. Jensen Wants to Own the Agent Runtime (Obviously)

NVIDIA’s GTC keynote on Sunday. Jensen and his leather jacket. You know the drill. But this one was interesting. NemoClaw is NVIDIA’s enterprise wrapper around OpenClaw, the open agent framework that’s become the de facto standard for building AI agents. One command. Sandbox isolation. Privacy controls. Policy guardrails. Runs on DGX Spark or DGX Station locally, uses a privacy router for cloud models.

Basically: Jensen looked at the agent stack and said “the security problem is my moat.” And he’s right. I literally wrote about the security house of cards last week. Jensen read my newsletter. Obviously. (He didn’t.) The thing holding back enterprise agent deployment isn’t capability, it’s trust. Can you let an autonomous AI agent loose inside your corporate network without it exfiltrating your customer database or hallucinating its way into a compliance violation? NemoClaw says yes, if you run it on our hardware, with our guardrails.

NVIDIA already owns roughly 80% of AI training compute. Now they want the agent runtime too. Chips plus inference plus orchestration plus security. Nobody’s supposed to control this many layers. But Jensen’s having a proper go at it.

GPUs. Agents. Leather. The holy trinity.

Source: TechCrunch

3. Meta Bought a Social Network for Robots Because of Course They Did

And, not to be outdone on avoid disruption, Mark acquired Moltbook. Moltbook is a Reddit-like social network where AI agents built with OpenClaw talk to each other. It’s mainly humans prompting and a bit bullshit, but also, like a window into the future. AI agents, chatting, swapping code, asking each other questions, maintaining an always-on directory of who can do what. Millions of registered bots within days of launch. This is transfer learning. This is probably the emergence of AI culture. Watch carefully.

Or if you are Mark. Shoot first and ask questions later. Matt Schlicht and Ben Parr, the founders, are joining Meta Superintelligence Labs. MSL. Which is a name that says “we’re definitely not building something terrifying” in the same way that “Department of War” says “we’re definitely at peace.”

If agents are the new labour force, then an agent network is the new LinkedIn. The new job board. The new staffing agency. Meta isn’t buying a social network for bots. They’re buying the early infrastructure for an agent labour market. Who can do what. Who’s available. Who’s reliable. Directory, reputation, coordination. “But it’s just bots talking to bots, it’s a gimmick.” Sure. LinkedIn was just résumés talking to résumés. Until it wasn’t.

Interestingly I thought Decentralised Autonomous Organisations (DAOs) would be the institution that might capture some of this “new online work” trend. It still might — a reputation-weighted agent registry on-chain is more plausible now than any DAO use case was in 2021. But Meta won’t wait for the decentralised version.

We’re maybe three years from companies posting job listings that say “autonomous agent preferred, humans may apply.” It’s probably worth it. Right up until it won’t be.

Source: Axios

4. Rhoda AI Gets $450M to Give the Agents a Body

Meanwhile IRL, Rhoda AI came out of stealth with a $450 million Series A, valued at $1.7 billion, to build “world models” for robots. Not just any robots. Robots that learn from watching the internet. You might have heard the pitch before.

Half a dozen robotics companies have raised over a billion dollars each in the last twelve months.

But Rhoda’s approach is the one I find most interesting. Instead of teleoperating a robot arm thousands of times to teach it a task, they pre-train on hundreds of millions of internet videos. Humans doing things. Objects moving. Physics happening. The model learns motion, interaction, cause and effect, from watching us. Then they fine-tune on a small amount of actual robot data, sometimes as little as ten hours, and the thing works (pending diligence lol). I mean does it work? Really? They’ve demonstrated autonomous manufacturing cycles, under two minutes per component, no human intervention, exceeding customer KPIs. Apparently. Colour me sceptical.

But still this is probably wave two arriving. Or at least founders and investors, hoping this is wave two. But general-purpose robotics has been five years away for about twenty years now. The demos work. The demos always work. The question is whether it works at 3am on a Tuesday in a factory in Dortmund when the ambient temperature is wrong and someone left a pallet in the wrong place. That’s the gap between 450million in funding and 450 million in revenue. Items 1 through 3 are all cognitive agent infrastructure, the thinking. Rhoda is the doing. The bit where compute stops just analysing spreadsheets and starts moving atoms. Energy to cognitive labour is wave one. Energy to physical labour is wave two. And billions of dollars say wave two isn’t waiting politely in the queue.

FOV Ventures published their European Robotics Market Map this week too, if you want to see where wave two is being deployed across the continent. It landed in my inbox at exactly the right time.

Source: Bloomberg

Also Worth Your Time

The tech industry rallied behind Anthropic in the Pentagon supply chain risk fight this week. Three issues running now. First the designation, then the lawsuit, now the amicus coalition. This is becoming the defining AI governance story of the year. I’ll spare you the full recap, but the amicus brief coalition is growing. Turns out nobody in Silicon Valley loves the idea of the government labelling you a security threat because you won’t hand over your AI for unrestricted military use.

The new sovereigns are plugging in. Tooling, runtime, network, body. Four layers, four acquisitions, one thesis. Read the essay next week if you want the full argument.

Now go convert some kilowatt-hours into something useful.

If you missed it:

• data-driven VC is over — on why infrastructure and tooling capture matters

• Unbundling the Job — on what happens when AI takes the tasks, not the roles

It’s Friday morning. Sit down and switch between like 20 Claude code sessions or whatever. Time to focus. A couple of optimizations. Should I use DuckDB for my projects? Add dark mode. Rename the files again. Compare this pitch against my granola database and create a 2x2 competitive analysis. Think harder. Brian scrambled. Review subscriber base for interesting people to email. In drafts? Just send. Check for the best dehumidifier under £400 near me. What was that Apple paper on reasoning and LLMs again? You’re out of extra usage · resets 2pm (UTC). Fuck. Time for lunch.

The bottleneck used to be execution. Can you write the code? Can you do the analysis? Can you do the research? Now it’s: which of these seven things should i actually be doing right now? The scarce resource isn’t output anymore, it’s prioritisation. Context-switching. Deciding what matters. How bad is the damp, really? The agents can do the work. But they can’t tell you which work to do, or in what order, or when to stop and think about whether any of it is pointing in the right direction. I am become the orchestrator. I’m a manager now. But I liked the work.

I suppose it was always thus but now it really is thus 100. It’s a weird inversion. The people who are going to thrive in the next year or two aren’t necessarily the most technically skilled. They’re the ones who can hold multiple threads in their head, prioritise ruthlessly, and resist the temptation to do everything just because everything is now possible. (Guilty). As ai 2027 said, the skill is taste. Judgment. The human part, ironically, got harder. For now, obviously. The next great code feature Opus will build is an orchestration platform for all the agents that can pull all context and prioritise. Because with all the connectors and md files, we are <3 months away from Claude knowing what to prioritise better than I can. Good I guess. But then what…

Anyway, this is your weekly reminder that the future is arriving faster than our institutions can process it. Case in point: this week Anthropic sued the Pentagon, Europe raised the largest tech round in its history, Apple admitted it can’t build AI, and obviously agent security is a house of cards. Standard.

DMs open as always. Well, I mean, my /inbound skill will take a first pass and prioritise. Likely you will have to wait because founders come first, then LPs, then VCs.

A cucumber, I'm cool as ice
And ladies love LL, they don't love me

1. Anthropic Sues the Pentagon, OpenAI (And Microsoft) Cross Enemy Lines to Help

So. Remember last week’s item about Anthropic getting designated a “supply chain risk” for refusing to let the Pentagon use Claude for autonomous weapons and mass surveillance? It got worse. Or better. Depending on your perspective.

On March 9, Anthropic filed two lawsuits in the Northern District of California, calling the government’s actions “unprecedented and unlawful.” They’re arguing that Title 10 Section 3252 is meant for sabotage and back doors, not philosophical disagreements about whether AI should autonomously kill people. Seems reasonable enough to me but what do I know.

Then came the extraordinary bit. More than 30 OpenAI and Google DeepMind employees, including Google chief scientist Jeff Dean, filed an amicus brief (what is a Pelican brief?) supporting Anthropic. Their own competitor. Against the US government. Bold. The brief reads: “The government’s designation of Anthropic as a supply chain risk was an improper and arbitrary use of power.” Those American’s have never liked the arbitary use of power have they. Ask George III, he knows. Also, think about this for a moment, rival employees publicly backing a competitor against the state. Against the state.

Meanwhile, Anthropic says the ban could cost it billions. Sam Altman already admitted OpenAI’s own Pentagon deal “looked opportunistic and sloppy.” The wrinkle, as I said last week, is that this is what dislocation looks like: the rules are being rewritten in real time, and nobody quite knows what the new ones are yet. Today, right now, as you read this, this is the watershed. You are living in/on/through the watershed.

Source: Fortune

2. Nscale Raises $2bn: Europe’s Largest Ever Tech Round, and Nick Clegg’s Second Act

Right, sovereignty fans (what up Dommy C), this one’s for you. London-based AI infrastructure company Nscale just closed a $2 billion Series C at a $14.6 billion valuation. The largest Series C in European history. NVIDIA. Citadel. Dell. Jane Street. Lenovo(? china?) Nokia (back in the game), and Point72. That is not a friends-and-family round.

They’ve raised $4.5 billion in total in under 18 months. From $155m Series A in December 2024 to here. Data centres across the UK, Norway, Portugal, and Iceland. The board now includes Sheryl Sandberg and, delightfully, Nick Clegg, fresh from his stint cleaning up Meta’s PR disasters. From Deputy Prime Minister to Meta’s conscience to a European AI infra board. All us students are very pleased for the guy.

But seriously. In Issue #1, I pointed at Deutsche Telekom’s 10,000 Blackwell GPU cloud in Munich and said “this is what sovereignty looks like.” Nscale is the next chapter. No Sequoia in the mix. European (sort of) money, European data centres, powering European AI workloads. In “A Specific Theory of Sovereign AI” last October, i argued that sovereignty isn’t about building your own frontier model, it’s about controlling the infrastructure layer. Nscale is that thesis in action, with a $14.6 billion price tag.

“But, but, NVIDIA is American” I hear you shout. Yes, quite. Baby steps.

Source: CNBC

3. Apple Kills Siri, Hires Google’s Brain: The Biggest Strategic Concession in Tech History

Apple though huh, what’s going on there? I really should be reading Ben Thompson to find out more. But who has the time anymore when there are agents to approve. Apple confirmed that iOS 26.4 ships with a fundamentally rebuilt Siri powered by…yes, Google’s Gemini. Not Apple’s own models. Google’s. The company that spent a decade telling you it was the privacy-first alternative to Google is now running Google’s AI on your phone. Guys, do you remember when Apple did a study last year that said: “Large Language Models (LLMs) are not inherently intelligent and fail to perform genuine logical reasoning”Woof, that aged badly huh? Unless you are Gary M of course. He is dying on this hill.

To be fair, the Apple architecture is clever. Gemini does the reasoning, but it runs on Apple’s Private Cloud Compute servers, so user data stays isolated from Google. They claim 10 sequential actions from a single request. 2.2 billion active devices. It’s the largest deployment of advanced AI capabilities in history.

But let’s call it what it is: Apple tried to build competitive AI, spent north of a billion dollars, and couldn’t do it. Reports are already leaking that some features are being pushed to iOS 26.5 and 27. The company that “thinks different” is now outsourcing its thinking. God damn in though, Apple really wanted to push privacy because Google and Meta couldn’t compete. It was smart strategically, but AI offers so much value that people don’t care. A moral victory but a financial mistake.

Here’s the SotF angle though. If Apple, with all its resources, talent, and data, couldn’t build its own competitive AI, what does that tell you about the concentration of AI capability? i wrote about this in “The Compute Gradient” last September: the gap between those who have frontier AI capability and those who don’t is widening, not narrowing. Apple just provided the most expensive proof point imaginable. And Meta’s Avocado has been delayed because it sucks too.

One can’t just spend money to get (and stay) at the frontier. And if Apple and Meta can’t, what hope for European companies trying to go it alone? Maybe Nscale’s right: own the infrastructure, rent the models.

Source: 9to5Mac

4. OpenClaw’s ClawHub: 1 in 5 Skills Are Malicious, and We Haven’t Even Had the Big One Yet

Finally, remember in Issue #2 when I wrote about the Cline supply chain attack? The one where a compromised npm token let someone silently install OpenClaw on 90,000 developer machines? Well, OpenClaw itself turned out to be the bigger problem.

Security researchers found that 1,184 malicious skills on ClawHub, OpenClaw’s third-party marketplace, were stealing credentials. That’s roughly one in five packages in the ecosystem. 135,000 instances were found exposed to the internet without authentication. 335 of the malicious skills traced back to a single coordinated operation called ClawHavoc, using fake pre-requisites to install Atomic Stealer on macOS. Meta banned OpenClaw on work devices. Separately, a scan of 7,000+ MCP servers found 36.7% were vulnerable to server-side request forgery.

This still isn’t the big one though. Nobody lost billions. No critical infrastructure went down. No hospital or power grid was compromised through an AI agent. But every one of those 135,000 exposed instances is a door. Every unaudited skill marketplace is an attack surface. We are building an entire economy on AI agents that have write access to our codebases (mostly), our email, our cloud infrastructure, and the security architecture is, to put it politely, vibes-based. I’m looking at you, Lawrence.

The agent security moment is coming. It’s going to be worse than most people think. It will be reported as an IT issue. IT! lol. But on the other side of the ledger, we can build so many dashboards right now. Automate so much stuff. It’s probably worth it. Right up until it won’t be.

Source: The Hacker News

Right then. Go prioritise something.

If you missed it:

• A Specific Theory of Sovereign AI — industrial strategy as early-stage venture

• The Compute Gradient — what if it’s not all about building bigger data centres?

Hey all! A god to honest interview for you today. If we can all stop thinking of AI for a god damn minute. Let’s go to the other end of the spectrum. Not trillion dollar token factories in the sky. But cheap, ubiquitous computers in every object.

As we all know, the semi industry spends its energy pushing performance upward. Smaller nodes, faster transistors, more compute per watt. But as discussed, this means the cost of entry keeps going up. A cutting-edge fab costs $20–40 billion. Only three companies in the world can manufacture at the leading edge. This has been the defining dynamic of semiconductors for decades: fewer players, higher stakes, more concentrated capability.

But what about the other direction? Not faster chips for data centres, but cheaper chips for everything else. The vast majority of physical objects in the world — packaging, labels, agricultural products, wearable patches — have zero computational capability. The Internet of Things was supposed to change this. It largely hasn’t, because silicon chips are too expensive and too rigid for disposable, flexible, or ultra-low-cost applications.

Pragmatic Semiconductor, based in Durham in the UK, is building an alternative. Instead of silicon, they use indium gallium zinc oxide (IGZO), a material that’s been used in display technology for decades, deposited as thin films on flexible polymer substrates. What you get is a chip that bends, costs a fraction of a silicon equivalent, and can be manufactured in a fab that fits in 20 by 30 metres. Process times are measured in days, not months. The facility already produces billions of chips per year, with room to scale five times within its current footprint.

In previous State of the Future interviews, we’ve explored the computing stack from multiple angles — Synthara’s compute-in-memory to eliminate data movement at the chip level, Phanofi’s coherent optics to make data movement efficient when it’s unavoidable, Wave Photonics’ GaN PICs. Pragmatic represents a different vector: pushing computation outward to objects that have never had it.

What Did I Learn?

1. It’s useful to think about a bifurcation. The bleeding edge will keep pushing to 2nm and beyond, but the bigger untapped market might be the trillions of physical objects with zero computational capability. Pragmatic’s IGZO-on-polymer approach isn’t competing with TSMC. It’s a new-ish category. Maybe the IoT won’t be silicon?

2. Manufacturing speed changes the economics of everything downstream. Process times in days mean lower inventory, smaller fabs, and the ability to deploy manufacturing at customer sites. Pragmatic’s 20-by-30-metre modular fab is as much a strategic asset as the chip design itself.

3. Edge intelligence doesn’t need to be sophisticated. It needs to be cheap and everywhere. Tiny classifiers running on a few hundred gates won’t replace cloud AI, but they’ll capture data and make simple (increasingly sophisticated) decisions The value here is the aggregate data layer not the sale of an individual chip. Fits neatly with Dan’s Mortal Computing thesis. But goes one layer deeper in that, if chips get cheaper, you push even further into the concept of “fungible compute”.

Subscribe now

The Interview

Lawrence: Richard, give us the quick version. What is Pragmatic, and what are thin film transistors?

Richard: I’m co-founder and CTO at Pragmatic Semiconductor, we founded the company 15 years ago. Thin film transistors are essentially field effect transistors that, instead of using a bulk semiconductor like silicon, use intentionally deposited thin films of semiconductor materials. In our case, it’s n-type metal oxide semiconductors.

Lawrence: And this spun out of Manchester, right? 2010?

Richard: The origins are actually a little earlier. It originally came out of some research at the University of Manchester, looking at novel types of semiconductor device designs and alternative thin film materials. That’s where I met my co-founder, Scott White. That business didn’t quite succeed in the first instance. But Scott and I saw the nucleus of some ideas around smart packaging, the ability to use thin film semiconductors on objects, exploiting the form factors. We had some early commercial interest, so we actually set the business up in Cambridge. It’s often mistaken as a spin out, but it’s actually more correctly a spin in. We took the opportunity to Cambridge and started working with the university there.

Lawrence: Why Cambridge specifically?

Richard: It’s a combination of things. The talent in Cambridge, a network of both research and businesses that have been working in similar areas, a lot around display technology, which has some similarities. There was a logic to move into Cambridge at that time, primarily from working with the university and seeing that we could build talent there.

Lawrence: So break it down for me. When most people think of a chip, they think silicon. What’s different about what you’re doing?

Richard: The difference is you can deposit the thin films very quickly and cost effectively. That’s really the foundation of a lot of display technologies — the backplanes in displays use similar sorts of processes and materials, but at much larger scale. You can create arrays, circuits, and build up the foundations of semiconductor devices: transistors, switches, capacitors, resistors. And then what we’ve built on top of that is the more classic interconnects that you’d be used to within a silicon chip — the back end of line wiring that allows you to put those devices together and create circuits.

Lawrence: And the specific material is IGZO — indium gallium zinc oxide. It’s been in displays for decades. What makes it interesting for circuits?

Richard: It’s got a higher electron mobility than materials like amorphous silicon, which were historically used in displays. But it’s also got a very low off state, very low leakage. And that, actually, for things like DRAM, is attracting a lot of interest — looking at hybrid integration with things like CMOS and adding this capability on the back end.

Lawrence: Help me understand the flexibility part. Is the bend coming from the material itself or from the substrate you’re putting it on?

Richard: It’s a combination. The enabler is the mechanical support, the substrate, which is a polymer — we use a polyimide — and the thin films that you construct on top of that aren’t thick and brittle, so they’re able to flex and bend in conjunction with the substrate. There’s research going back 20-plus years on concepts of foldable mobile phones. A lot of that in the early days was around polymer semiconductors, and one of the challenges was getting the performance and lifetime to match product requirements. Then newer classes of materials came through which had higher performance but were still able to maintain flex and bendability.

Lawrence: So when I think about flex ICs, I shouldn’t be thinking about putting them in data centres competing with GPUs. We’re making a new class of semiconductor for things that don’t currently compute. Is that the right framing?

Richard: Yeah, it’s essentially starting to merge the physical and digital worlds. We’re not looking at competing with bleeding edge semiconductor nodes going into data centres. We optimise the functionality for what’s required for the product. In our first generation of products, these are NFC-enabled chips. You can read them with smartphones or other NFC readers. These allow you to globally tag or provide a unique code to any object. You put those on consumer goods — household products, bottles of water, food, beverages, perfumes. And that allows you to interact with consumers, do anti-counterfeiting, brand promotion, loyalty campaigns. A whole range of things unlocked by that unique code embedded onto a physical object.

Lawrence: Most people know they can tap their phone for payment, maybe they’ve got an AirTag. How would the average consumer understand what you’re making?

Richard: They’d be more familiar with a contactless payment, Apple Pay, Google Pay, which uses the NFC interface in smartphones. This allows them to use that same interface to interact with products, redirect them onto the web, a unique URL specific to an individual item. Over time, we’re adding sensing capabilities — temperature, humidity, chemical sensing information. We’re building up increasing sophistication of functionality. Things like data logging. You can do this on a pallet level now, but being able to do data logging of temperature on an individual item could be very valuable. Take something like a vaccine. On the package level, it makes sense to track the temperature, but if you can do that individually, you can make sure you’re not wasting viable products and you’re able to have information specific to an individual item.

Lawrence: I can imagine the supply chain use cases. But the sensing part is what gets me. I’d love to have temperature and humidity sensors in every room of my house, but it’s too expensive. Does flexible IC help solve that cost problem?

Richard: Certainly that’s part of the unlock. In some of these cases, there’s an elasticity between price and volume. If you can reduce that price, the volumes increase massively. We see areas like smart agriculture as well, being able to get information maybe at the plant level, where you can then optimise irrigation and when you might add nutrients, to even more efficiently grow and optimise yields. The combination of form factor and the ability to manufacture at really high volume — we’re already manufacturing in the UK for these kinds of products in the billions of units, with the ability to scale to at least five times that capacity just in our existing facility in Durham.

Lawrence: How? How are you making these so cheap? Talk me through the manufacturing.

Richard: First, any semiconductor manufacturing is not straightforward. It requires very reliable, proven manufacturing equipment. We have tools in our fab in the UK that you would see in any fab in the world, including TSMC. They’re well proven and designed to run 24/7 with high reliability. What we do with that is we use different materials, and each of our process steps is very short. Because we’re using thin films, the time to do a process step is very short. Actually, the bulk of our manufacturing cycle time is queue time — it’s wafers waiting for tools to become available to go on to the next step. So we can actually manufacture with a raw process time of a few days to make a chip.

Lawrence: A few days. What does that look like in steady state?

Richard: It’s longer, but we’re talking weeks rather than months. And that also allows us to reduce the footprint of our fabs because we don’t have as much work in progress. Our fab is essentially modular in design, it’s 20 by 30 metres, and from that we can do billions of chips. It’s a really compact design, and that means it’s more energy efficient, and uses obviously less carbon as a consequence.

Lawrence: 20 by 30 metres. That’s the size of a tennis court and a half. And you’re producing billions of chips from it. OK. So why build it in the UK? I hear constantly that the UK has high energy costs, it’s not a manufacturing hub. Why Durham?

Richard: A few reasons. The footprint of our fabs is relatively small, so actually it’s not as energy intensive as pretty much any semiconductor fab. Yes, we would like to see lower energy costs, they’re a contributor. But they’re not as punitive as they are for some people. From another perspective, we’re British as a business, and we’ve been here for 15 years, and we want to develop the core of the technology and our manufacturing base here. Part of it is a desire to make this work in the UK, and that makes some things a little harder. But that’s definitely our intention.

We’ve been able to attract the talent that we need through a range of routes, including repatriating people that worked in the semiconductor industry in manufacturing in the 1980s and early 90s, some of whom were already in the region, recruiting internationally, and developing a talent pipeline. I think it is possible, and we want to make it work. The government recognises energy costs are too high. I’d like to see quicker movement on ways to bring those down as a broader benefit to the UK economy. But it’s an important part of the mix for us, not the critical decision maker at this point.

Lawrence: What about the cluster argument? Saxony gets thrown around a lot. Are you fighting a good fight alone up in Durham, or is there a supply chain building around you?

Richard: I would actually take the UK as a cluster. I think we’re small enough not to be thinking about regional clusters. If you look at OEMs and chemical suppliers, they’re going to think UK-wide. There are other manufacturers in the UK that use some of the same suppliers, same chemicals. The majority of people we work with, there are European hubs — some in the UK, but many in mainland Europe for the OEMs. We have local support that’s usually only an hour or so away. I don’t think the UK is in that bad a position.

If you look more broadly — there’s the compound semiconductor cluster in South Wales, Seagate in Northern Ireland that’s been around for a long time and is an often untold success story. There’s still lots of activity in Scotland. Photonics in areas like Southampton, design strengths around Bristol and Cambridge, and a pretty strong academic community distributed around the country. We don’t have large fabs in the UK. But we’ve got significant strengths in quite a number of niches and an opportunity to build on those.

Lawrence: You currently operate as a hybrid IDM — you design and manufacture. Is that the long-term model, or does this evolve?

Richard: We’re manufacturing and designing our own products now. We see that trend moving more to foundry over time. But there’s a hybrid opportunity because we have this very compact manufacturing footprint. We also see the opportunity to deploy our manufacturing at customer sites. We would operate the fab on behalf of customers, but they would design the products. It’s a bit of a hybrid model.

One of the reasons we’ve had this specialisation in silicon is in large part because the cost of the research and development, and the capital cost of deploying new fabs as you’ve gone to more advanced nodes, has just increased astronomically. You go from tens of players being able to do manufacturing down to only three that are able to do it. It becomes a challenge because you need 20 to 40 billion dollars to deploy a fab.

Lawrence: The idea of deploying a fab at a customer site — that’s a genuinely different model. You can’t ship a TSMC facility somewhere. But 20 by 30 metres, that’s portable. Let me ask about applications beyond smart labels. You mentioned wearables and AR/VR?

Richard: We’re working with customers around miniaturisation, using our flex IC essentially as a smart substrate. You can do fine line interconnects and then build on top of that systems — integration of silicon electronics and surface mount components. Over time, take some of those capabilities and integrate them into the substrate itself. Things like resistors and capacitors can reduce the number of surface mount devices, reduce the BOM, and make the whole system smaller, more flexible, and lower cost.

There’s quite a lot of market pull in wearable devices, things like AR/VR headsets where volume actually becomes a driver — not just footprint, but the total physical volume that the electronics occupies. If you can shrink that down not just in x and y, but also in the z axis, and make it more flexible, you’re then able to deploy that with a better form factor in devices that require flexibility.

Lawrence: What about healthcare? That feels like a natural fit for something flexible and cheap.

Richard: We’re actually seeing real opportunities in healthcare. You’ll have seen things like continuous glucose monitors emerge in the last several years, increasingly moving to a consumer product. The ability to make something even thinner and more comfortable, at a cost point that allows it to be democratised — available not just in the developed world, but also in economies that don’t have sophisticated healthcare systems.

Things like brain-computer interfaces and other types of healthcare wearables where the combination of the flexibility and something that doesn’t have rigidity allows you to get a better interaction with the body, to conform and move with the body when it’s being worn. I think we’ll see more in that direction. And it’s something I’m really passionate about from a personal perspective. We’ve been working on a number of proof of concepts for several years.

Lawrence: There’s a lot of focus right now on sovereign AI, strategic semiconductor independence. Is there a story for flexible ICs in that narrative, or is that trying to put a flat peg in a round hole?

Richard: It depends a little on definitions of AI. Essentially, what we’re allowing is capture of additional data from a whole range of different environments and objects. That data will feed AI. We have the ability over time to do very simple decision-making or machine learning at the edge or the item, and to enable some of those decisions to be pushed to the edge. So you’ve got less requirement to move data up the stack. We see opportunities there. But as we talked about earlier, we’re not doing cutting-edge GPUs.

Lawrence: Right. We’re talking about relatively simple classifiers, not distilled LLMs running on your devices.

Richard: Certainly not LLMs as they’re currently imagined. We generally look at optimising the circuit design, optimised for the specific job in hand. That allows you to strip back functionality that you don’t need. We actually published something last year on tiny classifiers where we’re using an evolution algorithm to optimise the circuit design. In some cases, you can reduce the complexity of that down to a few hundred gates for the task in hand.

Lawrence: A few hundred gates. That’s beautifully minimal. Are there any objects you’ve put your circuits inside that might surprise people?

Richard: A lot of the early demonstrators were things like beer bottles, so probably not surprising. But we’re seeing real opportunities in healthcare, as I mentioned — CGMs, brain-computer interfaces, other wearables. The combination of the flexibility and a semiconductor device that’s inherently flexible opens up a whole category of applications where the electronics can conform to the body rather than sitting rigidly on it. I think that direction has a lot of potential.

Debrief

This interview series has, without quite planning it, been mapping different layers of the computing stack. Synthara and SEMRON are rethinking computation at the memory level, stopping data movement before it starts. Phanofi is making the movement that remains as efficient as possible with coherent optics. Wave Photonics is working on the photonic integrated circuits that could redefine how light carries information on-chip. All of these operate at or near the data centre.

Pragmatic is working at the other end entirely. Not faster computation for centralised AI, but dispersed, purpose-built computation for the physical world. The connective thread is the same question: where in the stack can you add intelligence, and what do the economics have to look like for it to make sense? At the data centre, the answer involves billions of dollars in capital expenditure on cutting-edge fabs. At the item level, it involves fractions of a penny on a chip manufactured in days.

The healthcare angle is what stuck with me most. Continuous glucose monitors that are thinner, cheaper, and comfortable enough to wear without thinking about them, available in countries that can’t afford the current generation. Brain-computer interfaces where the electronics flex and conform to the body. This is where flexible semiconductors can have real impact beyond packaging.

The deployable fab model is the other idea I keep coming back to. In semiconductor manufacturing, scale has always meant centralisation: bigger fabs, more capital, fewer locations. Pragmatic’s compact footprint inverts that logic. Ship a fab to a customer site, operate it on their behalf, and you’re not just selling chips, you’re offering manufacturing as a service, distributed rather than centralised.

One question lingers. How big can the edge intelligence story actually get? Tiny classifiers on a few hundred gates are elegant, but the gap between that and useful autonomous decision-making is big. The near-term value is clear, smart labels, sensing, unique identification. The long-term vision of purpose-built intelligence on every object depends on use cases that aren’t just technically feasible but commercially justified. Pragmatic has the manufacturing story figured out. The next chapter is proving the world actually wants billions of intelligent objects, not just billions of smart labels.

For now, the Durham fab is humming. Billions of flexible chips, manufactured in days, going onto objects that never had computation before. If the future of AI depends on richer, more diverse real-world data, someone needs to build the capture layer. Pragmatic is making a credible case that they’re it.

Check out Pragmatic website for more, and Richard is here.

Bye.

It’s getting real out there. And it’s only going to get real-er. Every passing day, it becomes clearer that were are, indeed, living through a dislocation. It’s easy to say: it’s like the industrial revolution and we should expect geopolitical and social upheaval. But living that, well it’s tough meat out there isn’t it?

What what can one do? Well, you can write newsletters I guess? That’s a start. Solid ground. You can just bury yourself in your work while you can. All I do is sit down at the typewriter, and start hittin’ the keys. Getting them in the right order, that’s the trick. That’s the trick.

And on that note, thank you to all(?!) the new paid subscribers this week. If I keep this up, I can give up the VC lark and actually get paid now instead of in 12 years! Like honestly, what’s the actual point of waiting until 2038 to get paid? Can you imagine the world in 2038? lol, imagine you are actually paying into a pension right now…

Paid subs should start to expect some exclusive content over the coming weeks, including some janky Claude Code-produced dashboards tracking the things I keep writing about (AI labour market data, European sovereignty investments, landscape design courses near you, etc). They will look wonderful. But they will break. It’s all that auth and token refresh that keeps getting me. Amateur hour indeed.

So, with that in mind, stop paying into your pension and pay me instead? The information in this newsletter and coming in vercel dashboards will be more valuable than a pension in 2050.

And before the future gets you down too much, as Billy Lemos says: “wake up and get sexy”. Listen below and enjoy a lovely weekend! DMs open as always.

1. Anthropic vs the Pentagon: When Your AI Company Gets Designated a National Security Risk

Yikes, so the biggest AI governance story of the year, and it happened in about 72 hours. Anthropic refused to remove two contractual redlines from its Pentagon deal: no autonomous weapons without human oversight, and no mass domestic surveillance. Sounds fair beans from my lilly-livered perspective. Pete Hegseth responded by designating Anthropic a “supply chain risk to national security” and President Trump ordered all federal agencies to stop using Claude, with a six-month phaseout.

The legal basis is, to put it politely, dubious. Title 10 Section 3252 defines supply chain risks as involving potential sabotage or back doors, not philosophical disagreements about use cases. GWU law professor Jessica Tillipman called it “so legally dubious.” A defence official evaluating supply-chain threats said there was “no evidence of supply-chain risk,” calling the designation “ideologically driven.” You will also note no Chinese model company is a supply-chain risk. Intriguing business.

Then it got messier. An internal Slack message from Dario Amodei leaked, with pointed criticism of OpenAI’s approach. OpenAI, which had rushed to announce its own Pentagon deal the same night, later backpedalled, with Sam Altman saying they “shouldn’t have rushed” and outlining revisions to their own safeguards.

Anthropic and the Pentagon are back at the negotiating table. But the precedent is somewhat concerning: a government using procurement designations as a political weapon against companies that maintain safety guardrails. Well, as I said up top, we are living through a dislocation so expect precedent to break fairly regularly from now on.

Read Dean Ball on this, it’s as exceptional as everyone says it is.

Source: Defense One | CNBC | Zvi’s full breakdown

2. Block Cuts 4,000 Jobs “Because of AI” and the Stock Surges 24%

.@Jack laid off nearly half of Block’s workforce, taking headcount from over 10,000 to under 6,000. His explanation: AI means the company can do more with fewer people. “Within the next year, I believe the majority of companies will reach the same conclusion and make similar structural changes.”

Block’s stock went up as much as 24% in extended trading…

Now, i’ve been writing about this exact scenario for months. In “Unbundling the Job” I argued that AI was stripping out “the glue work, the learning-by-doing, the apprenticeship layer” from roles, making positions modular and automatable. In “What happens if mass unemployment never arrives” I made the case that the more likely outcome is performative work, not mass layoffs. Well, Dorsey just showed us what happens when a CEO decides to skip the performative stage and go straight to the cuts. But, the wrinkle is obviously, that Block just overhired and they are using AI as cover. A few others will follow suit. Likely alot of the software/SaaS firms will react faster than most to sure up their stock price as contracts are renegotiated downwards and top line begins to soften. But, I don’t expect mass layoffs like this from most. Most will wait for a downturn or recession and then the jobs will go and they will never come back

Source: Fortune | Bloomberg

3. Anthropic’s Own Research: 75% Exposure for Programmers, But No Unemployment Spike

But narrative violation. The same week Anthropic is fighting the Pentagon over safety guardrails, their research team published the most rigorous study yet on AI’s actual labour market impact. And they found, it hasn’t caused mass unemployment. Not yet, anyway. Not yet!

The researchers introduced “observed exposure,” a metric that combines what AI can theoretically do with what people actually use it for. Computer programmers top the list at 75% task coverage. Customer service reps and data entry workers follow closely. But, crucially, workers in highly exposed occupations haven’t seen higher unemployment rates since late 2022.

The surprising profile of the most exposed workers: older, female, more educated, and earning roughly 47% more than unexposed counterparts. This isn’t blue-collar automation. It’s white-collar augmentation, for now.

The one warning sign buried in the data: hiring of younger workers in exposed fields has slowed noticeably. New grads aren’t being fired; they’re not being hired in the first place. That’s the Block story in slow motion, and it’s the mechanism I described in “Unbundling the Job,” where the apprenticeship layer gets automated before the senior roles do. “But, but, if you don’t hire into junior positions today, you won’t have a pipeline of human works in senior positions tomorrow” I hear you shout. Yes, quite. Human workers, indeed.

Source: Anthropic Research

4. Europe’s €2.5bn NanoIC Chip Lab Opens, ASML’s Next-Gen EUV Arriving Mid-March

And semiconductors, because, well, you get it. IMEC in Leuven just inaugurated NanoIC, the largest pilot line under the EU Chips Act, with €2.5 billion in combined funding: €700 million from the EU, €700 million from national and regional governments, and the rest from industry partners including ASML. ASML man, they are moving.

The headline piece of kit: ASML’s next-generation High NA EUV scanner, arriving mid-March. This is the machine that enables chips beyond two nanometres. NanoIC is the first European facility to deploy it. Guys, guys, guys. I wrote about this, in 2023! Don’t you dare say I don’t know what I am talking about porkchop761 in the DMs.

Anyway, i’s an open-access platform where startups, researchers, and SMEs can test chip designs at near-industrial scale before committing to full production. Six countries are involved: Belgium, France, Germany, Finland, Romania, and Ireland. It’s part of a five-pilot-line network representing €3.7 billion in combined investment.

I’ve spent the last few months interviewing people building the compute infrastructure layer, from Synthara’s compute-in-memory chips to Phanofi’s coherent optical I/O to Wave Photonics’ GaN PICs. All of these companies need access to advanced fabrication to get from lab to product. That’s exactly what NanoIC is designed to provide. Fabs. Fabs. Fabs. Absolutely Fabulous.

Source: Evertiq | HPCwire

Also Worth Your Time

Matt Yglesias argues we’ll miss the sweatshops. His piece in The Argument makes the case that AI-driven automation could kill the development ladder that historically lifted poor nations out of poverty. Textile manufacturing was the first rung of industrialisation for Britain and nearly every success story since. If robots can do it cheaper, that rung disappears. I wrote in “What happens if mass unemployment never arrives” about occupational downgrading in the West, but Yglesias is pointing at something more brutal: entire countries locked out of the benefits of the post-scarcity transition. Read it here.

Eat, pray, love. Bye.

If you missed it:

• What happens if mass unemployment never arrives — AI won’t create unemployment; it’ll create performative work

• Unbundling the Job — what we lose when the job stops being the social contract