Fund Frontier Expeditions otherwise just buy the "Magnificent Seven"
There be Monsters. Big dreams need money. But who gives these dreamers money?
Portuguese ships chasing spices. Spanish galleons hunting gold. Google organising the World’s Information. Amazon offering the Everything Store. And 30-minute delivery of Milk?! Behind every audacious dream lies the same question: Who funds the dreamers?
Venture capital, from which I make a living, is basically giving merchants and entrepreneurs money today for the promise of a lot of money in the future. 3x in 10 years, that’s the professional version. Old school is was bring back gold or spices quick snap. There are plenty of less risky ways to make a return on your money. There is yield from bonds, which pay interest; dividends from established companies; rental income from real estate. But VC is different, it’s is not money management. It is risk capital.
Over the past five decades, VC has evolved from a niche investment category into a mainstream asset class, driving the growth of startups that have transformed entire industries and culture. VC's share of U.S. private sector investment went from less than 1% in the 1980s to roughly 5% in 2023. It’s still a niche sure, but a growing one. Or rather it was. Last week, the FT reported that more than half of the $71 billion raised by U.S. VCs in 2024 was secured by just nine firms, a seven-year low and less than two-fifths of the total haul in 2021.
Venture capital is shrinking and consolidating because of higher interest rates and longer time to IPO. But I’m out here claiming it will be structural not cyclical. Interest rates and longer time to IPO are likely to be cyclical on some timeline. But there is a structural change that has occured and will be consolidated by AI.
Disruption theory has been so well internalized by Big Tech that it is now longer predictive of the tech industry. The tech industry is now so large, enchrenched and profitable that yes one or two scrappy startups might make it, but they will be so rare as to be worthy of their unicorn name. And AI brings with it high capex but more importantly high opex that those with the deepest pockets are likely to win.
For VC to survive it must become risk capital again and fund risky expeditions.
Higher interest rates
The era of near-zero interest rates, lasting over a decade, fueled an unprecedented boom in venture capital. With traditional fixed-income investments offering negligible returns, institutional investors and high-net-worth individuals turned to venture capital in search of yield. This influx of capital drove down the cost of funding for startups, enabling them to raise massive sums at sky-high valuations. Startups prioritized rapid growth over profitability, aka blitzscaling, and the number of unicorns exploded from fewer than 40 globally in 2013 to over 1,200 by 2024. Cheap money allowed founders to pursue ambitious goals (blood tests in seconds? reimagine how the workplace can empower people to be impactful and fulfilled? cold-press juicing devices?) but it also fostered excessive risk-taking and inflated expectations.
This abundance of capital reshaped the dynamics of venture investing. Startups delayed liquidity events, opting to raise successive private funding rounds rather than going public. With investors willing to lock up capital for extended periods, timelines to exit lengthened, and valuations often outpaced value. The VC ecosystem expanded dramatically, but it became increasingly detached from traditional measures of financial sustainability. This period exhibited the hallmarks of a bubble: overvalued assets, unprofitable companies commanding premium valuations, and a frenzy of capital chasing growth at any cost.
When interest rates began rising post-2022, this dynamic reversed. Capital became more expensive, valuations contracted, and many startups struggled to raise funds. The prolonged liquidity timelines and lack of near-term profitability left some startups vulnerable to collapse. Limited partners, now presented with more attractive opportunities in fixed-income and public markets, grew reluctant to commit to venture capital funds with decade-long lockups. The end of the zero-interest-rate era has forced a painful reset for the VC industry, underscoring the extent to which the boom relied on cheap capital. The boom in AI came at exactly the same time the interest rate environment changed so we haven’t yet seen a full return to sense. It will likely take a few years for the funds already raised to be deployed for the true reckoning to arrive. Expect H2 2025 and onwards to be full of bad news stories.
Longer time to IPO
One of the biggest challenges facing venture capital is the growing delay in liquidity timelines. In the 1990s, startups typically reached an exit in about seven years. Today, this often exceeds a decade. Stripe, founded in 2010, remains private after 13 years at a $95 billion valuation, while SpaceX has stayed private for over 20 years, now valued at $125 billion. But I mean I don’t think we can generalise anything from Elon tbh. This contrasts sharply with earlier IPO waves: Apple and Microsoft went public within a decade of founding, and internet-era companies like Google and Amazon went public in under seven years. Such delays strain the traditional 10-year venture fund model that once aligned with quicker exits.
Several factors drive the elongation of liquidity timelines. First, an abundance of late-stage private capital allows companies to raise significant funds without turning to public markets. Mega-rounds from venture and crossover investors provide the runway needed for sustained private growth. Second, regulatory changes, such as the Sarbanes-Oxley Act, have increased the costs and complexity of going public, making IPOs less attractive. Third, founders often prefer the relative stability of private ownership, avoiding the volatility and scrutiny of public markets while maintaining greater control over their companies’ strategic direction. These dynamics collectively shift the incentives, delaying the transition to public markets.
For VC firms, longer liquidity timelines mean delayed returns and greater difficulty recycling capital into new investments. Limited partners (LPs) also face extended durations before realizing gains, prompting questions about whether venture capital still offers a compelling risk-reward tradeoff compared to other asset classes with shorter horizons. As startups stay private longer, the traditional 10-year venture fund model increasingly struggles to align with the extended timelines required for meaningful exits.
Magnificant Seven Advantage in AI
Now the more structural trend. In the past decade, we’ve seen sustained massive equity value growth from most tech stocks to a level previously thought impossible due to the law of diminishing returns. Not only have we crashed through the $1 trillion dollar market capitalisation for the first time, but $1 trillion soon became three. And there is no reason to believe it will stop there. In the last decade, the Magnificant Seven stocks have been on a tear: NVDA 12,000+%, TSLA 2,200%, MSFT 900%, AAPL 840%, AMZN 635%, META 545%, and GOOGL 390%. So if you held these stocks for the last decade, assuming equal weighting, you would be looking at roughly a 2,500% return. Nvidia skews the returns, but even if you remove it, you end up at about 900% return. To be in the top decile VC fund you should return 300% over a 10-year fund. Even without Nvidia, you would be 3x better off not allocating to VC.
Now past performance is no guarantee of future results, and the law of diminishing returns should constrain growth to some extent, but I find it hard to make the case these companies will get weaker. In fact, the received wisdom of disruption in the technology sector seems to grow weaker every year. The general view is that every platform shift from the mainframe to PC to Internet to smartphone disrupts the value chain. Previous winners like IBM, Intel, Nokia and Blackberry go out of business, and new startups take over. At this point this theory is closer to fiction than fact, with Apple, Microsoft and Nvidia all outlasting multiple platform shifts successfully. Meta, Google and Amazon, all successfully navigated the shift from PC to smartphone. The next platform, VR/AR/spatial computing, may see a new device maker win, but it’s hard to look past Apple, Google, Meta, or Microsoft. Especially when you consider their advantages in distribution, ecosystem and cash reserves .
The shift from smartphones to AR/VR will take a few years, but the real bull case for incumbents winning over startups is in AI. AI services, unlike Internet services like search or social media, is capital intensive and does not offer zero marginal costs. The Internet’s impact was facilitated by a relatively low capital expenditure burden for startups, largely because the groundwork—such as extensive fiber-optic cabling—had already been laid during the telecom boom before the dot-com bust. This allowed internet-based businesses to scale rapidly on pre-existing infrastructure with near-zero marginal costs. In contrast, o3, OpenAI’s most recent “large reasoning model,” shifts the paradigm for AI. Unlike traditional LLMs like ChatGPT, Claude, and Gemini, which rely primarily on static, pre-trained models, o3 incorporates dynamic, real-time reasoning and fine-tuning during use. This innovation increases operational costs, emphasizing test-time compute and ongoing inference over one-time training. No longer is software instantly scalable at close to zero-marginal costs.
Consequently, AI demands both high upfront capital expenditures for infrastructure and sustained operational investments, creating barriers to entry. Microsoft’s capital expenditures are already expected to exceed $60 billion annually, with Google, Amazon and Meta all in and around the same numbers. I wouldn’t be surprised to see more than one of these hyperscalers spending $100 billion in AI capex in 2025. And with test-time compute, opex will matter again.
There will always be opportunities for startups to out-execute incumbents, and OpenAI and Anthropic exist, but these capex and now opex numbers are a different world to the Cloud and SaaS Capex-light startup world the VC industry has been structured around for the past decade. if people thought AI was going to be the next SaaS and save venture capital, they are going to be solely dissapointed. Unicorns are going to get rarer as delivering competitive services and reaching customers becomes expensive. AI will reshape the tech industry to look much like the automotive or oil&gas industry. High capex and opex with high barriers to entry, tighter margins, and fewer pathways to disruption. Watchwords will be return on invested capital (ROIC) and gross margins. If you want exposure to AI, you can buy in the public market and get the 300%+ return over a decade, with the massive advantage of liqudity.
The End of Venture Capital as We Know It
The abundance of cheap capital that fueled VC’s golden age, driving record valuations and an explosion of unicorns, is over. With higher interest rates making long-duration investments less attractive, LPs are turning to more liquid and predictable opportunities in public markets. At the same time, extended liquidity timelines, the dominance of tech incumbents makes the VC asset class ever more uncompetitive.
A return to the near-zero rates that once defined the VC boom appears unlikely. Geopolitical headwinds—including persistent inflation, the slowdown of globalization, and the end of China’s rapid economic growth—are creating structural pressures that support a higher-for-longer interest rate environment. While a global recession or deflationary technological breakthroughs could shift this trajectory, the most probable outcome is that venture capital continues to contract and consolidate. This is not the end of venture capital altogether, but it is the end of the era where easy money fueled rapid growth and transformative returns. What remains will be a leaner, more selective industry, focused on fewer, larger bets in an economic climate that favours incumbents over disruptors.
So what to do? Well it’s obvious isn’t it? Fund merchant expeditions to new frontiers. VC is *risk* capital. Otherwise you are just managing money.
Back entreprenueurs early to go on adventures. Back them to build galleons and buy compasses. Find new materials. Test nuclear reactor designs. Build quantum computers. And literally explore new frontiers with solar sails.