Part 4: the new frontier

Hello friends, and welcome to The Unsophisticated Investor! Brought to you by Mark, Head of UK Expansion & Operations at Shuttle!

If you want to invest alongside the VC funds who've backed breakout companies like Revolut, Monzo, Asana, JustEat, Bolt, Lets Get Checked, Loom, Runna, Charlotte Tilbury, Deel, Aircall, AngelList, Carta, TransferWise and many more, regardless of your knowledge, network or net worth, then sign-up today using the link below.

Now, let’s get into it 👇

For seven decades the architecture of venture capital, from the ten-year lock-ups to the layers of liquidation preference, rested on one premise: capital was scarce, and access to it was a guarded privilege. Over the last three episodes we have tracked VC’s humble origins from the hardware syndicates of the 1950s, through the unlocking of US pension funds in the late 70s, and into the zero-rate era that allowed sovereign wealth funds to set global valuations.

That premise no longer holds. Capital has become abundant and, in many parts of the market, undifferentiated. When money is no longer the scarce input, the traditional $1bn mid-market fund loses its economic rationale, and the industry has separated towards its two ends: sovereign-scale managers financing infrastructure, and a new class of small, specialist funds writing the first cheque. This is the new frontier, and it is a more open one than what preceded it.

Section 1: what venture built

Venture capital is still a relatively small industry by any macroeconomic measure. In a strong year it deploys less than half of 1% of US GDP. Yet companies financed by it account for a substantial share of the public market value created in the US since the late 1970s. The list of what it funded is essentially the list of what changed the modern economy: the semiconductor industry, the personal computer, the internet, mobile computing, biotechnology and the platform technologies that underpin drug discovery today. No other financing structure has matched that ratio of capital consumed to value created.

The reason is that venture solved a problem other capital could not. Banks lend against assets and cash flow, and a company with neither cannot borrow. Public markets price businesses with a history. The limited partnership was engineered to fund the specific case of a technically credible team with no revenue, no collateral and a long path to proof, and it did so by accepting that most attempts would fail and organising itself around the few that would not.

That is difficult to do, and the returns to getting it right accrue well beyond the fund. Venture-backed companies employ people at above-average wages, concentrate private research and development spending, and generate the supply chains and tax payments that follow. The failures matter too, since the engineers and operators they train tend to reappear in the next cohort of companies. An asset class consuming a fraction of national output has repeatedly underwritten the technologies that later became general infrastructure, and it did so at a stage when no other pool of capital would take the risk.

Europe's difficulty has never been an excess of this activity; it is a shortage. European venture investment reached about $44bn in 2025, its strongest level since 2021, with deeptech taking 36% of the total. Set against the US, Atomico estimates European technology companies were underfunded by roughly $375bn over the past decade. The continent's problem is not that too much capital chases its founders. It is that too little does, and that the capital which exists reaches too narrow a set of people.

Section 2: when capital stopped being scarce

The change of the past six years was monetary before it was anything else. Zero-interest-rate policy, sustained from 2008 and reimposed in 2020, functioned as an instruction to institutional capital to move along the risk curve. When the risk-free rate approaches nothing, an allocator with a fixed liability, such as a pension fund promising 5-7%, must seek duration and illiquidity to meet it. Venture sits at the absolute end of that curve. (The same conditions expanded private credit from a post-crisis niche market into one now measured in trillions, as banks retreated under tighter capital rules.)

US venture investment ran at roughly $170bn in 2020 and reached about $345bn in 2021. Several hundred private companies crossed a $1bn valuation in that single year, more than in the preceding decade. Tiger Global industrialised deployment, completing diligence in days and backing more than one company every working day at the peak. The logic was internally consistent: if you cannot pick winners in advance, buy the index. So long as entry prices leave room to absorb the losses. That condition depended entirely on discount rates remaining at zero.

Speed replaced price as the competitive variable, and two disciplines weakened as a result. Pre-emptive rounds, offered before a company had decided to raise, removed the negotiation in which valuation is ordinarily established. Governance provisions followed; an investor asking for board seats and information rights in a market clearing at that pace lost the allocation to one who did not. Neither was surrendered through carelessness. Both were competed away by participants who understood the trade and judged access more valuable.

Rates normalised in 2022, rising from near zero to above 5% in roughly eighteen months. Long-duration assets repriced hardest, and venture is the longest-duration asset in the equity complex. Silicon Valley Bank, through which much of the sector banked, failed in March 2023 when the same rise destroyed the value of securities it had bought with boom-era deposits. Klarna raised at roughly 85% below its 2021 valuation, Instacart marked itself down from about $39bn towards $13bn, and Stripe cut its internal valuation from $95bn towards $50bn (though that has now recovered three times in value since).

Those markdowns were less dramatic than they sound. A private valuation is not struck in a market. It is agreed between a company and whoever writes the newest cheque, and what they agree depends on the rest of the contract. An investor entitled to be repaid first in a sale can afford to be relaxed about the headline. The founder gets a number to announce, and the investor gets terms to rely on. Higher rates did not so much invalidate those valuations as expose how little of the company they had actually described.

The correction also did something useful. Burn rates fell, pricing improved, and businesses that could not have survived on 2021 economics reached profitability on 2023 economics. Venture exists to fund experiments, and experiments are supposed to conclude when the evidence turns.

Section 3: the deployment trap

The mid-market fund's difficulty is arithmetic rather than character.

A $2bn fund generates $40m a year from a 2% management fee, and justifying it requires deploying $2bn. Deployment capacity does not stretch to meet the money. A partner can hold eight to ten board seats and monitor them properly, so ten investing partners support perhaps eighty positions and in practice concentrate on fewer, since reserves must be held for later rounds. Spread across forty core positions, the average cheque approaches $50m.

Everything follows from that cheque. An investor targeting 20% ownership cannot deploy $50m into a company valued at $30m without taking most of the equity and removing the founder's incentive. The cheque size therefore pushes the fund towards companies already valued in the hundreds of millions, where the product exists and much of the risk has been priced by whoever took it earlier. Late-stage investing at this scale is not a preference. The fee base compels it.

The return requirement compounds it. A $2bn fund returning three times gross must generate $6bn. A company exiting at $1bn which most founders would consider a triumph returns $200m on a 20% stake, or just 10% of the fund. Scale converts a business that benefits from rare successes into one that entirely depends on them.

Section 4: sovereign-scale funds and infrastructure

Andreessen Horowitz (since rebranded to a16z) recognised this arithmetic early and followed the logic to its natural conclusion. In 2019, it surrendered its status as a venture capital firm in the eyes of the SEC and registered as an investment adviser, trading a regulatory exemption for the freedom to hold anything: tokens, listed equities, secondary positions, stakes in other funds. Its media operation addressed a different constraint. When founders can raise from anyone, founder attention replaces capital as the binding input, and distribution becomes worth more than the balance sheet. Both moves were rational responses to a market flush with cash.

By the mid-2020s, a16z managed roughly $45bn, and in 2025 moved to raise a single artificial intelligence vehicle of around $20bn, close to a fifth of all US venture capital raised that year. At this scale, venture moves further up the capital curve into the realms of capex financing used by private equity firms. Deploying tens of billions into frontier AI is ultimately a physical capital expenditure programme, closer to building national grid capacity, semiconductor fabrication plants or an aerospace programme than to backing startups.

Training a frontier model consumes processors, data centre shells, transmission capacity and power, in quantities measured in tens of billions per company and spent years before revenue arrives. The relevant expertise is procurement and energy contracting. There are no ten-year funds raised from pension capital that can realistically write cheques at that scale.

As a result, the capital comes from balance sheets instead. Microsoft committed well over $10bn to OpenAI, much of it as Azure credits. Amazon and Google committed comparable and then larger sums to Anthropic. Nvidia took equity in companies that rank among its largest customers. This gave way to a circularity of investment flows within the sector, and a source of concentrated risk for the years ahead. A hyperscaler supplies compute recorded as investment, the developer consumes it, and the consumption returns as cloud revenue supporting the next round. The commercial rationale is genuine, since compute is the binding constraint. But aggregate investment statistics now include vendor financing that no external market has priced.

The valuations reflect this dizzying reality. OpenAI moved from about $28bn in early 2023 to around $850bn by early 2026, and Anthropic from about $60bn in early 2025 toward a staggering $1tn by mid-2026. Established firms appear on these registers alongside strategic and sovereign investors who set the terms. By the first half of 2026, close to 90% of US venture dollars went to artificial intelligence, according to PitchBook. Compute has acquired the strategic character of energy or defence manufacturing, and states now treat it as industrial policy rather than portfolio construction.

The traditional venture partnership retains the layer above, where companies built on foundation models remain conventionally fundable, and that is where the next decade of venture returns most plausibly sits, but also where some of the greatest concentrated risk will lie too.

Section 5: the case for the emerging manager

The more interesting development sits at the other end of the barbell, and it is a genuinely positive one for founders.

A $25m fund generates $500,000 a year in fees, enough to support a lean investor with admin support or third party support if required. Investors, though remunerated well, are tied to the carried interest that comes from outsized returns. The alignment such managers describe is not a marketing position but a strict condition of the fee structure: they are paid when their founders succeed and not otherwise.

Cheque size inverts too. A $25m fund writing $500,000 cheques takes thirty positions and can enter at a $5m valuation, before the product exists, purely on the strength of knowing the founder and understanding the problem. One company reaching $500m returns twice the fund. The mid-market fund needs a decacorn to register; the small fund needs a good outcome.

Consequently, the earliest and hardest cheque, the one written when there is nothing to underwrite but a person and an idea, is now being written by people with the specialist knowledge to judge it. The infrastructure that made this possible was administrative. Platforms like ours at Shuttle that reduce the of running a fund and deal to something one person can manage, via rolling funds and special purpose vehicles allowed capital to be raised in increments. About half of new US venture funds raised in 2020 came in under $25m and by 2024, roughly seven in ten did.

Founders are the clearest beneficiaries. A specialist manager who has built in the same category can assess a technical claim that a generalist committee cannot and can do it in days rather than weeks. Decisions come from one person with conviction rather than a committee optimising for consensus, a dynamic which naturally favours unusual companies over legible ones. Because this class of manager is far more varied in background, geography and network, capital reaches founders who were previously outside the blast radius of a few firms clustered in a few cities.

There is a systemic function here as well. As the large funds moved upmarket, they left a genuine gap at the earliest stage, where cheques are small, diligence is qualitative and the economics simply do not work for a firm managing billions. That gap is where companies are actually formed. The managers filling it are taking the first risk on which every later round depends, performing the price discovery that institutional capital relies on when it arrives two or three rounds later. A healthy pipeline for growth funds, and eventually for public markets, requires somebody to fund the stage before any of it is legible. The emerging manager class has quietly become that somebody.

But this model carries distinct structural risks. Small funds hold no reserves to defend their winners in later rounds, so their best positions dilute. Dispersion at this end of the market is extremely wide, and reported outperformance carries heavy selection bias, since funds that fail quietly stop reporting. A first-time manager has no fees from a prior fund to absorb a slow period, and when allocators retrench they cut exposure to newer names first. Many who closed a first fund in 2020 or 2021 never closed a second. This is a demanding way to make a living, and the managers who have raised again have generally done so on a demonstrable edge in one sector or community.

The music industry offers the closest parallel. Labels once advanced capital, controlled distribution, decided who reached an audience, and took ownership as payment. Artists accepted the terms because the label held the only route to a listener. Then, streaming and social platforms commoditised distribution and discovery. Artists with direct followings began keeping their masters, trading mass reach for smaller, higher-intensity communities.

The intermediary did not disappear, but it lost pricing power. The economics moved toward whoever held the direct relationship. Venture is at exactly this point. The solo manager occupies the position of the independent artist: a narrower constituency, higher intensity per relationship, and the economics accruing directly to the person doing the work.

Next week

We move from the structure to the operators, and specifically to Europe, where the shape of the market differs.

The US barbell formed because private capital concentrated at the top, with the ten largest funds absorbing roughly 40% of the market’s investable capital. Europe does not concentrate that way. Public institutions supplied about 38% of European venture commitments in 2025, between six to eight times the US proportion, and the European Investment Fund's mandate heavily favours diversification across many managers. The ten largest European funds take about 20%. Europe is, by design, a small-fund market, which makes it the most interesting place in the world to examine how this new model actually works.

Next week we profile the solo GPs and emerging managers changing the model throughout Europe and bringing highly specialised experience to the next generation of breakout founders. We will look at what they back, how they decide, and what founders get from them that they cannot get anywhere else.

What we’ve been working on at Shuttle

• Final touches to our new payment partner integrations 💸

• A lot of work on application and infrastructure observability ahead of our UK launch 🔐

• Working with US counsel to map our path for supporting US based investors and founders 📝

The Unsophisticated Investor is brought to you by Mark, Head of UK Expansion & Operations at Shuttle. Shuttle's mission is to break down the barriers to private markets and make wealth-building opportunities accessible beyond the ultra-wealthy elite. To do it, we're building Europe's private market infrastructure platform for running investment operations. By stripping away the legal friction and admin for deal managers, we make it easier than ever for high-performing opportunities to open up to a wider, modern network of investors.

Mark
Shuttle’s Head of UK Expansion & Operations