David Clark on Venture Fund Returns, Liquidity and Choosing Managers

David Clark says his firm’s data includes 45 investments that each returned at least $1 billion to a single fund and also returned that fund. He argues that today’s fund sizes should be considered against possible exit values 10 to 15 years later, while acknowledging that future outcomes are uncertain.

Clark has spent 32 years at VenCap, investing in venture funds as a limited partner. He grew up in Northumberland without knowing about venture capital and applied for a finance job advertised in an Oxford newspaper in 1992.

He became fascinated by the field as internet companies emerged. A distribution of Netscape stock gave him a reason to speak with its finance team, and he found the LP role offered a view into technologies changing society.

Selecting managers is difficult when there are many funds and paper valuations make almost everyone look successful. Clark says VenCap learned from the late 1990s, when it expanded its manager roster and backed young funds with impressive early marks that later looked very different.

His response is focus: VenCap does not need to find every great emerging manager, he says, but wants the managers it does select to be strong. The firm waited for clearer signals rather than adding many new relationships during the recent boom.

After the financial crisis, liquidity shortages and the denominator effect constrained some LPs, creating openings with established managers. Clark says a fund of funds that has already raised its capital can still invest through such a period; that is a description of VenCap’s model, not a guarantee of access.

Clark describes venture as a power-law industry. He says roughly 1% of exits, or around 30 companies a year in his analysis, account for more than half of global venture exit value. VenCap seeks managers that repeatedly gain meaningful exposure to these companies rather than trying to identify every possible emerging manager.

Stebbings questions whether multibillion-dollar funds can still generate strong multiples. Clark cites a PitchBook analysis of approximately 1,200 funds from vintages in the 2000–2014 period, although he is not certain of the exact endpoint: more than half had distributed less than 1x capital, 6.6% had reached 3x net DPI and 2.6% had reached 5x DPI.

Those are distributed-to-paid-in capital figures, not total value including unrealized holdings. Stebbings asks whether venture justifies illiquidity; Clark argues that consistently selecting top-quartile managers can outperform public equities. The cited data and benchmark comparisons are historical claims, not guaranteed returns.

Fund size remains a concern for VenCap. Clark says its approximately dozen core managers received about 90% of the firm’s capital over the prior decade-plus, and reports an aggregate net multiple of roughly 3.5x for their mature funds. His figure comes from VenCap’s portfolio and is not representative of every venture fund.

He also says fewer than 3% of those mature funds showed total value to paid-in capital, or TVPI, below 1x. TVPI includes remaining reported value, unlike DPI, so it should not be read as cash already returned. The portfolio and time frame differ from the broader PitchBook comparison.

VenCap found 45 investments in its data that returned at least $1 billion to an individual fund and were fund returners; Clark says many occurred in the previous seven or eight years and one returned $15 billion. These are figures from his firm’s dataset, not evidence that such outcomes are common.

Clark argues that comparing present fund sizes only with present exit sizes misses the long holding period. He expects technology to capture a larger share of economic activity and therefore expects larger exits over the next 10 to 15 years, while noting that valuation multiples can fluctuate. This is a forecast, not a known future outcome.

Stebbings worries that large incumbents have unprecedented advantages in data and computing. Clark accepts their strength but believes technological shifts can still disrupt them. He cites past platform changes and speculates that AI, and potentially blockchain, could create new openings.

Stebbings argues that AI’s dependence on compute and data ties the new platform closely to existing incumbents. Clark points to emerging companies including ByteDance and OpenAI, but acknowledges that markets differ and the final outcomes are still unknown.

Liquidity often arrives in short windows. Clark says LPs need managers who understand cycles and know when to realize value, while LPs should maintain commitments across vintages instead of trying to time the market. He notes that some 2019–2021 liquidity came from investments made around 2010–2011.

He expects antitrust scrutiny to make some acquisitions more difficult in the near term. Founders should therefore build durable standalone businesses rather than assume an early sale. He thinks a more constrained exit market could further concentrate venture returns, a prediction rather than an established result.

Clark does not claim to know the next winning geography or sector. His stated task as an LP is to find managers who can identify exceptional companies wherever they emerge. At the time, he says, around 70% of VenCap’s portfolio was in the US and 10% each in Europe and China; China’s share had fallen over the previous decade.

He says managers VenCap backed had exposure to 85% of the top 20 or 30 recent exits in its analysis, emphasizing meaningful exposure rather than a token stake. The statistic is the firm’s own retrospective measure and does not establish future access or performance.

Clark says VenCap’s last new manager relationship was around 2018, though it continued reinvesting with existing managers. Stebbings challenges whether a fund of funds that adds managers so infrequently earns its fees; Clark replies that access and net risk-adjusted outcomes are the product clients seek.

Stebbings argues that investment decision-making process is itself the product and may reveal future quality before outcomes are known. Clark responds that successful venture partnerships use many structures and backgrounds, and says VenCap has found it difficult to predict reliable determinants independently of results.

Clark acknowledges that venture feedback takes many years and that past output can be a lagging indicator. His approach is to return to underlying performance data, while Stebbings warns that a strategy proven in an earlier market may not transfer to new models.

Rather than test every new theory, VenCap aims to enter once it sees stronger evidence. Clark says the firm was looking at new managers again as valuations reset and expected differentiation to become clearer. Historically it often entered at a manager’s third fund, when more evidence about the first fund’s investments was available.

Clark wants more than a tiny check in a winning company: he looks for meaningful ownership and an investor’s role in supporting the founder. He says a good investor should know when to provide advice and difficult feedback and when to get out of the way.

VenCap’s sourcing is predominantly outbound, Clark says. The team tracks what it considers the top 1% of companies, identifies their early investors and researches unfamiliar names. It uses references to deepen its understanding after that screen.

Access can arise when established firms go through succession or other difficult periods and existing LPs are constrained. Clark emphasizes persistence and timing, but neither a challenge at a firm nor a liquidity shortage guarantees an available allocation.

He says VenCap had a good distribution year in 2023 partly because managers were still distributing shares of companies that went public in 2021. The inventory of public stock was falling, and he expected replenishing it through new IPOs and subsequent distributions to take time.

VenCap generally sells public stock it receives because its investors have separate public-equity managers. Clark does not require venture managers to distribute immediately: he thinks they should use judgment, especially when they know of a significant future opportunity at a company.

He prefers managers to distribute most holdings while making selective exceptions. Some venture-backed public companies continue compounding after an IPO, he says; his support for retaining select positions depends on manager judgment and does not imply that investors can reliably identify every future winner.

Stebbings asks whether exceptional later growth, such as Nvidia’s AI-driven rise, is predictable at IPO. Clark concedes it need not be. He recalls investors holding Square after listing because they knew about a potential additional product opportunity, illustrating the kind of specific reason he considers relevant.

Succession requires senior partners to make room for the next generation and share economics, Clark says. He looks for a continuing flow of talent, recognition for people who source important investments and a genuine path to influence firm strategy.

He acknowledges that LPs may learn about partnership problems a fund later, while excessive rumor can also obscure what is material. He cites Accel and Sequoia as firms with experience of generational transitions and respects Foundry Group for choosing to stop rather than pretend it wanted perpetual succession.

VenCap monitors existing managers continuously and still completes formal diligence before each recommitment. Clark says most managers are likely to be renewed, but the review and references help identify missed issues and document stewardship of investors’ money.

He says poor performance and succession problems are the main reasons the firm may decline to recommit. VenCap shares anonymized peer benchmarks across neighboring vintages, including IRR, TVPI and DPI, to help managers understand where outcomes or distributions lag.

The same data can have different explanations: a manager may deliberately retain a promising company rather than distribute, or may find other opportunities for liquidity. Clark favors honest conversations while recognizing that the manager may disagree with VenCap’s decision.

One lesson from a 1999-vintage fund that invested its capital in 15 months was the importance of time diversification. VenCap aims to deploy over roughly three years and encourages its managers to do likewise, though it recognizes that they make the final investment decisions.

A single poor fund does not automatically end a relationship, he says; he wants managers to understand and learn from mistakes. Persistent poor performance, however, can lead VenCap to stop investing with them.

On management fees and carried interest, Clark focuses on net returns and would prefer performance-linked carry tiers. He accepts that sought-after managers have negotiating power. He also monitors whether fund size leaves a plausible path for an early-stage investment to return the fund.

VenCap targets roughly equal allocations to early and growth funds, with some flexibility. Clark reports that, for its core managers since 2005, aggregate early-stage, growth and non-US fund TVPI had been similar, while growth funds had distributed sooner. These are portfolio-specific historical observations.

He remains cautious about overallocating to growth because entry valuations and exit prices may make later-stage investments more sensitive to market conditions. Nonetheless, he says companies staying private longer can create opportunities for strong later-stage managers.

At the time of the interview, Clark expected further losses and valuation adjustments in existing portfolios. He says historical early-stage loss ratios in VenCap’s data were around 60% of companies failing to return cost and expects recent vintages to move back toward that level. This is a prediction, not a completed accounting.

Stebbings notes that different investors can carry the same company at widely different values. Both describe the difficulty LPs face in interpreting private marks; Clark says some established managers had written holdings down earlier than some newer firms in his experience.

Clark sees parts of late-stage venture becoming a capital-allocation business where competition can compress returns. He believes earlier stages retain more room for judgment about founders and markets. This is his interpretation of the industry, not a forecast of any particular fund.

VenCap reviews investment decisions several years later. One change was to seek references not only from investors who worked alongside a manager, but also from relevant peers who did not. Clark says such comments need triangulation and context because a single source can be biased.

He has become more open to selective direct co-investments in established companies, although he remains cautious about LPs’ ability to access the exceptional few. He also wants broader access to venture for investors and a wider range of people entering the profession.

Stebbings challenges the tension between that goal and VenCap’s selective approach to new managers. Clark says the firm must safeguard clients’ capital, while describing placements and other initiatives intended to broaden career access. The exchange leaves the tension visible.

Clark recalls missing an early Benchmark investment when a pension client would commit $4 million rather than the $5 million minimum. He says UK institutions could benefit from venture exposure but worries that geographic restrictions could prioritize local managers over performance.

Looking ahead, he wants younger colleagues to have a path to lead VenCap. He would be happy to support them if useful, but believes they should drive the firm’s future. Both participants close by appreciating a discussion in which they did not agree on everything.