Capital Calls and Concentration Risk: Limited Partners Navigate the AI Funding Surge

The Compression of Time

Limited partners - the institutional investors, endowments, and foundations that supply capital to venture funds - are recalibrating their expectations around deployment velocity. What was once a 24 to 36-month investment cycle for many funds has contracted sharply, with some managers now returning for fresh capital after just twelve months.

The shift reflects a structural change in round sizes and follow-on mechanics. When a single portfolio company raises billions in a financing round and a fund allocates 20 per cent of its capital to that opportunity, pacing models built for gradual deployment break down. Seed rounds, too, have roughly doubled in size over the past two years, forcing funds that lead rounds to write proportionally larger cheques from day one.

Beezer Clarkson, a partner at LGT Capital Partners in Zurich, describes the current environment as moving toward one-year pacing for some funds - a tempo that historically correlates with lower vintage performance. "It's typically harder to do a one-year vintage and get it right," she notes, pointing to the compressed decision windows and reduced time for market feedback.

Yet the anxiety around pacing is tempered by what limited partners are seeing on the ground. Unlike previous periods of accelerated deployment - particularly the 2021 vintage, when capital flowed into early-stage crypto projects with no revenue - current investments are often going into companies scaling from zero to $100 million in annualised revenue within a year or two. That kind of growth trajectory was functionally impossible in earlier technology cycles.

Concentration as Feature, Not Bug

Across the portfolios of institutional investors, the same names now appear repeatedly: OpenAI, Anthropic, Databricks, SpaceX, Stripe. This concentration, normally a red flag in portfolio construction, is being reframed by limited partners as a natural expression of venture's power law dynamics.

Amit Bhatti, a partner at TrueBridge in Chapel Hill, North Carolina, argues that consolidation into a handful of breakout companies is exactly what mature venture portfolios should exhibit. "If we were getting to the maturity of one of our funds and there were still lots of tiny positions, something's gone wrong," he explains. In funds that are a decade or older, the majority of unrealised value has typically coalesced into two or three holdings.

What differentiates the current moment is the speed at which this consolidation has occurred. Normally, it takes seven to ten years for a fund's winners to emerge and dominate the portfolio. In the case of OpenAI and Anthropic, valuations have appreciated so rapidly that funds just two or three years old already show extreme concentration in those positions.

The diversification challenge is compounded by the fact that nearly every venture fund - regardless of stage focus or sector thesis - has found a way to participate in the same handful of artificial intelligence companies. The traditional venture norm of avoiding competitive conflicts within a single fund has been set aside, and limited partners have largely acquiesced, betting that the upside justifies the correlated risk.

At the early stage, Clarkson notes, there remains more diversity in opportunity sets, with smaller cheque sizes and a broader range of companies. But as funds scale and follow their winners, capital aggregates into the same names across the entire limited partner portfolio.

The Mid-Cycle Upsize

Another departure from historical norms: funds are increasing their committed capital mid-deployment, rather than waiting to raise a successor vehicle. This practice, once rare, has become more common as round sizes balloon and managers seek to maintain or increase their ownership stakes in breakout companies.

The mid-cycle upsize allows a fund to deploy more capital without the optics of launching a new fund on an accelerated timeline. For limited partners, it creates logistical strain. Most institutions build multi-year commitment calendars to manage cash flow and maintain diversification across vintages. When a fund manager moves off that schedule, it forces a reallocation of capital that may have been earmarked for other managers or asset classes.

Limited partners face a difficult choice in these situations. Declining to participate risks signalling a lack of confidence in the manager, which can damage the relationship and make it harder to secure allocations in future funds. Yet participating on an accelerated timeline compresses the institution's own pacing and increases exposure to a single vintage.

The typical compromise, according to both Clarkson and Bhatti, is to reduce the size of the commitment rather than exit altogether. In 2021, when many funds came back early, limited partners would offer half the capital they had originally planned, adjusting for the shortened deployment period.

Liquidity Beyond the IPO

The traditional venture liquidity model - a seven-to-ten-year hold culminating in an initial public offering - has been under strain for years as companies delay going public. Secondary markets have emerged as a third pillar of exit options, alongside IPOs and mergers.

Bhatti describes secondaries as "the third leg of liquidity now," noting that the practice has shed much of its earlier stigma. A decade ago, selling a position in the secondary market was often interpreted as a distress signal. Today, it is used strategically by managers who want to return capital to limited partners without waiting for a public listing, or who wish to de-risk a concentrated position while retaining exposure to upside.

Private equity has long had a well-developed secondary market; venture capital is catching up. Limited partners themselves are now more willing to sell stakes in funds or strips of portfolios, and the taboo around doing so has diminished. The 2021 vintage, in particular, left many investors holding illiquid positions at high markups, reinforcing the lesson that partial liquidity can be prudent risk management.

Secondary activity is concentrated in the top handful of companies - precisely the names that dominate portfolios today. For a fund holding a large position in a company valued at tens of billions, selling 20 per cent of the stake can return the entire fund to limited partners while preserving the majority of the upside. That calculus has made secondaries more attractive even to managers who remain bullish on the underlying company.

Founder and employee secondary sales, including tender offers, have also proliferated. For companies that have been private for a decade or more - Stripe and Databricks are frequently cited examples - offering partial liquidity to employees is widely seen as reasonable. But when early-stage companies conduct tenders soon after raising capital, limited partners express more caution.

The concern, Bhatti notes, is motivation. If employees can access liquidity quickly, their incentive to remain with the company through difficult periods weakens. The analogy to cryptocurrency, where token-based compensation created immediate liquidity and contributed to high attrition, is often invoked. Lead investors, in some cases, have pushed back on employee tenders at companies they consider too early or too fragile to absorb the turnover risk.

Yet enforcement depends on leverage. When a founder has exceptional bargaining power - because the company is growing explosively, or because competing investors are eager to participate - objections from existing investors carry less weight.

The 2021 Vintage and the Yellow Light

Five years into the 2021 vintage, limited partners are beginning to form judgements about funds raised during the pandemic. The consensus is cautious. Valuations during that period were extreme, and many investments were made at markups that now appear difficult to justify.

Clarkson argues that the 2021 vintage demonstrates the risks of one-year pacing. Funds that deployed quickly in 2021 had less time to evaluate opportunities and were more likely to overpay. Some will be rescued by exposure to artificial intelligence companies - OpenAI had already been founded by that point, and Cursor began development earlier than many realise - but the overall vintage is not expected to rank among the best of the past 15 years.

Liquidity, however, has begun to flow. SpaceX remains the most prominent example, but Cerebras completed a significant exit this year, and mergers and acquisitions have resumed at scale. Bhatti points to $10 billion-plus acquisitions and at least one transaction valued above $60 billion. The IPO market, while not yet fully reopened, is no longer the only path to realising gains.

The current environment, both limited partners emphasise, is not a red light - but it is a yellow one. The pace of fund deployment, the concentration of capital in a small number of companies, and the compression of vintage cycles all introduce risks that were less pronounced in earlier periods. Yet the underlying companies are often showing revenue growth and product-market fit at speeds that were unimaginable a decade ago, which provides some justification for the aggressive pacing.

Emerging Managers in a Concentrated Market

As a handful of established venture funds capture the majority of committed capital from limited partners, the question of access for emerging managers has sharpened. Both Clarkson and Bhatti report that first-time and early-stage funds continue to be raised, but the bar has risen and fund sizes have contracted.

A differentiated thesis and a clear reason for being are now prerequisites. Fundraising has become harder, but limited partners still allocate to funds one through three, particularly when the manager can demonstrate sector expertise, network access, or a novel investment approach. The typical first fund today ranges from $25 million to $50 million, down from the $100 million-plus vehicles that some emerging managers attempted to raise in 2021.

Bhatti advises emerging managers to start small and build a track record. "It's much better for an emerging manager to build the track record, prove out the thesis, your access and your judgement with whatever capital you have, and then fund two will be easier to raise," he says. Attempting to raise a large debut fund creates expectations that are difficult to meet and increases the risk of a failed fundraise.

Even in San Francisco, where housing costs have surged alongside tech valuations, new managers are raising $5 million to $10 million funds, often formalising years of angel investing. Many well-known venture firms began with funds of similar size, and the model remains viable for those willing to operate at smaller scale.

The density effect in the San Francisco Bay Area continues to exert a gravitational pull, according to Clarkson. While some activity has shifted to New York and other hubs, the concentration of talent, capital, and operational expertise in the Bay Area has proven difficult to replicate. The question of whether soaring housing costs will eventually price out the next generation of founders remains open, but historical precedent suggests that the ecosystem adapts - perhaps by shifting activity back to the suburbs, as it did in earlier eras when Y Combinator was based in Mountain View.

Capital Calls Without Distributions

One structural pressure facing limited partners is the mismatch between capital calls and distributions. When venture funds deploy capital quickly but exits remain years away, institutions must meet their commitments without receiving corresponding cash back. This can strain operating budgets, particularly for endowments, hospitals, and foundations that rely on distributions to fund current expenses.

The risk is that limited partners rotate out of venture altogether, either because they cannot meet the cash flow demands or because they lose confidence in the asset class. That dynamic, if it accelerates, could reduce the pool of capital available to venture funds and force a recalibration of fund sizes and pacing.

For now, limited partners appear willing to tolerate the strain, particularly given the paper returns showing in portfolios with exposure to artificial intelligence companies. But the patience is conditional. If distributions do not begin to flow in meaningful volume over the next few years - whether through IPOs, acquisitions, or secondaries - the pressure on institutions will intensify.

The Next Wave

As OpenAI and Anthropic mature and employees begin to access liquidity, either through secondary tenders or eventual public listings, a new generation of capital allocators and company builders will emerge. Limited partners have seen this pattern before, most notably after Google's IPO, when a cohort of newly wealthy employees went on to become angels, limited partners, and general partners in their own right.

The wealth creation from the current wave of artificial intelligence companies is likely to be larger in absolute terms, which suggests a corresponding increase in the number of individuals with the capital and networks to participate in the venture ecosystem. Whether they become competitors to existing funds or partners depends on their appetite for the long-term, illiquid commitment structure that venture requires.

Clarkson notes that being a limited partner is not simply a matter of having capital - it requires signing up for a multi-decade business and navigating the operational and strategic complexities of portfolio construction. Many newly liquid employees will choose to angel invest or start companies rather than commit to institutional venture investing.

But some will make the leap, and when they do, they will bring the networks, product intuition, and technical fluency that come from having built or worked inside the breakout companies of this era. That infusion of talent and capital could reshape the emerging manager landscape, much as earlier waves did in the decades following the IPOs of Google, Facebook, and other defining companies of their time.