$677 billion sitting still: how committed capital piled up in the 2020-2021 fund vintages

Venture capital is living with a paradox. The pool of capital that limited partners have committed but that has not yet reached startups is very large, while the actual pace of deployment is well below its multi-year average. According to PitchBook research, there is $677 billion of global VC dry powder, and 53% of it sits in funds three to five years old - the highest build-up in that cohort since the 2008 financial crisis.

The 2020-2021 vintages and the money they never spent

PitchBook traces this dry powder back to the 2020-2021 boom, when LPs committed to venture funds at record speed. The prolonged downturn that followed left many funds with too much cash on hand and too few opportunities they were willing to price - the point made by Kyle Stanford, PitchBook's director of venture capital research, in the same report.

The deployment numbers say the most. PitchBook calculates that since 2010 venture funds have called and deployed an average of about 37% of outstanding dry powder a year. In the 12 months to September 2024 that rate fell to 18%, more than halving. In private equity, the comparable figure fell from about a third to 26% - a far shallower contraction than in venture.

Why sit on the money? PitchBook's 2026 US venture capital outlook gives a professional reason: investors hesitate to work through dry powder without meaningful markups in the portfolio, because going back to LPs for a new fund without evidence of value creation is a weak negotiating position. The same report notes that 2024 seed deal count was still 27% below 2021, even though the early stage is described as the most active part of the market.

The bottleneck is the exit door, not the fundraising door

PitchBook describes a closed loop: a lack of IPOs means few distributions back to LPs, LPs then become less enthusiastic about committing to new funds, and new fundraising slows even while old capital remains unspent. In other words, $677 billion is a sign of congestion, not of abundance.

On the deployment side, the picture is highly concentrated. Crunchbase data for 2025 shows close to 60% of all invested capital went to 629 companies that raised rounds of $100 million or more; 68 companies raising rounds of $500 million or more took more than a third of global funding, against 24% in 2024. The two largest private rounds ever recorded both occurred in 2025 and both were AI-related: OpenAI at $40 billion and Scale AI at $14.3 billion.

• Dry powder is concentrated in the 2020-2021 vintages, the cohort under pressure to show results before raising again.
• The call-down pace fell from a 37% annual average to 18% in the 12 months to September 2024.
• The money that is being spent flows into a small set of very large companies, mostly in AI, so the rest of the market feels thinner than the headline totals suggest.

What could thaw the market in 2026

PitchBook expects fundraising to increase in 2026, supported by improving liquidity and gradually recovering LP sentiment. Crunchbase notes the IPO market opened up in 2025 and argues larger listings of venture-backed companies look more likely - precisely the catalyst the loop above is waiting for. In Crunchbase's forecast survey, one investor expects global deployment in 2026 to rise from the low $400 billions to the high $400 billions, implying roughly a 10% increase, with AI still the driver.

For founders, the practical implication is that the money exists but the conditions for releasing it have changed. A fund in year three to five faces pressure both to deploy and to deploy only into deals that can show a clear markup at the next round. Those two pressures pull in opposite directions, which explains why diligence takes longer while the deals that do get chosen close very fast.

This briefing summarises market data from the sources listed below and is not investment advice.