Venture Capital Is Shrinking. Here’s What That Means For You.

The fundraising collapse, the zombie VC problem, and what founders should do about it.

According to the Q1 2026 PitchBook-NVCA Venture Monitor, US venture capital fundraising has fallen 70% from its peak. In 2022, venture funds raised $223.5 billion across 1,793 funds. In 2025, they raised $66.7 billion across 626 funds. First-time fund formation has been hit even harder: only 106 first-time funds closed in 2025, down 78% from 478 in 2022. The 2026 NVCA Yearbook reports that the number of VC firms in existence fell to 2,984 in 2025 from 3,054–the first outright decline in that figure ever recorded.

And don’t forget inflation. Prices have risen roughly 18% since the start of 2022, so the real figures look meaningfully worse.

Things could turn around. But this is what it looks like when an asset class has become broadly disfavored by investors.

OK that’s bad, you might say, but the headline deal numbers look fine. According to PitchBook-NVCA, US VC deal value rebounded to $321.6 billion in 2025, close to the 2021 record of $358.1 billion. And Q1 2026 alone posted $267.2 billion in deal value. So if deals are getting done, how can VC be shrinking?

Because the headline numbers are a mirage. The Venture Monitor reports that the top 5 deals represented 73.2% of all quarterly deal value. The top 5 exits represented 86.6% of exit value. This is not broad-based prosperity. The market is being propped up by a tiny number of enormous transactions, mostly in AI. Strip those out and you’re looking at a much smaller, much quieter market for the vast majority of startups.

The same pattern shows up in VC fundraising from limited partners. In Q1 2026, the median fund size fell to $15.3 million while the average jumped to $289.7 million. That combination (falling median, soaring average) is a classic sign that a handful of megafunds are holding up the entire market.

The concentration is now extreme. According to the NVCA Yearbook, the top 10 funds captured 32.9% of all VC capital in 2025, up 2.5x from 13% in 2021. In Q1 2026, 73.1%(!) of committed capital went to just 5 firms. The Yearbook says 2025 fundraising was the lowest in nine years. There’s no other way to interpret this other than this is a market in which a small number of elite franchises are absorbing an increasing share of a shrinking pie, while most of the industry slowly dies.

Why the money isn’t recycling

I wrote in early 2024 about why VCs weren’t deploying their record dry powder. My main point was that VCs don’t actually have the money–their LPs do. And the LPs weren’t eager to have it called because they could earn 5%+ risk-free in Treasuries. Dry powder is a contractual commitment, not cash in the VC’s bank account. If LPs hint that they’d prefer the money stay where it is, VCs tend to listen, because crossing your LPs is a good way to never raise another fund.

The new data adds a dimension I didn’t have then, and it makes the picture considerably worse. How so?

According to Cambridge Associates’ US VC benchmark data, from 2022 to 2024, even as the overall investment pace slowed from the 2021 peak, VCs called 1.5x more capital from LPs than they returned. In 2023, managers called $30.3 billion and returned only $19.4 billion. In 2024, they called $46 billion and distributed $27 billion. It’s not only that LPs are choosing not to recommit. Many cannot recommit because the money hasn’t come back.

This is the self-reinforcing cycle that the NVCA Yearbook describes explicitly: companies stay private, exits don’t happen, LPs don’t receive distributions, LPs can’t recommit, fundraising collapses, and fewer new funds get raised.

Consider the exit numbers from the Venture Monitor. Total exit value collapsed from $865.1 billion in 2021 to $151.2 billion in 2022 and $117 billion in 2023. It recovered to $286.9 billion in 2025, which sounds better until you realize that a handful of mega-exits dominated the total. The broad middle of the portfolio–the Series B software company, the growth-stage consumer business–is not getting liquid. IPO activity remains a fraction of what it was: roughly 50 VC-backed IPOs in 2025, compared to over 300 in 2021.

There was also a denominator effect that compounded the problem. As PitchBook has explained, when public markets fell in 2022, institutional portfolios suddenly looked overallocated to private assets–not because private allocations grew, but because the public side of the portfolio shrank (and let’s just say that private capital is slower to mark down its assets). Even after public markets recovered, many allocators still hadn’t received enough capital back from their private commitments to feel comfortable making new ones. The result is that even LPs who want to recommit to venture are constrained by portfolio mechanics.

VC has been disrupted at the top

While the broad middle of the VC industry is slowly dying, something equally important is happening at the high end: VC has been made irrelevant for the most important companies of this era.

With some exceptions, the biggest AI companies–OpenAI, Anthropic, xAI–are not primarily being funded by venture capitalists in any traditional sense. The anchor capital is coming from public market hyperscalers and a small number of very large strategic investors. Microsoft has invested up to $13 billion in OpenAI, which recently raised $122 billion at an $852 billion valuation. Google and Amazon have made similarly massive commitments to Anthropic. These are not venture-scale checks, even for the largest growth funds. Very few, if any, VC funds can write $13 billion into a single company.

The traditional VC model goes like this: invest early, add value, ride the company to an IPO or acquisition, return capital to LPs. That model works when the generational winners need tens of millions or even low hundreds of millions to reach profitability. Google raised roughly $25 million before its IPO. Facebook raised $2.4 billion. Those are numbers a large VC fund can participate in meaningfully.

But when the generational winners need $50-100 billion before they’re profitable–as the frontier AI companies appear to–the primary entities with the balance sheet to fund them are the Mag 7 tech companies and a small set of sovereign and strategic capital pools. And those companies aren’t doing it as financial investors. They’re doing it as strategic investors who integrate the AI models into their own products. Microsoft doesn’t need OpenAI to IPO to get a return. It gets its return through Azure and Copilot.

So VC is being squeezed from both ends simultaneously. From below, the fundraising collapse is killing the broad middle of the industry–the first-time fund managers, the generalist funds, the firms that raised once in 2021 and won’t raise again. From above, the companies that would historically be the generational VC winners–the Googles and Facebooks of this cycle–are being capitalized by big tech, not Sand Hill Road. VCs are increasingly left with the middle: important enough to need institutional capital, but not important enough to attract hyperscaler balance sheets.

The zombie problem

What does this mean practically for founders and startups?

There are roughly 3,000 VC firms in the US. Given the fundraising data above, a significant number of these firms are, for all practical purposes, zombies. Their existing fund is mostly or fully deployed. Their LPs have not given them permission to raise again, or have quietly signaled they won’t commit to the next fund. They may have some reserves for follow-on investments in existing portfolio companies, but they cannot lead a new deal.

Here’s the part that founders need to understand: in my experience, many of these zombie VCs are still taking meetings. The junior partners especially have strong incentives to keep showing up. They need to stay current on deal flow to remain employable. They need to maintain their network. They need, frankly, to look productive. If you’re a junior VC at a firm that can’t invest, what exactly are you supposed to do all day?

So they’ll take your pitch. They’ll ask smart questions. They’ll give you thoughtful feedback. They might tell you they’re interested and ask for a follow-up meeting. But when it comes time to actually lead a round where actual money enters your bank account, they can’t do it.

This is an enormous waste of founder time, and I think it’s happening way more than people realize. Founders should be diligencing their VCs at least as hard as VCs diligence them. Tactfully, ask when the firm last closed a fund (you can also look this up). Ask how much dry powder they have. Ask when they last led a round and at what size. If the answers are vague or the last fund closed in 2021, you are quite possibly talking to a zombie.