Q1 2026 shattered the global VC funding record: $297 billion deployed in a single quarter — up 150% year-over-year. The headline reads like a boom. The reality inside the unicorn layer is more complicated, and significantly less healthy.
Most of that $297 billion went to a small number of AI mega-rounds. OpenAI, Anthropic, xAI, and a handful of AI infrastructure companies absorbed the majority of the capital. The rest of the startup ecosystem — the 750+ companies valued above $1 billion — is experiencing something very different.
I'd call it the quiet crisis of cap table paralysis.
What Cap Table Gridlock Actually Is
A "cap table" is a company's capitalization table — a record of who owns what, at what price, and with what rights. For early-stage startups, it's simple. For a company that has raised 6–8 rounds over 8–10 years, it becomes a governance structure that can make the company effectively ungovernable.
Here's the mechanism: each successive round adds preferred shareholders with specific rights — liquidation preferences, anti-dilution provisions, pro-rata rights, and board representation. By the time a unicorn is approaching Series F or G, the cap table may contain investors with conflicting interests that cannot be simultaneously satisfied in any single transaction.
A Series A investor with a 2x liquidation preference needs the exit price to be above a certain threshold before they participate in upside. A late-stage investor who came in at the peak valuation needs an even higher exit price to avoid writing down their position. A founder with heavily diluted equity has different incentives than any of the institutional holders.
When these interests can't be aligned — when no single sale price, merger structure, or IPO valuation satisfies enough of the table — the company simply can't move. It can't sell. It can't merge. It can't IPO. One shareholder or one share class with blocking rights holds veto power over any transaction.
That is cap table gridlock. And it's now the most common structural problem I hear about in late-stage private company discussions.
The Zombie Unicorn Count
Approximately 25% of current unicorns — companies marked at $1B+ on their last funding round — have almost certainly declined below a $1 billion valuation in real terms. They haven't marked down publicly because they don't have to. Private company valuations are self-reported, based on the last funding round price, not on an ongoing market clearing mechanism.
These are zombie unicorns. They carry the $1B+ designation on CB Insights and Crunchbase. Their investors carry the asset at cost on their fund books. But if you asked those companies to raise a new round or execute a secondary sale today, the clearing price would be well below the headline valuation.
The zombie count has grown for three reasons: the 2021 funding bubble inflated entry valuations to levels that can't be sustained, IPO windows have been effectively closed for non-AI companies since 2022, and M&A as an exit has been complicated by antitrust scrutiny and buyer balance sheet caution.
The Concentration Problem
The $297B Q1 funding record masks a distribution problem that should concern anyone allocating to the asset class.
When you strip out the ten largest AI rounds in Q1 2026, the remaining $200+ billion of venture deployment looks much less impressive relative to the number of companies competing for it. Median deal sizes at Series A and B are down. Bridge rounds — companies raising to survive rather than to grow — are up significantly.
The capital is there. It's just not reaching the companies that need it. Founders outside the AI super-cycle are experiencing a funding environment that looks nothing like the headlines suggest.
What This Means for LP Returns
For limited partners in venture funds raised in 2019–2022, the return picture is increasingly difficult. Fund life is typically 10 years. For a 2020-vintage fund now in year six, the clock is running. If the portfolio contains multiple companies in cap table gridlock — companies that can't exit through any practical mechanism — the fund's DPI (distributions to paid-in capital) stays at zero while the J-curve clock runs.
Fund managers facing this problem are increasingly looking at secondary sales of their LP positions, direct secondaries on specific portfolio company stakes, or in some cases simply waiting and extending fund life. None of these are great outcomes for LPs who expected distributions in year six through eight.
My Read on the Venture Market Right Now
The AI funding boom is real and the companies at the center of it — true AI infrastructure, model providers, and specialized application layers with genuine distribution — are worth the attention they're getting.
Everything else in the unicorn layer requires much more caution than the headline funding numbers suggest. The gridlock problem doesn't resolve on its own. It requires either a market clearing event (a wave of down rounds and restructurings that resets cap tables at lower prices), a sustained IPO window (which requires stable equity markets and investor appetite for growth stories), or continued bridge financing (which buys time but doesn't fix the underlying structure).
I'm not deploying into late-stage private companies outside the core AI infrastructure layer right now. The risk-reward is poor when the exit mechanisms are this constrained.
— Ruslan Averin, averin.com