AI Companies Have Captured 87.5% of the Venture Market — Everyone Else Is Fighting Over Scraps
This is an AI world, and the rest of us are just living in it.
PitchBook released its U.S. venture valuations report for the second quarter of 2026 this week, and the numbers reconfirmed a now-familiar and deeply lopsided market picture: investors are willing to pay almost any price for a handful of "chosen" AI companies, while nearly everything else gets either passed over or valued as an afterthought. In the first half of the year, 87.5% of all U.S. venture funding went to AI mega-deals.
Valuations are climbing outside AI too — it's just that the "AI premium" is far more pronounced. According to PitchBook, the median valuation of non-AI companies rose 1.6x, while AI companies saw a 2.2x increase. The gap widens sharply at Series D and later rounds, where valuation growth for AI companies reached 6.6x.
"That kind of jump at Series D+ is clear evidence of just how strongly AI is distorting venture valuations," says PitchBook senior analyst Emily Zheng. "In 2025, the pace of value creation at that stage was $108.9 million; in 2026, it topped $1 billion — nearly a tenfold increase. Leading AI companies like Anthropic are driving that growth: its valuation climbed 5.3x in just eight months. Venture returns typically follow a power law, but AI has set a new standard for abnormally high valuations."
Liquidity remains a major problem in this market. The IPO window is technically open, but it isn't exactly convincing — aside from SpaceX and Cerebras, no one has proven that going public actually pays off.
M&A activity looks encouraging at first glance: total deal value hit $375.4 billion in 2026 — the highest in a decade — with valuation multiples climbing from 1.2x last year to 1.9x, according to PitchBook. But look at individual deals and the picture gets murkier: ServiceNow's acquisition of Armis came in at $7.8 billion, a solid step up from the cybersecurity company's previous $6.1 billion valuation. On the other hand, Capital One's $5.2 billion price tag for Brex marks a steep drop from the fintech unicorn's earlier peak of $12.3 billion.
The secondary market has become an especially precise mirror of the situation: if your company is young and AI-related, demand is strong; if not, you're in trouble. According to PitchBook, on the Forge platform, startups that raised funding this year or last are trading at a median discount of just 0-5%. For startups that last raised money in 2021 or 2022, that discount climbs to 54% and 59%, respectively.
"The winners have never been this big — they're overshadowing the entire rest of the venture market," Zheng says. "Companies that can't raise on strong terms right now generally can't raise at all."
To be clear, I'm not saying every startup that raised money in 2021 is a great company left behind — that era had its own share of irrational exuberance (remember NFTs?). Still, some genuinely solid companies are simply gathering dust because they fell out of trend. What's most striking is the speed of this shift: yes, the data on AI's dominance of the venture market has looked roughly the same for a while now. But the world has actually already turned in a completely different direction.
Related articles

E-Commerce Is Growing, But Walmart's Stores Matter More Than Ever
Walmart's second-quarter results show that even as online sales grow their share, physical stores are becoming the central fulfillment hub for delivery.
Not Tomorrow, Not Even an Hour — Now Goods Arrive in Minutes
Major retailers — Target, Walmart, and Home Depot — are turning delivery within hours, or even minutes, into the new standard. What's driving the shift?
What Is an AI Persona, and How Do You Make One Actually Trustworthy?
An AI persona is a tool that models your target audience so you can question it directly. But what actually makes that persona trustworthy?