TLDRocket
Sign in

87.5% of venture dollars went to AI. The rest fought over scraps

Fortune Allie Garfinkle

VC money is flowing hard into AI: 87.5% of U.S. venture dollars in H1 went there. Everyone else is stuck with tiny valuation bumps, or big markdowns on secondaries.

Based on reporting by Fortune, Allie Garfinkle — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

PitchBook’s latest U.S. venture data makes the market feel brutally simple: if a company is AI, investors will stretch. If it isn’t, they usually won’t. In the first half of this year, AI megadeals absorbed 87.5% of all U.S. venture dollars. That is not a typo-sized edge. It is the market.

The valuation gap shows up everywhere PitchBook looked. Non-AI companies saw median valuation step-ups of 1.6x. AI companies got 2.2x. By Series D and later, the spread becomes almost absurd: AI step-ups hit 6.6x. PitchBook senior research analyst Emily Zheng said the median velocity of value creation at that stage jumped from $108.9 million in 2025 to over $1 billion in 2026, and pointed to Anthropic as one example, saying its valuation rose 5.3x in eight months.

That kind of paper wealth does not make exits any easier. IPOs are still more promise than proof, with PitchBook saying only SpaceX and Cerebras have offered much evidence that going public is worth the trouble. Acquisitions have been more active, with 2026 deal value reaching $375.4 billion so far, a decade high, and valuations at 1.9x versus 1.2x last year. But the headline number hides the mess underneath. ServiceNow paid $7.8 billion for Armis, above its prior $6.1 billion valuation. Capital One’s $5.2 billion deal for Brex was a hard cut from Brex’s $12.3 billion peak.

Secondary trading tells the same story without the conference polish. On Forge, startups that raised this year or last were trading around a zero to 5% discount. Companies that last raised in 2021 or 2022 were at median discounts of 54% and 59%. Zheng’s blunt read was that companies unable to raise on strong terms right now usually are not raising at all. The market is not closed. It is just very selective, and very fond of AI.

My take — AI-written commentary, not fact-checked reporting

This is what a hype cycle looks like when it grows teeth: capital stops pretending to be patient and starts acting like a searchlight with tunnel vision. The inconvenient part is that a lot of decent companies will get starved simply because they don’t wear the right label. Venture loves to call that discipline; it mostly looks like crowd behavior with a spreadsheet.

Read more about this at: Fortune

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.