AI Venture Capital Is Booming While Smaller Rounds Disappear

Venture money is surging upward. Crunchbase data shows investors have poured around $48 billion into financings near the Series B stage so far in 2026, the highest tally in years. The catch is brutal: fewer deals are getting done, and the money is concentrating in companies already large enough to attract enormous checks.

Round counts sit below year-ago levels and are on track for a multiyear low. Less than a third of deals in the sample set were for under $30 million, compared with the period from 2021 through 2025, when rounds around $30 million represented about half of deals.

The smaller end of the market has nearly vanished. Only 26 rounds under $10 million appeared in the sample set, while at least 110 companies raised $100 million or more around the Series B stage—roughly a quarter of all deals.

AI Owns the Series B Conversation

AI startups captured about two-thirds of Series B funding so far this year, up from about half of investment in AI categories during 2025. More than half of all reported U.S. Series B deals in 2026 have involved AI companies, creating a bullish climate for startups with demonstrated traction and a particularly favorable one for companies that already raised substantial sums at seed and Series A.

The outliers explain the totals. Prometheus, described as Jeff Bezos’ physical AI startup, raised a $12 billion round after launching last year. World Labs, led by AI pioneer Fei-Fei Li, closed on $1 billion in February, while Valar Atomics secured $1 billion in Series B funding across tranches with Sequoia Capital as lead investor.

That leaves a clear message for founders seeking smaller Series B rounds: the market is not closed, but it is not designed for you. Investors are focusing on big checks for high-valuation companies, which makes fundraising look healthy in aggregate while making the path for ordinary companies much narrower.

Founders Face a Market Built for Extremes

At TechCrunch Disrupt, Shailendra Singh of Peak XV described the fundraising environment as one of the hardest he has seen. “Seed and venture, I’ve never seen it as dysfunctional as it is now,” Singh said, adding that companies can now try to close a round at three different prices at once.

Jahanvi Sardana of Index Ventures gave the arrangement a name: “We’re calling it the ‘tranchopalooza.’” She was referring to rounds split into separate chunks, a structure that reflects how investors are managing risk while still chasing companies that might break out.

Singh said founders who attract very high-quality talent have an advantage, but he also warned against polished metrics that conceal weak performance. “If you ever show a chart which shows cumulative revenue or cumulative ARR, that is a massive red flag,” he said, calling cumulative data “a big lie” and urging companies to present themselves with intellectual honesty during bull cycles.

His broader warning was less comfortable than the funding totals. The market is full of “false positives” because “it can’t be that every company is going to be successful.” Singh expects 2027 to look different and said, “I think we will see a correction.” His advice to founders was blunt: do not rush, watch for weak product-market fit, hold the bar high, and remember that building a company is one of the “most excruciatingly difficult things” one can do—“deal with it.”

Sardana also warned that founders may be choosing what to build based on what gets funded, calling that “really scary.” Janelle Teng Wade of Bessemer offered the opposite fundraising posture, saying, “I would say just be audacious in your ask,” then advising founders to see how the market responds and tailor their requests from there.

Neolabs Turn Capital Into a High-Stakes Bet

The same concentration is driving huge valuations for AI research companies. Jeff Dean, Google’s former chief scientist, is seeking a valuation as high as $50 billion for his new AI science startup; OpenAI cofounder Ilya Sutskever’s Safe Superintelligence is worth $32 billion; and Mira Murati is raising a round for Thinking Machines Lab at around $40 billion.

Anthropic is preparing an IPO at a valuation of around $2 trillion and has told investors it was profitable on an adjusted basis, though that measure excludes some costs. The example captures the appeal of neolabs: enormous compute and talent bills today, paired with the possibility that a rare winner becomes a foundational AI company worth hundreds of billions or more.

The economics remain fiercely disputed. A venture capitalist at a top firm called neolabs the most debated investment in the market, with some investors arguing that nobody should fund them and others saying investors should fund almost nothing else because software is dead and companies must own models or hardware—even if a company entered at a billion dollars and reached a trillion-dollar value.

Matt Murphy of Menlo Ventures called neolabs the most overhyped sector and compared Google’s success, from his time as an early board advisor, to a match thrown into a bonfire. “The opposite thing you want to do is start a match with a bonfire,” Murphy said, warning that huge rounds create exactly that risk.

Investors know most of these companies may burn enormous sums on compute and talent without becoming profitable. Still, Rosenberg said, “We’re confident that the best neolabs will attract ongoing support,” while Andrew Schoen said neolabs he knows are finalizing billion-plus-dollar equity rounds that are moving smoothly and quickly.

The scale is the point—and the warning. Only 130 U.S. tech companies, public or private, carry valuations of $10 billion or more, with roughly 20 to 30 new companies reaching that level each decade across the past five decades. Yet 42 decacorns have been minted privately this year with a quarter still remaining. As Das put it: “Like all startups, many will fail, some will work.”

Stay connected via Google News

Follow us for the latest AI News