Jumbo Series A Rounds Surge as AI Claims 70% of $100M Deals

A new class of Series A

A Series A round once meant a startup had found early traction and needed a few million dollars to prove it could scale. In 2026, a growing slice of these rounds look more like late-stage megadeals.

Crunchbase counts at least 114 Series A rounds worldwide this year at $100 million or more, the highest annual total in years and on pace to surpass the all-time peak 2. Together these rounds have raised roughly $33 billion, and at least a dozen came in at $500 million or above 2. As of September 23, the year was tracking toward a record 3.

Overwhelmingly an AI story

The surge is concentrated in one sector. More than 70% of the $100 million-plus Series A rounds went to AI-focused startups 23. The largest is a $1.2 billion raise for River AI, a Silicon Valley platform that lets developers train and serve custom models 2. China-based Xpeng Robotics, which builds AI-enabled humanoid robots, raised $900 million 23. The United States leads the jumbo category, accounting for roughly half of these deals according to Crunchbase's geographic breakdown 2.

One framing of the trend is that investors are paying early for companies they see as likely category leaders, not waiting for the usual milestones 3. That interpretation matches the valuation data. PitchBook-NVCA figures put the median pre-money valuation for AI Series A companies at $78.0 million in Q1 2026, compared with $42.4 million for non-AI companies 1. Carta data shows a wider gap for foundational-model startups, whose Series A post-money valuations run around $300 million against $55 million for non-AI peers 1.

The broader Series A market is inflating too

The jumbo rounds sit at the top of a market where typical round sizes have also grown, though benchmarks differ depending on who is counting. Carta reports a median software Series A of $14.4 million at an $80 million post-money valuation in the six months to July 2026 1. That is roughly double the $7.4 million median raised at a $48 million pre-money in Q1 2025 1. PitchBook-NVCA puts the median US Series A higher, at $19.6 million on a $62 million pre-money. That pre-money figure is nearly triple the $21 million median of 2020 1. PitchBook-NVCA's average deal size of $39.6 million is about double its median, which suggests a handful of very large rounds are pulling the mean upward 1. Crunchbase's 2025 quartiles put the US median at $15 million, with an upper quartile of $25 million 1.

These gaps likely reflect different datasets, sector mixes and time periods, not contradictions. Still, they matter for founders comparing themselves to "the market." A software founder benchmarking against Carta and an AI founder benchmarking against PitchBook could reach very different conclusions about what a reasonable round looks like.

That divergence shows up in ground-level deal tracking as well. One tracker of Series A activity logged 46 rounds totaling about $1 billion in January and 47 rounds worth about $1.45 billion in February, its peak month 4. That works out to average rounds in the low tens of millions. The same list includes modest European raises such as France's Kheops at €12 million and Madrid-based Reveni at $8.8 million 4. For most startups, the $100 million Series A remains far out of reach.

Stage labels are blurring

The jumbo trend is also blurring what funding stages mean. The same 2026 list cites MatX raising a $500 million Series B, and Fireworks AI and Baseten each raising $1.5 billion in Series D and Series F rounds respectively 4. When a Series A can top $1 billion, as River AI's did, the letter attached to a round says less about a company's maturity and more about the order in which it raised money 2.

What to make of it

The data points toward a two-tier venture market. At the top, AI companies with credible claims to infrastructure or model leadership are raising capital at the earliest priced stage on a scale once reserved for pre-IPO companies. Below that, the Series A market is healthier than it was in 2025 but looks far more ordinary.

In my reading, the jumbo-round boom reflects concentrated conviction more than broad exuberance. Investors are compressing several rounds into one to lock in ownership of perceived winners before valuations climb further. The risk is that much of this money is priced for outcomes that only a few companies can deliver. If AI category leadership proves less durable than investors expect, many 2026 Series A valuations will be hard to grow into. For now, though, the capital is flowing, and it is flowing overwhelmingly to AI.