AI Startups Raised $364 Billion This Month — Here's Where the Capital Went

Three hundred forty-seven venture capital deals closed for AI startups in the past 30 days, representing a pace that dwarfs every other technology sector. The numbers are staggering: $364 billion deployed, with average deal sizes reaching $1.5 billion. But averages obscure the story — the median round sits at $25 million, a sign that while mega-rounds make headlines, the bulk of capital is flowing into early-stage founders solving concrete problems.

What emerges from this data is not chaos, but rather a clear bifurcation. On one side, a handful of infrastructure plays and mature startups capturing generational funding. On the other, a wave of early-stage founders — seed and Series A companies — building atop the large language models and foundation models that are now, finally, stable enough to bet on.

The Seed Round Is Inflating, But Not How You'd Expect

Seed rounds and Series A rounds each account for roughly 15% of all AI funding activity this month. But here is the shift: the seed round is no longer the scrappy $2 million moment. A Seed round in AI now lands, on average, at nearly $10 million. Series A averages $233 million — a number that would have been unthinkable for enterprise software five years ago.

This shift is not arbitrary. The bar to credibility in AI has risen. Founders need to demonstrate, early, that they can actually build something that works. A chatbot prototype running on OpenAI's API is no longer sufficient. Investors expect founders to have proprietary data, proprietary models, or a defensible application that actually solves a billion-dollar problem. That requires capital — hence the inflation.

The data also shows that Pre-Seed rounds are almost a rounding error now: 27 deals in a cohort of 347. Founders with pre-product ideas are being told to come back when you have paying customers. Or, more likely, to bootstrap longer or find a technical co-founder with deeper pockets.

Vertical Platforms Are Eating the Market

Of the eight subsectors we track, AI Vertical Platforms dominate, accounting for 180 of the 347 deals. This is the category that covers industry-specific AI applications: healthcare diagnostics, legal document review, financial underwriting, supply chain optimization. Narrow, deep, defensible.

SaaS follows at 149 deals — the broadest category, capturing everything from admin tools to workflow automation. Then AI Infrastructure at 125 deals: the companies building inference engines, serving models, managing compute costs. These are the picks-and-shovels plays that will remain valuable regardless of which application companies survive the inevitable shakeout.

Notably, AI Agents captured 113 deals. This category barely existed two years ago. It represents the bet that language models can be given agency — goals, memory, access to APIs — and thus can automate entire workflows, not just parse text. Early validation is strong, though the failure rate will be brutal.

America Still Sets the Pace, But Global Capital Is Rebalancing

The United States accounted for 146 of the 347 AI deals — 42% of the total. This is still overwhelming dominance, but it masks a key shift: capital is moving outward. The United Kingdom (39 deals), China (35), France (18), and Germany (13) are all funding AI startups at scale. This is not venture tourism; these are local founders building for local markets.

France and Germany are particularly noteworthy. Both countries have implemented AI regulation (the EU AI Act) ahead of the US, which has paradoxically accelerated funding rather than chilling it. Founders believe that building to EU standards first is a competitive advantage: if you can pass Europe's gaze, you can pass anyone's.

China's 35 deals suggest the market has moved past the initial regulatory chill. Beijing has been quietly allowing AI funding to resume, particularly in applications that do not involve geopolitical sensitivities — semiconductors, inference, agent systems.

The Seed Boom and the Middle-Stage Crunch

Series C and beyond capture only 15 deals combined. This is the crunch point: founders who cannot raise growth capital are stuck. There is a wall between the $100-million-valuation round (Series B) and the multi-billion-dollar outcome. Many companies will not make it across.

Interestingly, Series G saw 12 deals — mega-rounds for companies that have already proved unit economics. These deals are for acceleration, not survival. They suggest a handful of AI companies have passed through the valley of death and are now scaling predictably. Mistral, Anthropic, Scale AI, and a few others have demonstrated that frontier models and supporting infrastructure can generate revenue.

But these winners are the exception. The 347-deal wave is a lottery. Perhaps 5% will hit a $1 billion outcome. The rest will either plateau at $50-100 million revenue, find acquirers, or shut down.

What This Pace Means for Q2

The volume of capital deployed in the past 30 days is not sustainable at current burn rates and customer acquisition costs. Expect a rotation: less capital for "AI-washed" applications (every startup is an AI startup now), more scrutiny of unit economics, and a migration of capital toward infrastructure plays and vertical applications with clear defensibility.

For founders, the message is clear: having AI in your pitch deck is no longer enough. You need to show that your AI actually solves something that was previously unsolvable, or makes something dramatically cheaper or faster. The bar has moved. The capital has not run out — but it has gotten more discerning.