American venture capital has never seen a six-month stretch like this. US investors deployed a record-shattering $412.7 billion in the first half of 2026 ā nearly 30% more than they put to work in all of 2025 ā according to the Q2 PitchBook-NVCA Venture Monitor. The even more remarkable figure sits inside that total: AI companies absorbed $355.9 billion of it, roughly 86 cents of every venture dollar spent, confirming that the US startup economy has effectively become an AI economy.
A Market of Megadeals
The headline number masks extreme concentration at the top. Megadeals of $100 million or more captured 87.5% of all capital deployed in the half, and the second quarter alone saw seven rounds above $1 billion close for a combined $87.2 billion ā five of them AI companies. The largest was Anthropic's $65 billion round, which lifted the company's post-money valuation to $965 billion. Together, OpenAI's $122 billion raise and Anthropic's cumulative $95.6 billion accounted for a stunning 43% of all global startup funding.
Crunchbase's independent tally corroborates the trend, putting global startup funding at a record $510 billion for the first half ā already ahead of the $440 billion raised worldwide in all of 2025.
The concentration extends to who is writing the checks. Three firms ā Andreessen Horowitz, Thrive Capital, and Founders Fund ā took in 48.1% of all capital raised by venture funds in the period. Zoom out slightly and the pattern holds: the five largest venture managers captured 73.1% of fund capital, and the top 15 captured 88.5%.
Boom at the Top, Squeeze Underneath
Fortune's analysis of the data carries a pointed warning: almost none of the record capital is trickling down. Beneath the mega-rounds, the market is contracting in nearly every segment. First-time fund formation is on pace for its weakest year since 2016, meaning fewer new managers ā and by extension fewer new theses, geographies, and founder networks ā are entering the ecosystem even as headline totals smash records.
The exit market shows the same barbell shape. SpaceX's $1.7 trillion IPO in the second quarter raised $75 billion and generated more value than every US venture-backed exit of the previous decade combined. Its $250 billion acquisition of xAI in the first quarter was the largest purchase of a venture-backed company on record, and a $60 billion all-stock deal for AI coding startup Cursor is pending. PitchBook notes that both Anthropic and OpenAI have filed confidentially to go public ā transactions the firm expects to produce two more trillion-dollar exits and reshape venture return math for a generation of funds.
Where the AI Money Is Flowing
Beyond the frontier labs, mid-July brought a steady drumbeat of substantial rounds across the AI stack:
- Neko Health raised a $700 million Series C for its AI-powered preventive health diagnostics platform
- Emergent, an AI coding platform out of India, hit a $1.5 billion valuation with a $130 million Series C
- Taktile landed $110 million from Goldman Sachs Alternatives for its agentic decision platform serving banks and insurers
- Radical Numerics closed a $50 million seed round ā a size that would have counted as a late-stage round only a few years ago
Sector data shows AI agent startups alone attracted $1.8 billion across more than a dozen July deals, with median post-money valuations up 40% from the first quarter. Crunchbase data also shows nearly 88% of global AI startup funding this year has gone to US-based companies, underscoring how geographically lopsided the boom has become.
Why It Matters
The first half of 2026 settles one debate and opens another. Venture capital is unambiguously all-in on AI ā but the structure of that bet is historically narrow. When 86% of capital flows to one sector, when three firms gather nearly half of new fund commitments, and when two companies account for 43% of global funding, the ecosystem's fortunes become tightly coupled to a handful of balance sheets and a single technology thesis. For founders outside AI, fundraising math has become brutal; for enterprises, the flood of capital means an unprecedented wave of well-funded AI vendors competing for their budgets. And for limited partners, the looming OpenAI and Anthropic IPOs will be the first true stress test of whether paper markups at trillion-dollar scale can convert into distributed returns.
The second half will reveal whether this is concentration before consolidation ā or the top of an extraordinary cycle.
