Venture Capital Funding Falls to $159B as AI Mega-Rounds Dominate

A pullback that isn't quite a retreat

Global venture funding fell to $159 billion in the third quarter of 2026, according to Crunchbase data. That is roughly 25% below the second quarter's $212 billion, but still up 53% from a year earlier.2 The quarter-over-quarter decline makes a tidy headline. Underneath it, the market looks more concentrated than ever.

AI startups took 64% of the quarter's dollars, about $102 billion.2 A record 27 companies closed rounds of $1 billion or more, and those 27 accounted for roughly a third of all capital deployed.2 At the very top, the market showed no sign of cooling. Databricks and Safe Superintelligence each raised $5 billion.2

North America followed the same pattern. The region drew $92 billion in Q3, down 35% from the prior quarter but 50% higher year over year. AI companies absorbed about $61 billion of that, or roughly two-thirds.3 Crunchbase attributes much of the sequential drop to the absence of fresh, outsized financings for OpenAI or Anthropic. Their rounds had inflated the first half to record levels.3 Deal counts at both early and late stages stayed close to Q2 levels.3 That suggests the dollar decline reflects a few missing mega-deals, not founders across the board suddenly finding doors closed.

Concentration is the real story

The numbers look different depending on who is measuring, but they all point the same way. Crunchbase puts AI at 64% of global Q3 dollars.2 PitchBook-NVCA data cited in industry newsletters shows AI accounting for 82.7% of US venture deal value through the first three quarters of 2026, with total US deal value reaching $515.8 billion.4 Earlier, PitchBook-NVCA figures had AI capturing 88.8% of deal value in Q1 2026 while making up only 42.5% of deal count.1

The gap between Crunchbase and PitchBook likely comes from differences in geography, timeframe and how each firm classifies an "AI company." Neither should be read as the definitive figure. The trend they share is what matters: a minority of deals is carrying most of the capital.

The same skew shows up on the fundraising side. US funds raised $108.5 billion across 699 vehicles, and megafunds captured 78.1% of that money.4 Large funds need large checks, and that dynamic tends to feed the mega-round cycle.

What founders outside the top tier are seeing

The broad recovery metrics look healthy. Global VC reached $425 billion in 2025, a 30% increase. The median seed pre-money valuation hit $18.4 million in Q1 2026, and down rounds fell to 11.4%, roughly where they sat in 2019–20.1 But much of that rebound is explained by AI concentration rather than a broad return of capital. Many founders who benchmark against these medians are measuring themselves against a market they are not actually in.1

The concentration ratio may matter more than the headline total for companies raising a Series B or C outside the AI elite.2 If roughly a third of capital goes to a few dozen companies, the pool left for everyone else shrinks quickly, even in a quarter that looks strong on paper. In North America, the money that did flow went mostly to more mature companies, with late-stage and technology-growth businesses taking the bulk.3

AI's gravitational pull on adjacent sectors

"Non-AI" startups are not uniformly losing out. Aerospace, robotics, data centers, semiconductors and energy each cleared $10 billion in Q3.2 This suggests the AI boom is lifting the infrastructure around it rather than simply crowding it out.2 Recent deals fit that pattern. VEIR raised a $110 million Series C to scale superconducting power delivery for AI data centers, and Universal Quantum raised more than $100 million in Series A funding for modular quantum computing.4

Smaller managers are still raising money too. Graduate Ventures launched a €125 million second seed fund, Linse Capital raised $150 million for early-stage deep tech, and LAUNCHub Ventures hit a first close above €65 million for pre-seed and seed investing in Central and Southeast Europe.4

The read

Taken together, Q3 looks less like a cooling market than a widening split. The quarterly dip mostly reflects timing, because a handful of frontier labs did not raise this quarter. Deal activity held steady, and the top of the market remains intensely competitive.

For founders, the practical takeaway is that sector matters more than ever. A company tied to AI's physical or computational buildout, such as power, chips, data centers or robotics, can still find serious capital. A company outside that orbit faces a smaller share of the pie, regardless of what the aggregate numbers say. Concentration ratios, not quarterly totals, are the indicator to watch.