US Venture Capital Hit $19.27 Billion in June 2026 as AI Infrastructure Captured 60% of All Funding

US venture capital activity totaled $19.27 billion across 429 companies in June 2026, according to AlleyWatch’s monthly funding report. AI-focused startups captured 59.8% of all capital deployed, with inference infrastructure emerging as the month’s defining investment thesis.

Inference as Capital Category

Two raises anchored the month. Baseten closed a $1.5 billion Series F at a $13 billion valuation, led by Altimeter Capital, Conviction Partners, and Spark Capital. The company announced it processes more than 1 billion inference calls per day across 87 clusters on 18 clouds, with revenue growing 20x year over year. It was Baseten’s fourth fundraise in 18 months. Groq raised $650 million in the same month for its custom inference chip and serving platform.

Combined, Baseten and Groq accounted for $2.15 billion, more than 11% of total US venture activity in June, according to AlleyWatch. The simultaneity reflects a market consensus: as enterprises move AI from pilot to production and agentic systems multiply the number of model calls per task, inference capacity has become the critical bottleneck. Capital is chasing the serving layer, not the training layer.

Concentration at the Top

Late-stage capital accounted for 58.8% of national venture deployment, with a $218 million average deal size across 52 rounds. Forty-six companies raised $100 million or more. The top 10 deals alone represented 34.6% of all capital deployed nationally, per AlleyWatch.

The rest of the top 10: AppsFlyer ($1B Series E), Ramp ($750M), Supabase ($500M), Impulse Space ($500M), Flourish ($500M), Helion Energy ($465M), Ionic Digital ($400M), and Suno ($400M).

Series A Bifurcation

Despite 98 Series A deals, the stage’s $37.6 million average was heavily skewed. Eight Series A rounds exceeded $100 million, pulling the average up; the true median sat at $20 million. That gap signals a bifurcated market: a handful of high-conviction bets alongside a larger volume of conventional first institutional rounds operating in a more cautious environment, according to AlleyWatch.

Early-stage activity was healthy in deal count (248 rounds) but modest in capital at $1.83 billion, consistent with a national market that has re-priced seed and pre-seed rounds downward while reserving outsized capital for proven late-stage platforms.

New York’s Record Share

New York startups raised $4.70 billion in June, capturing 24.4% of US venture dollars. AlleyWatch noted this was the city’s highest national share in at least a year. California maintained structural dominance at 47.9%.

The Inference Economics Question

The concentration of capital in inference infrastructure has direct implications for teams building on top of these platforms. Baseten’s pitch to customers centers on post-training and optimization of specialized models served with low latency at production scale. As co-founder Tuhin Srivastava wrote in the Series F announcement, “open-weight models have become strong enough that enterprises can now use them as serious alternatives to closed APIs.”

That framing positions inference serving as the enabling layer for enterprises that want to own their intelligence stack rather than rent it from frontier API providers. Whether the $2.15 billion bet on inference infrastructure produces returns depends on whether enterprises actually follow through on that migration at scale, or whether closed API convenience continues to win on ease of integration.