AI Swallows 86% of US Venture Capital in H1 2026, as PitchBook Data Reveals a Two-Company, One-Trade Market
US venture capital has entered territory with no historical precedent. The second-quarter PitchBook-NVCA Venture Monitor, released on July 9 and reported on by SiliconAngle and Fortune, shows that American investors deployed $412.7 billion in the first half of 2026 — nearly 30 percent more than the entirety of 2025. Of that sum, $355.9 billion, or 86 cents of every venture dollar, went to artificial intelligence companies. Mega-rounds of $100 million or more captured 87.5 percent of everything deployed. In the second quarter alone, seven rounds exceeding $1 billion closed, five of them AI companies, totalling $87.2 billion. The headline numbers are records in every dimension. The substance underneath them is more complicated and, for most founders, considerably less encouraging.
The market is, in practice, a two-company trade. According to Crunchbase data, OpenAI and Anthropic together attracted $217 billion — 43 percent of all global startup funding in the first half of 2026. Anthropic's $65 billion round in the second quarter was the single largest of the period, pushing its post-money valuation to $965 billion, up from a $350 billion pre-money mark just three months earlier, per the PitchBook-NVCA report. OpenAI had previously closed a $122 billion Series F in March at an $852 billion valuation — the largest private funding round ever completed — with Amazon, Nvidia, and SoftBank anchoring the deal, as Forbes reported. Both companies have confidentially filed for IPOs, with PitchBook and Crunchbase analysts noting that their public offerings could collectively generate more exit value than the entire IPO market has produced in years and rank among the largest listings in history.
Below the mega-rounds, the picture bifurcates sharply. Rounds below $100 million, which still constitute the bulk of the market by deal count, drew just $51.4 billion across the entire half-year period, according to SiliconAngle's reading of the PitchBook data. First-time fund formation is tracking for its worst year since 2016, as the PitchBook-NVCA report noted. Seed-stage capital totalled $4.9 billion against late-stage's $101 billion. Three firms — Andreessen Horowitz, Thrive Capital, and Founders Fund — captured 48 percent of every dollar raised by venture funds in the half. Geographic concentration is equally stark: 68 percent of second-quarter US venture dollars landed in the San Francisco Bay Area. PitchBook director of US venture capital research Kyle Stanford told Fortune that founders seeking large capital allocations essentially need to be present in that geography to access them. The result, as Fortune concluded, is a VC ecosystem setting records at the very top while contracting in almost every other segment beneath it.
The exit market tells a similarly lopsided story. Total US venture-backed exit value reached $2.2 trillion in the first half, a figure that appears extraordinary until its composition is examined. SpaceX's initial public offering in the second quarter generated $1.7 trillion of that total, what PitchBook called the largest IPO of all time and the largest US venture-backed listing by a factor of 17. The company's $250 billion acquisition of xAI in the first quarter added further to the tally. Strip out SpaceX-related activity and quarterly exit value would sit near the constrained levels of recent years, per the SiliconAngle analysis. Cerebras Systems completed a $34.3 billion IPO after cancelling a 2025 attempt, though shares opened at more than double the offer price before sliding back. Of the ten largest US tech IPOs excluding SpaceX and Cerebras, only three traded higher a year after listing — a sobering statistic for the broader venture-backed pipeline.
For founders operating outside the frontier-AI and infrastructure layers, the H1 2026 data offers a mixed strategic signal. The sheer volume of capital flowing into AI is compressing valuations in non-AI categories while simultaneously driving up talent and compute costs that non-AI startups must pay. PitchBook characterises the shift as structural rather than cyclical, with AI coding tools lowering the cost to build software and foundation models giving founders a base layer that removes the need to train their own systems. Investors at firms including Goldman Sachs Alternatives and Tiger Global are directing capital toward agentic AI in regulated sectors — fintech decision automation, healthcare workflow, and compliance infrastructure — rather than broad-based software. Whether the pending IPOs of OpenAI and Anthropic validate the private valuations that have defined this era or trigger a repricing that cascades downward through the funding stack may be the defining question for the second half of 2026 and beyond.