Crunchbase's H1 2026 funding report puts global venture capital at a record $510 billion for the first six months of the year -- already well ahead of full-year 2025's $425 billion total. The headline is unambiguously bullish for venture as an asset class. The composition underneath it is more complicated: OpenAI and Anthropic alone accounted for $217 billion, or 43%, of that entire global total.
That means two companies -- both AI labs racing toward IPOs, both burning enormous amounts of capital on compute -- represent nearly half of all the venture and growth capital raised by every startup on Earth in the first half of 2026. The remaining 57%, roughly $293 billion, is spread across every other sector, every other geography, and every other stage of company, from pre-seed to late-stage growth rounds.
This concentration is a meaningfully different story than 'AI funding is booming broadly.' It's closer to 'two frontier AI labs have capital needs so large they are distorting the denominator of every global venture statistic,' which is a very different signal for allocators trying to read broader market health from the H1 2026 headline number.
“Strip those two companies out, and H1 2026's roughly $293 billion in remaining global funding is a healthy, but far less extraordinary, number relative to recent years.”
Context helps here: OpenAI closed a round at an $852 billion valuation in Q1, and Anthropic raised $65 billion at a near-$1 trillion valuation in May -- both a function of the sheer compute costs required to stay competitive at the frontier, not evidence that early-stage or mid-stage venture funding broadly has doubled. Strip those two companies out, and H1 2026's roughly $293 billion in remaining global funding is a healthy, but far less extraordinary, number relative to recent years.
For LPs benchmarking venture as an asset class, this is an important caveat on any 2026 vintage-year comparison: returns and deployment pace data that includes OpenAI- or Anthropic-adjacent funds (or funds that got allocation in either company) will look structurally different from funds with no exposure to either name, regardless of how well either fund picked its other portfolio companies.
For founders outside the frontier-AI-lab category, the practical takeaway is that the 'record year for venture' headline doesn't necessarily reflect more capital chasing your specific stage or sector -- competition for the remaining 57% may feel just as tight, or tighter, than in prior years, even as aggregate statistics look historic.
The bear case against reading too much into this: concentration in a handful of category-defining companies isn't new to venture -- previous cycles saw similar concentration in companies like Uber, WeWork or SpaceX at various points -- and outsized capital needs at the frontier don't necessarily crowd out capital for everyone else, since much of OpenAI and Anthropic's capital comes from strategic and sovereign sources unlikely to otherwise flow into early-stage venture funds anyway.
What to watch: whether H2 2026 shows continued concentration in OpenAI/Anthropic-scale rounds as both approach potential IPOs, or whether capital broadens out across more companies as frontier-lab fundraising needs are increasingly met by public and strategic capital instead of private venture rounds.