VC Just Had Its Biggest Half Ever. Two Labs Took 43% of It

Global venture funding just posted its best half-year on record. According to Crunchbase, startups worldwide raised $510 billion in the first half of 2026 — a figure that, remarkably, exceeds the $440 billion invested in all of 2025. Six months of 2026 outran twelve months of the year before. On the surface, it's a story of a market on fire. Read one layer down, it's a story about two companies.

The headline number hides the real one

Of that $510 billion, OpenAI and Anthropic alone accounted for $217 billion — 43% of all startup funding in the entire period. Anthropic raised $65 billion in Q2 by itself. Two labs, out of the tens of thousands of companies that raised money this half, captured well over four-tenths of the world's venture capital. That is not a healthy distribution curve; it's a spike with a long, thin tail.

The AI concentration extends past those two names. More than 70% of Q2's global startup capital went to AI-focused companies, up from just under 50% a year earlier. In the space of twelve months, "AI startup" went from being half the venture market to being roughly three-quarters of it. Non-AI founders raising in mid-2026 are competing for a shrinking slice of a record pie — the pie got bigger, but their wedge got thinner.

A tale of two quarters

The half was front-loaded in a way that matters. Q1 pulled in $305 billion; Q2 pulled in $205 billion. That's still an enormous quarter by any historical standard, but the sequential drop of $100 billion is worth sitting with. The record wasn't a steady climb — it was a Q1 that shattered every prior benchmark, followed by a Q2 that, while huge, came back to earth somewhat. When a market's all-time-high half is built on a first quarter that the second quarter can't match, the trend line inside the record is pointing the other way.

The Q2 detail sharpens the concentration story further. Sixteen companies raised billion-dollar rounds in the quarter, totaling $108.6 billion — about 53% of all second-quarter funding. Sixteen deals took more than half the money. For the thousands of other companies that raised in Q2, the reality was the opposite of the headline: they were splitting the leftovers after a handful of megadeals ate the buffet.

The exit window finally opened

If the funding side is lopsided, the exit side is the genuinely bullish surprise — and the part that makes this cycle different from the paper-valuation booms of the past. Twenty-four companies were acquired at $1 billion or more in Q2, totaling $113 billion in value — the highest quarter for billion-dollar acquisitions on record. Liquidity, the thing the venture market has been starved of for years, came roaring back.

The two anchor events tell you how big the tickets have gotten. SpaceX went public at a $1.77 trillion valuation, raising $75 billion — the largest IPO in history. And less than a week later, it confirmed its intent to acquire Anysphere, the maker of the AI coding tool Cursor, for $60 billion. An IPO that resets the record books and a $60 billion acquisition of a coding-tools startup, inside the same fortnight, from the same company. The money isn't just being raised at unprecedented scale — it's being recycled at unprecedented scale, from public markets back into private AI assets, almost immediately.

Concentration is the risk hiding in the record

Records feel like good news, and for founders with an AI story and a hot round, this is the best fundraising environment in history. But the shape of the money should unsettle anyone reading past the top line. When two companies take 43% of a market, the market's fortunes are effectively pinned to two balance sheets. If OpenAI or Anthropic stumbles — on a product cycle, a safety incident, a capex overrun, a regulatory turn — the reverberation isn't confined to one cap table. It reprices the entire asset class that the last two years have been built around, and it does so for every fund that marked its portfolio to the frontier-lab comparables.

There's a subtler cost, too. Capital this concentrated distorts everything downstream. Talent, compute allocation, media attention, and follow-on funding all flow toward the two gravity wells, which makes it harder — not easier — for the next genuinely novel company to reach escape velocity. A market where 43% of the money goes to two incumbents is a market that has, at least temporarily, stopped optimizing for the next OpenAI and started optimizing for the current one.

The $510 billion is real, and so is the exit window that finally cracked open after years of drought — that combination is the healthiest thing to happen to venture in a while. But the record isn't the story. The story is that the biggest half in the history of startup investing was, to a degree the industry has never seen, a bet on two companies. Everyone else raised in the margins of that bet.