SAN FRANCISCO: Anthropic has agreed to terms on a $30 billion funding round at a $900 billion pre-money valuation, surpassing OpenAI’s $852 billion mark and becoming the most valuable private AI company in the world. Four firms, Dragoneer Investment Group, Greenoaks Capital, Sequoia Capital, and Altimeter Capital, will co-lead the round with each lead expected to commit at least $2 billion. The deal is on track to close as soon as the end of May 2026.
The terms were first reported by Bloomberg on May 12, and the four co-lead investors were named by The Information shortly after. No term sheet has been publicly signed, but the parties have agreed on terms and the round is expected to close within days.
How Anthropic nearly tripled its valuation in three months
Anthropic closed its Series G in February 2026 at a $380 billion post-money valuation. The new round prices the company at $900 billion pre-money, an increase of more than $500 billion in roughly three months. The Series G was, at the time, the largest single AI funding round in history. The new round breaks that record by a wide margin.
The valuation jump tracks the revenue trajectory. Anthropic’s annualized revenue is projected to exceed $45 billion as of May 2026, up from roughly $4 billion a year earlier, a 10x year-over-year increase. Eight of the Fortune 10 are enterprise customers. More than 1,000 customers spend over $1 million annually on Claude, the company’s AI assistant.
Three of the four named co-leads also back OpenAI. Sequoia, Altimeter, and Dragoneer hold positions in both companies. Greenoaks has positioned itself as the Anthropic-aligned bet inside the top tier of growth equity. The willingness of the same firms to underwrite both companies at near-trillion-dollar valuations is the signal worth reading. The largest growth funds in the world are no longer picking a single winner at the foundation-model layer. They are running parallel positions in both.
What does Anthropic’s $900 billion valuation mean for AI founders?
For AI founders, the round is a clear repricing signal: enterprise revenue is now valued more highly than consumer reach, and the foundation-model layer is closed to new entrants. Anthropic and OpenAI together capture about 89% of revenue among top AI startups, leaving every other AI company to compete for the remaining 11%.
The most direct lesson is in the contrast between the two leaders. OpenAI built its valuation on consumer mindshare, with ChatGPT serving as the cultural face of generative AI. Anthropic built its valuation on the enterprise, with API contracts, Fortune 10 customers, and ACVs above $1 million. The market has now repriced the enterprise playbook as the more durable one. The same investors who funded the consumer-first story at $852 billion are paying more for the enterprise-first story at $900 billion.
The 89% concentration carries a second implication. The capital required to train a frontier foundation model is now measured in the tens of billions, and the named lead investors at that tier are committed to Anthropic and OpenAI. The window for a new venture-backed foundation-model company has functionally closed. Application-layer and vertical-AI startups remain the open territory, but they sit on top of a platform layer that is now a duopoly. Pricing, rate limits, terms of service, and competitive product launches from Anthropic or OpenAI are structural risks that every AI application company carries on its balance sheet.
The hedge at the top of the AI market
The investor overlap is the second story buried in the round. Three of the four Anthropic co-leads also hold positions in OpenAI. The era of “pick a horse” venture investing at the foundation-model layer is over. The largest growth funds have decided that the cost of being wrong about which of the two companies wins is higher than the cost of owning both at near-trillion-dollar valuations.
For founders raising in adjacent AI categories, the same firms writing $2 billion checks into the model layer are the firms most likely to lead growth rounds at the application and infrastructure layers. Those rounds will be underwritten with one question: does this company benefit from the success of Anthropic or OpenAI, or does it depend on a third foundation-model winner emerging? Companies in the first bucket get funded. Companies in the second do not.
The enterprise-first repricing also reframes the AI capex cycle. Anthropic’s $100 billion AWS commitment and its SpaceX/Colossus compute partnership are now backstopped by a $45 billion revenue base and a $900 billion valuation. Bets that looked aggressive twelve months ago look proportional today. The companies that sell into Fortune 500 procurement and survive a security review are being valued at multiples consumer-facing AI products cannot reach.
What to watch through the rest of 2026
Three questions sit immediately downstream of the round. The first is the close itself. Both Bloomberg and The Information report the deal is on track to close by the end of May. If it slips, that is the first signal that the price is hitting resistance.
The second is the IPO timeline. Bloomberg has reported that Anthropic is weighing a public listing as early as October 2026, though some bankers privately estimate a public-market valuation between $400 billion and $500 billion. That is a meaningful discount to the $900 billion private mark, and it would put Series H investors underwater on paper from day one. Whether Anthropic files in October will tell the market how confident the company is in defending the private valuation in public.
The third is the OpenAI response. OpenAI’s own IPO setup has been complicated by exactly the enterprise-versus-consumer pricing dynamic on display in the Anthropic round. The pressure to demonstrate enterprise revenue growth, not just user counts, will now intensify. The two companies that capture 89% of AI startup revenue are heading into the second half of 2026 with the same investor base, similar valuations, and very different product strategies. Founders building in AI will be planning around whatever the gap between them looks like by year-end.
