Anthropic has agreed terms on a $30 billion funding round at a $900 billion pre-money valuation, the Financial Times reported, citing two people familiar with the deal. The round is being co-led by Sequoia Capital, Dragoneer Investment Group, Altimeter Capital, and Greenoaks Capital Partners, each writing roughly $2 billion checks. Founders Fund and General Catalyst are participating from the existing cap table. The round is expected to close this week.
Anthropic passes OpenAI on paper
At $900 billion pre-money, Anthropic vaults past OpenAI's most recent valuation of $852 billion, making it the most valuable private AI company in the world. Three months ago, the company's Series G closed at $380 billion post-money on a $350 billion pre-money mark. The new round is the second $30 billion raise of 2026 and the third $30 billion-plus financing the company has closed since February. On an FT-cited annualized run rate above $45 billion, Anthropic is now tracking roughly five times where the business exited 2025.
The four-firm lead group is the real signal
Co-leading with checks of roughly $2 billion each is unusual at any stage; in growth, it's a structural choice that says no single fund could (or wanted to) absorb the concentration risk of a $30 billion check. Sequoia, Dragoneer, Altimeter, and Greenoaks all have existing positions in Anthropic, and all four sit inside the small group of growth investors with both the capital base and the conviction to keep marking up the same company through this much delta in three months. The pre-money jump from $350B (February) to $900B (May) is a 2.57x in 12 weeks. The market is pricing the model itself, not the existing revenue.
What the money is for
Anthropic has not disclosed deployment plans for the round, but the answer is the same as every Frontier-lab raise this year: compute. The company's most recent infrastructure agreements are a $100 billion-plus, multi-year Project Rainier commitment with AWS for Trainium2 capacity and a separate Google Cloud agreement for up to one million TPUs. At current training-run economics, $30 billion is the scale of an 18-to-24-month compute commitment. The Stainless acquisition the lab announced earlier this month — picking up the developer-tooling startup whose SDK generators power major frontier-lab APIs — points at the other half of the spend: closing the distance between model release and enterprise deployment. For context on how rapidly the consolidation cycle has tightened, see our piece on the dual-mode memory files architecture Anthropic shipped earlier this week.
Bloomberg first reported the talks on May 12 at the $900 billion target. The FT confirmation that terms have been agreed effectively closes the question. No term sheet has been signed publicly, and the deal could still slip, but at this point the closing is procedural.
Comments