// news · industry · compute2026-05-25source: anthropic / forge / letsdatascience

Anthropic in final stages of Series H seeking $50B at $900B valuation — would be the largest single AI-lab funding round in history

Anthropic is in the final stages of a Series H raise seeking $50 billion of new funding against a $900 billion valuation. That's a 137% jump from the $380B post-money Series G closed earlier this month, and a single round larger than OpenAI's entire $122B March 2026 funding total.

The valuation trajectory is what's eye-opening. Anthropic closed Series G at $380B post-money in early May. Three weeks later the company is in talks for Series H at $900B. The 137% uplift in 21 days is not driven by new financial disclosures (the company hasn't released updated revenue figures); it's driven by the market repricing AI-lab valuations against the SpaceX June 12 IPO target at $1.75T and the OpenAI IPO confidential filing aiming for $1T+. If the public-market peer set clears at those numbers, Anthropic at $900B is a discount-to-comps trade rather than a stretch.

The capital structure question that the round resolves: Anthropic has the [$30B → $50B] commercial ARR trajectory plus the SpaceX compute deal of $1.25B/month through May 2029 plus the enterprise JV with Blackstone, Hellman & Friedman, and Goldman Sachs from this morning's industry cycle. A $50B raise at $900B gives Anthropic the war chest to compete with OpenAI's Stargate-scale infrastructure ambitions without conceding the strategic-independence advantages of staying private through 2026-2027. The trade-off is dilution: existing Anthropic investors (including Amazon and Google as the largest external holders) accept ~5-6% dilution to fund the next 18 months of operations without needing to IPO under unfavorable market conditions.

See our analysis →

Forge Global — Anthropic Upcoming IPO and Private Stock Price → · Let's Data Science — OpenAI $10B Deployment Company Anthropic $1.5B Blackstone JV → · Axios — Why private equity is making deals with the AI giants →