Anthropic said on May 28 that it raised $65 billion in Series H funding, a round that values the company at $965 billion post-money and includes $15 billion of previously committed investments from hyperscaler partners, largely Amazon. The announced figure vaulted the Claude maker to the top of the frontier-valuation league table by latest disclosed marks, though it underscored that the race among frontier AI labs is increasingly a contest to finance the electricity and silicon needed to train next-generation models.
Anthropic said the proceeds would fund safety and interpretability research, scale compute to meet demand for its Claude assistant, and accelerate product and partnership development. The funding announcement did not disclose a chief executive statement beyond the formal use-of-funds language.
The financing pushes Anthropic’s last-posted valuation above OpenAI’s $852 billion post-money figure from March 31, when OpenAI announced a $122 billion committed-capital round. But by total dollars committed, OpenAI’s raise remains substantially larger—more than double Anthropic’s headline amount. The difference matters because the raw raise-to-raise comparison distorts the picture: OpenAI’s round was $122 billion in fresh committed capital, while Anthropic’s $65 billion includes $15 billion that was already tied to earlier cloud infrastructure agreements.
The round drew a broad mix of traditional growth-equity firms, sovereign wealth funds, and technology infrastructure partners. Lead investors included Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, with co-leads Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ, and XN. The cap table also added Blackstone, Brookfield, Fidelity, General Catalyst, Lightspeed, MGX, Temasek, and strategic infrastructure investors Micron, Samsung, and SK hynix—chipmakers whose memory products underpin the high-bandwidth compute clusters Anthropic consumes. Bloomberg reported, citing unnamed sources, that each lead investor contributed more than $2 billion; the investors either declined to comment or did not respond.
The valuation leap is stark. In February, Anthropic closed a $30 billion Series G at a $380 billion post-money valuation. Within a little more than three months, the company doubled its private-market mark, even as press reports in April and May suggested it was seeking between $30 billion and $50 billion at valuations around $800–$900 billion. The final $965 billion mark suggests demand exceeded expectations, but the heavy presence of pre-committed hyperscaler money makes the number less a test of open-market pricing than a reflection of embedded compute partnerships. Amazon, which had disclosed a $5 billion investment in April with up to $20 billion more in future compute-linked commitments, is both a funder and primary cloud and training partner.
Separately, Anthropic said its annualized run-rate revenue crossed $47 billion earlier in May. The figure is company-reported and not audited GAAP revenue; it offers a directional signal of enterprise and developer uptake for Claude but cannot be directly compared to the trailing twelve-month revenue of a public software company. The company remains private, with no public audited financials, and has not disclosed profitability or cash burn.
The precise split between newly arranged capital and existing hyperscaler pledges beyond the $15 billion disclosed is not public. Financing terms, including preferred-share provisions and liquidation preferences, are unknown. Anthropic has not filed a registration statement with the SEC, and the company has not announced plans for an initial public offering. The absence of those details means investors and competitors can measure the sticker price only in the context of a private, structured round—not a market-clearing public transaction.
For an industry where a single training run can cost hundreds of millions of dollars, the round confirms that access to capital and pre-secured compute capacity is now as much a moat as model architecture.