Anthropic, the San Francisco AI safety company behind the Claude models, raised $65 billion in a Series H round on May 28, 2026, at a $965 billion post-money valuation. That figure makes it the world’s most valuable AI startup, eclipsing OpenAI’s $852 billion by $113 billion.
The milestone inverts the narrative that defined the generative-AI era. The company founded by former OpenAI employees on a safety-first thesis is now, by private-market valuation, worth more than the lab it split from.
The round that passed OpenAI
Altimeter Capital, Dragoneer, Greenoaks, and Sequoia led the $65 billion raise. Capital Group, Coatue, D1 Capital, GIC, ICONIQ, and XN joined as co-leads — an investor list that reads like a roll call of the funds that set late-stage technology prices.
Roughly $15 billion of the round came from hyperscalers, including $5 billion from Amazon, deepening a partnership that already underpins much of Anthropic’s compute. The $965 billion post-money mark sits $113 billion above OpenAI’s most recent $852 billion valuation.
A billion revenue run rate
Anthropic is now operating at a $47 billion annualized revenue run rate, up from $30 billion just three months earlier. That is roughly 57% growth in a single quarter, a pace that justifies — to its backers, at least — a valuation approaching the trillion-dollar line.
The revenue story is increasingly an enterprise and developer story. Over half of Claude Code revenue now comes from enterprise customers, and Anthropic’s argument to investors is that paid, sticky business usage scales more predictably than consumer subscriptions.
Claude is taking ChatGPT‘s download share
The valuation reversal tracks a shift in usage. ChatGPT‘s share of global AI app downloads fell from 67% to 47% since the second quarter of 2025. Over the same window, Claude climbed from 1% to 14% of downloads.
OpenAI still leads in raw consumer reach, but the trend line matters more than the absolute number to the investors pricing this round. A 14x jump in download share is the kind of signal that moves late-stage capital.
Chipmakers bought in strategically
Three memory and chip suppliers — Micron, Samsung, and SK Hynix — invested strategically in the round. Their participation is not financial tourism. Frontier-model training and inference are bottlenecked by high-bandwidth memory, and a direct equity stake aligns the suppliers with the customer consuming their highest-margin parts.
It also signals where the supply chain expects demand to concentrate. When the firms that make the memory bet on the lab that buys it, they are pricing in years of capacity commitments.
First frontier model on all three clouds
Claude is now the first frontier model available simultaneously on AWS, Google Cloud, and Azure. That triple-cloud distribution is a structural advantage: enterprises that standardize on a single cloud can reach Claude without leaving their existing procurement and security perimeter.
The contrast with single-cloud competitors is the point. Anthropic is positioning Claude as the neutral, available-everywhere option for enterprise buyers who do not want to bet their AI strategy on one hyperscaler’s roadmap. Microsoft’s own AI chief recently argued the opposite case on cost — a debate we covered in Microsoft’s claim that Anthropic’s models are too expensive.
Why the safety-first company is winning the money race
The simplest read is that safety positioning and commercial dominance are no longer in tension. Anthropic spent years arguing that careful, honest models were a competitive feature, not a tax — the same posture behind the recent joint call from OpenAI, Anthropic, and Google for Congress to curb AI bioweapons risk.
The $965 billion valuation is the market’s verdict that the strategy works. Enterprise revenue, triple-cloud reach, and strategic chip-supplier backing have turned a research lab’s caution into a moat. The remaining question is whether OpenAI responds on price, product, or its own delayed fundraise — and how quickly.
