Anthropic Is About to Be Profitable. It Still Needs $30 Billion.

Anthropic Has Committed $45 Billion to SpaceX. That's the Real Story in Its Latest Fundraising.

Anthropic is in conversations to raise $30 billion at a valuation exceeding $900 billion, according to people familiar with the matter — a figure that would value the AI lab at roughly 2.4 times what it was worth just 90 days ago. But the more consequential number is the $45 billion.

That is what Anthropic has agreed to pay SpaceX over the next three years for computing power, disclosures from SpaceX's IPO filing show. The deal breaks down to roughly $1.25 billion per month through May 2029, with capacity ramping through this summer. It is, by any measure, a structural commitment — the kind that reshapes a company's balance sheet for years and changes what its competitors must become to survive.

The fundraising conversations and the SpaceX disclosure, taken together, tell a story the raw valuation number obscures: Anthropic is not raising capital to fund a startup. It is raising capital to fund an infrastructure empire.

The company is simultaneously generating extraordinary revenue and burning at a scale that makes that revenue almost irrelevant to its capital needs. Anthropic is on pace to post its first potentially profitable quarter, with Q2 revenue expected to reach $10.9 billion — more than double the prior quarter, according to figures disclosed to investors and confirmed by Bloomberg. Its annualized revenue run rate has crossed $30 billion. Eight of the Fortune 10 are now Anthropic customers, the company said in a recent funding announcement.

And yet it needs $30 billion in fresh capital.

The reason is compute. The $45 billion commitment to SpaceX is the cost of staying in the race — and it exceeds the total amount Anthropic just raised from investors in February, when it closed a $30 billion round at a $380 billion post-money valuation, according to Anthropic's own account of the raise. To put that in perspective: the SpaceX deal represents roughly 18 months of Anthropic's annualized gross revenue consumed entirely by a single compute supplier, before a single other operating cost is counted. That is the scale of the capital gap the new round is meant to close.

The new round would value the company at $900 billion or more, sources told Bloomberg and The Wall Street Journal. No term sheet has been signed, and the deal is not finalized; the round is expected to close as soon as the end of May. Whether the market will accept that valuation is an open question: the comparable benchmark for a frontier AI company at IPO or large secondary carries a significant premium to current revenue, and the gap between a $900 billion ask and where comparable companies are priced remains wide enough to matter.

Anthropic is not alone in this position. In the first quarter of 2026, OpenAI, Anthropic, xAI, Waymo, and Databricks together captured roughly 75 percent of all venture investment in the United States, the Journal reported. The infrastructure requirements of frontier AI development have compressed the field to a small number of players who can raise at sovereign-wealth-fund scale — and everyone else is now facing an existential question about what they are for.

The consolidation pressure is already visible below the top tier. Labs that cannot signal they are compute-secure are finding it harder to retain talent, close enterprise deals, or justify valuations to the next round of investors. The options for a mid-tier AI company in 2026 are increasingly stark: get large enough to raise sovereign-wealth-fund rounds, get acquired by a hyperscaler, or become a specialized vertical that cannot compete for frontier capabilities. The middle ground is disappearing, and the pressure is starting to show in specific names.

OpenAI illustrates the parallel exposure. Even with a revalued Microsoft partnership and massive Azure commitments, OpenAI's burn rate remains large enough that its own CFO has flagged infrastructure obligations as a going concern, Reuters has reported. The gap between committed capital and actual usage is the same pressure Anthropic faces — the difference is that Anthropic is showing faster revenue growth as its excuse, while OpenAI is still hunting for it.

For enterprise buyers and developers building on Claude, Anthropic's staying power is becoming more legible by the day. An Anthropic with $30 billion in fresh capital and compute secured through 2029 is a different partner than one burning through runway. It is also a different competitor to every other AI lab that cannot raise at these scales. The SpaceX deal effectively puts a floor under Claude's infrastructure story for the next three years — and that certainty has a value that does not show up in any revenue line.

The question is not whether Anthropic can afford to keep building. The question is what everyone else does when the answer is yes.