# Anthropic's Revenue Hits $47 Billion as Daniela Amodei Makes the Case for an IPO

The Vault — AI Edition | June 15, 2026 | 4 min read | 842 words

Category: business

Key Takeaway: Anthropic's annualized revenue has surged past $47 billion, and President Daniela Amodei is using that ramp to argue that the punishing upfront cost of training and serving frontier models is precisely why the company belongs in the public markets.

---

Anthropic has crossed a threshold that not long ago would have read like a typo. The maker of Claude saw its annualized revenue blow past $47 billion in May 2026, according to figures circulating ahead of the company's planned public offering -- a more than fivefold jump from the roughly $9 billion run rate it reported at the end of 2025. In a single year, one of the fastest commercial ramps in software history has handed Anthropic the kind of top-line momentum that founders normally spend a decade chasing.

The number arrives at a charged moment. Just days earlier, on June 1, Anthropic confirmed it had confidentially submitted a draft S-1 registration statement to the Securities and Exchange Commission, setting the stage for what could be one of the largest technology listings ever. And on June 4, speaking at Bloomberg's Tech conference in San Francisco, co-founder and President Daniela Amodei tied the two threads together: the revenue is real, the costs are enormous, and the public market is where she believes a company like Anthropic now belongs.

The Case Amodei Is Making

Amodei was direct about the economics driving the decision. "It's a really big upfront cost to train the models and to serve inference on them," she said. "And I think that my guess is that over time, the sort of core set of companies that are working to advance the frontier are just going to need access to capital. And I think the public market is very well suited to that."

It is an argument that reframes the IPO not as a victory lap but as a financing necessity. Frontier AI is brutally capital-intensive: each new model generation demands a larger training run, and every additional Claude user adds to a serving bill that scales with usage rather than shrinking with it. Public equity markets, in Amodei's telling, are simply the deepest and most durable pool of capital available to companies that intend to keep spending at that frontier for years.

Asked point-blank whether Anthropic was racing to become the first AI lab to go public -- ahead of OpenAI and SpaceX, both of which are reportedly weighing their own listings -- Amodei declined to comment. The reticence is notable. The Financial Times has framed the three companies as a potential three-way sprint toward debuts that would rank among the biggest in market history.

A Valuation in the Stratosphere

The IPO conversation is happening against a freshly minted private valuation that itself strains comprehension. On May 28, Anthropic announced a $65 billion Series H round that valued the company at $965 billion post-money -- knocking on the door of a trillion dollars and making it, on paper, the most valuable AI startup in the world. The round included previously committed investments of $15 billion from hyperscalers, among them $5 billion from Amazon, and Anthropic said the capital would fund safety research, expand compute, and scale the products its customers depend on.

That nearly trillion-dollar marker is the lens through which the $47 billion revenue figure has to be read. A 20-times multiple on annualized revenue is rich, but far from the stratospheric ratios the company carried when its run rate was a fraction of today's. The revenue ramp, in other words, is doing the work of making an extraordinary valuation look less extraordinary.

The Strategy: Buy Compute, Don't Build It

What separates Anthropic's pitch from the broader AI-infrastructure arms race is a deliberate decision not to pour its own balance sheet into data centers. Rather than committing to ground-up construction it might not be able to fill, the company leans on partners -- buying capacity from Amazon at a scale running into multiple gigawatts and, per disclosures tied to SpaceX's own filing, tapping xAI's Colossus cluster in a compute arrangement reported to run roughly $1.25 billion a month.

The logic is the same "do more with less" discipline Amodei has championed for months: avoid overcommitting to infrastructure you cannot productively use, and compete on capability per dollar of compute rather than raw spending. It is a pointed contrast to rivals making trillion-dollar headline commitments, and it lets Anthropic keep more of its capital aimed at models and research while outsourcing the cement-and-silicon risk to companies whose core business is building it.

What to Watch Next

The next markers are concrete. The SEC review of Anthropic's confidential filing will determine how quickly a public debut can follow, with reporting pointing toward a possible listing later in 2026. Watch whether the revenue curve holds its slope into the back half of the year, how the company frames its margins and compute obligations once the S-1 goes public, and whether OpenAI or SpaceX move to file first. For now, Amodei has planted her flag: the costs are the point, and the public market is the answer.

"It's a really big upfront cost to train the models and to serve inference on them. And my guess is that over time, the core set of companies that are working to advance the frontier are just going to need access to capital. And I think the public market is very well suited to that."
- Daniela Amodei, President & Co-Founder, Anthropic
$47B
Annualized revenue
$965B
Valuation
$65B
Series H round
$1.25B/mo
xAI Colossus compute