Back to storiesNewsGlobal Highlights

Anthropic's 335-page prospectus opens its numbers to the market

By Editorial Desk4 min read
Anthropic's 335-page prospectus opens its numbers to the market
The Claude maker's SEC filing shows revenue of US$ 16.286 billion in the first half of 2026 and an adjusted profit of US$ 1.618 billion, alongside long-term compute commitments that could exceed US$ 600 billion.

Anthropic has registered a 335-page prospectus with the SEC for a listing that, if it goes ahead at the scale reported, would be the largest initial public offering in history. The Claude developer has not set a date, price range or share count, but is expected to debut on the Nasdaq under the ticker ANT in the week of 16 November, according to Token Times. The company is reportedly seeking to raise at least US$ 100.000 billion, above the US$ 85.700 billion SpaceX raised in June, with investors projecting a valuation of between US$ 1,5 and US$ 2 billion. Goldman Sachs and Morgan Stanley are leading the placement alongside JPMorgan, Citi, Barclays, Bank of America, Deutsche Bank and UBS.

The financial disclosures show growth at a pace rarely seen. Revenue rose from US$ 41 million in 2023 to US$ 386 million in 2024 and about US$ 4.585 billion in 2025, then reached US$ 16.286 billion in the first half of 2026 alone, with a year-on-year rise of 1.368% in the second quarter. More than 11.900 enterprises now pay at least US$ 100.000 a year, up from fewer than a thousand a year earlier, and roughly 250 spend more than US$ 10 million. Net revenue retention is around 900%. The prospectus also flags concentration: the five largest customers account for about 25% of revenue, and many have no long-term contracts.

On paper the losses are vast — US$ 41.965 billion in 2025 and US$ 80.748 billion in the first half of 2026 — but the latter figure is inflated by a US$ 72.400 billion accounting charge tied to the rise in its valuation. The operating picture is different: the operating loss narrowed to US$ 693 million in the half from US$ 2.135 billion a year earlier, and on an adjusted basis the company posted a positive result of US$ 1.618 billion. It held US$ 131.700 billion in cash and short-term investments at 30 June, having raised US$ 141.100 billion over its history. XTB, whose analysis is cited in the coverage, cautioned that the net loss exaggerates recurring pressure but that the company is not yet structurally profitable, making operating income the variable to watch.

Anthropic earns mainly from usage of its models rather than subscriptions, selling Claude directly and through AWS, Google Cloud and Microsoft Azure, channels that accounted for 48% of half-year revenue. Compute and infrastructure spending reached US$ 7.326 billion in 2025 and US$ 8.945 billion in the first half of 2026, though the company says revenue is growing faster than that outlay. The larger weight sits in the future: after conceding it had underestimated capacity needs, Anthropic signed long-term, mostly non-cancellable commitments with Google, Amazon, SpaceX, Microsoft, Nscale and Lambda, among others. Estimates of the total vary — more than US$ 475.000 billion over the next decade, rising to about US$ 750.000 billion including chip leases with Broadcom and AMD, while XTB puts the figure above US$ 600.000 billion with a potential maximum of US$ 807.000 billion. The prospectus itself warns the expansion could force it to seek new financing sooner than planned.

Control will stay tightly held. The seven co-founders, including Dario and Daniela Amodei, will concentrate their rights in a vehicle with a class F share carrying 50,1% of voting power, while the Long-Term Benefit Trust, set up to protect the company's mission, will hold a class T share giving it the power to appoint and remove a majority of the board. Amazon and Google will each keep more than 5% of the capital but with non-voting shares, leaving minority investors little influence over strategy. Founders and large shareholders are subject to a 365-day lock-up, with a partial release of up to 29% at 180 days. The prospectus devotes 94 pages to risk factors, including frontier-AI hazards such as advanced systems causing catastrophic harm, autonomous model behaviour and capabilities outpacing safety, and warns that the insurance sector lacks capacity to cover a catastrophic incident. It names OpenAI, Google and open-weight models as competitors.

Related topics & keywords

  • #ai
  • #ipo
  • #anthropic
  • #nasdaq
  • #artificial intelligence
  • #capital markets

Related Stories

View all news →