Anthropic files S-1 for Nasdaq listing under ticker ANT: $80 billion loss, revenue triples
Anthropic, PBC, the company behind the Claude language model, filed its formal registration statement on Form S-1 with the U.S. Securities and Exchange Commission (SEC) on Monday (26), initiating the public process for an initial public offering. The company will list its Class A common stock on the Nasdaq Global Select Market under the symbol "ANT." The number of shares to be offered and the price range have not yet been determined.
The announcement, published on Anthropic's own website in an official statement, contained no surprises — the company had already confidentially submitted a draft S-1 on June 1, giving the market months to prepare. What has been surprising, however, is what the prospectus has revealed through the press in recent months: the sheer scale of the costs and the bet on infrastructure.
What the S-1 shows: numbers that defy conventional wisdom
According to the preliminary version of the prospectus, first reviewed by the Wall Street Journal and subsequently confirmed by the Financial Times, Reuters, and Bloomberg, Anthropic generated revenue of $16.3 billion in the first half of 2026 — a leap compared to the $4.6 billion it earned in all of 2025. In the same period, the net loss reached $80.7 billion.
The magnitude of the loss is staggering, but most of it reflects an accounting adjustment related to long-term investments in artificial intelligence assets, not traditional operating losses. Still, the figure serves as a stark warning about the financial cost of training and scaling frontier models.
In the third quarter alone, revenue reached $16.8 billion, a 45% increase from the previous quarter. In the second quarter, the number had already been $11.6 billion. Based on Bloomberg estimates, annualized revenue should exceed $100 billion in 2026.
The compute expenditure: $750 billion in cloud contracts
One of the most discussed aspects of the prospectus is the projected infrastructure costs. Anthropic has signaled it could spend more than $750 billion over the coming years on compute capacity, including contracts still under negotiation. This is not a fixed figure — it is the company itself admitting that it does not know how many billions will need to be allocated to remain competitive.
Reuters reported that, according to the prospectus, Anthropic may invest this amount in contracts with cloud providers — AWS, Google Cloud, Microsoft Azure — as well as agreements with third-party data center operators. The Financial Times detailed that the company is also building its own data centers in partnership with SpaceX: the Colossus I and II contract, financed by SpaceX, requires payments of $1.25 billion per month until May 2029, totaling $15 billion per year.
TechCrunch reported that Anthropic's cloud partnerships are becoming increasingly strained. With Anthropic growing at this pace, the very cloud providers begin to see the company not as a customer, but as a potential competitor in AI services. The company also fears that third-party data center operators have a "limited operating history," which could make them less reliable than the hyperscalers.
Customer concentration and dependency risk
A point that the S-1 highlights with the transparency required by the SEC is the concentration of customers. According to the WSJ, just two customers account for about one-fifth of sales. In other words, Anthropic still relies on a few large buyers to generate revenue — a risk that public investors need to consider.
Reuters also reported that some current investors were disappointed by the slowing sequential growth rate in the latest quarter. Revenue more than doubled from the first to the second quarter, but the 45% pace in the third quarter was slower than the previous rate. For a market accustomed to exponential growth, this sounds like deceleration.
The safety and governance debate
While the market prepares to receive Anthropic, there is no shortage of external debate. New York City Comptroller Mark Levine wrote to CEO Dario Amodei asking the company to pause the IPO until it addresses AI safety, governance, and development pace concerns. The letter raises a question that goes beyond Anthropic: to what extent can a traditional capital market evaluate the risks of systems that are not yet fully understood?
The prospectus also includes the company's own warnings about "catastrophic" risks associated with AI development — a rare element in public offerings and one that raises a question: how can a company sell its shares to the public when it itself acknowledges that its own products might pose existential threats?
What comes next
According to IFR (Investor Feed Report) reports, Anthropic plans to make the S-1 public this week, with an investor roadshow during the week of November 9 and a targeted IPO for mid-November. The price range, share count, and final valuation have not yet been determined. Previous reports indicated a valuation target of up to $2 trillion, although some investors are lowering their expectations to about $1.5 trillion.
Anthropic will be the first "frontier AI lab" to become a public company. Alongside OpenAI, which also filed a draft S-1 on June 8, Anthropic is shaping what it means for an artificial intelligence company to go public. The question remains: is the stock market prepared to invest in a company whose risk could be existential?
Sources: RuntimeWire, CNBC, TechCrunch
✓ Independent sources cross-checked and verified before publishing
