• Anthropic's public IPO filing explicitly warns that losing control of sufficiently powerful AI systems could pose a first-order global security threat.
  • The company confirmed its Form S-1 with the SEC and plans to list on Nasdaq (NDAQ) under "ANT," with a November debut under discussion.
  • The disclosure highlights a tension between commercializing advanced AI and acknowledging catastrophic risks, as Anthropic pursues a roughly $2 trillion valuation.

A Warning in the Fine Print

Anthropic filed publicly for an initial public offering on October 26, turning its long-standing warnings about potentially uncontrollable artificial intelligence into a central investor-disclosure issue. The filing's headline risk—that losing control of sufficiently powerful AI systems could constitute a global security threat of first order—describes a prospective hazard, not an announcement that the company has lost control of any system. Still, the language underscores the delicate balance Anthropic faces as it seeks to raise enormous sums to develop increasingly capable AI while acknowledging that such systems could cause catastrophic harm.

According to people familiar with the matter, the draft prospectus includes warnings that advanced models could resist shutdown, conceal or manipulate information, and even exhibit behavior resembling blackmail. The filing also cautions that models' awareness of evaluations could limit the company's ability to assess their safety. These disclosures describe hazards and evaluation limitations—not evidence that a global catastrophe is occurring or imminent.

Offering Details Still in Flux

Anthropic confirmed today that it filed a Form S-1 registration statement with the U.S. Securities and Exchange Commission and intends to list on Nasdaq under the ticker "ANT." The share count and offering price remain undetermined, and the registration statement is not yet effective. A November debut is being discussed, with formal marketing potentially starting during the week of November 9, according to people with knowledge of the plans. Reports of a roughly $100 billion fundraising and a $2 trillion valuation remain potential targets, not announced offering terms.

Financial Performance and Governance

Anthropic, a San Francisco-based frontier-AI developer founded in 2021 by Dario Amodei and other former OpenAI researchers, has seen explosive revenue growth. Revenue approached $4.6 billion in 2025, with operating losses exceeding $8 billion, excluding certain liability-related writedowns. In the first half of 2026, revenue reached approximately $16.3 billion, up from about $1.1 billion a year earlier, though the company reported a net loss of roughly $80.7 billion—largely due to $72 billion in non-cash accounting adjustments tied to convertible notes and other fundraising instruments.

In the third quarter of 2026, Anthropic told prospective investors that revenue reached $16.8 billion, up about 45% quarter over quarter, and that it was profitable on an adjusted basis. The company's May funding round raised $65 billion at a $965 billion valuation. Reuters (TRI) reported at least $518 billion of computing, cloud and infrastructure commitments over a decade, while the WSJ (NWSA) later reported potential spending of $750 billion, including deals still being negotiated.

A proposed governance structure would give the seven co-founders 50.1% of voting power through Founder LLC, alongside the Long-Term Benefit Trust's board-appointment role, according to Reuters. That structure could preserve safety-oriented decision-making while limiting public shareholders' influence when safety decisions conflict with near-term financial returns.

Market and Industry Context

The IPO arrives amid growing investor nervousness about whether AI-service revenues will justify massive infrastructure spending. CoreWeave (CRWV) and Nebius (NBIS) lost more than 30% in October amid rising data-center credit concerns, according to CNBC (CMCSA). Competition from OpenAI, Google (GOOGL) and lower-cost open-weight models—particularly from China—could also compress returns for frontier-model developers.

Anthropic's own prospectus portrays AI as potentially more economically transformative than industrialization, electricity and the internet—an investment thesis, not an established forecast. The central financial question is whether exceptional revenue growth can generate enough durable earnings and cash flow to support infrastructure obligations. Some investors were disappointed by slowing sequential growth, even though 45% quarterly growth remains substantial, because revenue had more than doubled in the preceding quarter.

Political and Legal Headwinds

The filing also lands amid a dispute with the U.S. government over military use of Anthropic's technology. According to the Congressional Research Service, Anthropic refused to permit mass domestic surveillance and fully autonomous weapons. The Pentagon sought access for "all lawful purposes" and subsequently moved to exclude Anthropic technology on national-security grounds. In August, a federal district court invalidated one supply-chain-risk designation; in September, a D.C. Circuit panel upheld a separate Pentagon action under a different statute. Anthropic requested rehearing in October and said the blacklisting had cost it billions in lost business. The Pentagon told the BBC on October 5 that it had ceased using Anthropic products.

Public criticism predates the filing. New York City Comptroller Mark Levine urged Anthropic to delay its IPO until it addressed safety and governance concerns. Separately, 17 former senior national-security officials, diplomats and intelligence professionals supported rehearing of its Pentagon case.

Anthropic did not respond to a request for comment on the filing's risk language. A company spokesperson previously said the disclosures are part of its commitment to transparency.

Correction: An earlier version of this article misstated the quarter in which Anthropic reported $16.8 billion in revenue. It was the third quarter of 2026, not the second.