Anthropic Files Public S-1 for Proposed IPO
Anthropic, PBC submitted its public registration for a proposed initial public offering on October 26, 2026, according to U.S. Securities and Exchange Commission records. The Form S-1 sets out a planned Nasdaq Global Select Market listing using ANT as the ticker, without specifying an offering price or final valuation.
The public filing follows a confidential draft submission that Anthropic said it made to the SEC on June 1, 2026. In its statement at that time, the company said an IPO remained conditional on regulatory review, market conditions and other factors. Neither the number of shares nor their price had been determined at that stage.
The registration documents put Anthropic’s revenue for January through June 2026 at $16.3 billion, against approximately $1.2 billion for the corresponding six months a year earlier. Its reported net loss for the 2026 period was $80.7 billion. Of that loss, the filing assigns $72.4 billion to a nonrecurring accounting adjustment associated with financing agreements.
For comparison, the company’s reported results for the whole of 2025 showed revenue of $4.585 billion and a net loss of approximately $33.9 billion. Spending on AI infrastructure reached roughly $8.9 billion during January through June 2026, according to the filing, compared with approximately $2.3 billion during those months in 2025.
Customer concentration figures cover a different period: Anthropic said approximately a quarter of its revenue during January through September 2026 came from its five largest customers combined. Their identities are not given in the registration statement.
In describing potential benefits of AI, Anthropic identifies biology and medicine, economic expansion, broader sharing of the gains from that expansion, and the provision of government services. The document presents these as opportunities identified by the company, rather than outcomes the filing establishes.
The prospectus also identifies potential harm involving misuse, a loss of control over AI, use by authoritarian authorities and disruption to the economy. According to Anthropic’s disclosure, technical protections may become outdated, offer incomplete coverage, be circumvented or not function as intended.
A separate account in the filing concerns a Claude coding tool. Anthropic said it believes a group backed by the Chinese state manipulated the tool in attempts to infiltrate approximately 30 targets. The targets included businesses in technology, finance and chemical manufacturing, as well as government agencies. The company’s account describes attempts to gain access, rather than confirming that those targets were successfully accessed.
* This article is based on publicly available information at the time of writing.