CNBC reported on 19 October that Anthropic has become the AI trade's last realistic candidate for a major public listing this year. No filing has been made and the company has not announced an intention to go public.
The timing is what makes it interesting. One day earlier, Anthropic said it intends to commit 5% of its shares to a new public benefit fund, subject to stockholder approval, with the stake able to grow as the valuation does. Those two stories pull in different directions. A pre-IPO company asking existing holders to approve 5% dilution is making a harder request than the announcement implied, and the stockholder vote is now the thing to watch rather than the press release.
The surrounding week does not help the case for a smooth listing. Anthropic received an FTC civil investigative demand on 14 October covering security incidents, safeguarding practices, marketing materials and privacy policies. It is in reported talks with Nscale for a second roughly $45 billion compute commitment. And it signed a 20-year lease for a C$10 billion Nova Scotia campus. Enormous contracted obligations plus an open federal investigation is a particular kind of S-1.