Anthropic has officially filed to go public

The filing itself is less surprising than its timing: Anthropic is entering public markets while the AI infrastructure investment wave is still cresting, meaning its S-1 disclosures will force the first real public accounting of what frontier safety research actually costs relative to revenue, a question private funding rounds have never required the company to answer cleanly.

SoftBank's $87.3 billion commitment to French AI infrastructure, covered here the same day, illustrates the capital environment Anthropic is stepping into: institutional money is moving fast and at scale, and public markets represent the next logical pool. That context matters because Anthropic's valuation will partly depend on whether investors treat it as a safety-first research lab or a compute-and-API business. The Import AI digest from June 1st, which surfaced the nascent economics of pricing extinction risk, is directly relevant here: public shareholders will need a framework for that kind of liability, and no such framework currently exists in standard equity analysis.

Watch whether OpenAI accelerates its own IPO timeline within 90 days of Anthropic's S-1 becoming public. If it does, the two filings together will establish competing valuation templates for how public markets price safety spending versus near-term revenue, and whichever model trades at a higher multiple will define the sector's capital logic for years.