Altman says listing in 2026 would be ill-advised given the state of safety work, parking the S-1 OpenAI filed confidentially in June - which also spares it from explaining its rogue-agent incidents in securities filings while Anthropic heads for a $2 trillion float.

Twenty-five Fields Medallists say the labs' goals and mathematics' goals are 'severely misaligned' - written in a week because they judged it urgent - and OpenAI has now pulled its Caltech sponsorship.

Thune, Cruz and Klobuchar are negotiating a duty of care that would let Washington bar an unsafe frontier model, with national-lab experts testing for nuclear and biological uplift and a court challenge as the only appeal - three Senate voting weeks remain before the midterms.