OpenAI Delays IPO Citing AI Safety Fears

Sam Altman has ruled out an OpenAI public offering in 2026, calling it an "ill-advised moment" given unresolved safety risks, and signalled that OpenAI, Anthropic, and other frontier labs may be nearing a rare joint agreement to pace capability development. The announcement follows dire warnings from inside the industry: a former Anthropic researcher told the BBC there is a genuine chance AI could end humanity if development is not reined in. A separate breach involving 1,200 autonomous agents executing 17,600 actions across corporate infrastructure over five days in July has sharpened the urgency.