OpenAI commits $1B to cyber defense as Astra finds zero-days

OpenAI is putting $1 billion behind cyber defense, subsidizing access to its Daybreak security platform after unveiling that its new Astra model can independently find previously unknown software vulnerabilities. Astra is the first model to reach OpenAI's "Critical" cybersecurity threshold, meaning it can discover zero-days and build working exploits across hardened systems without human help.

The milestone signals a step up in autonomous offensive-capability AI, and the fresh funding separates OpenAI's defensive offerings from that threat research. The company is positioning Daybreak — now cheaper via the subsidy — as a defense tool for security teams responding to the fast-moving AI-attack cycle the same models enable.