OpenAI is putting $1 billion into a new program to help critical infrastructure defend against AI-driven attacks.
Called Daybreak for Frontline Defenders, the program won't hand cash directly to organizations. Instead, the money is being converted into access to the Daybreak model, training, and technical support.
The first batch targets US water utilities, power grids, local governments, community banks, nonprofits, and open-source projects, with the plan expected to roll out over the next six months.
Daybreak is meant to help these organizations spot vulnerabilities, analyze anomalies, validate risks, and test patches. OpenAI says about 2,000 organizations and workspaces are already using it, and Daybreak is set to be integrated into more than 35 enterprise products and services.
The announcement comes on the same day as GPT-6 Astra, OpenAI's first model to reach the 'Critical' threshold for cybersecurity capability. With supporting tools, Astra can track down unknown vulnerabilities and research new exploitation methods without a human guiding it step by step. OpenAI says AI-driven attacks will only become more common and more sophisticated over the coming months.