OpenAI commits $1B to defend critical infrastructure from AI-driven attacks
What's the deal? OpenAI has launched a $1 billion programme, Daybreak for Frontline Defenders, to help operators of critical infrastructure protect themselves against cyberattacks. The money funds subsidised access to specialised AI models, training, and product integrations. It starts in the US via an initiative called Daybreak for America, with a rollout to partner countries planned over the coming months.
Who can access it? Water and energy network operators, local governments, regional banks, healthcare providers, open-source developers, and non-profits can apply for the budget, which OpenAI intends to deploy within six months.
What the programme includes: The initiative rests on three pillars. The $1 billion covers subsidised access to two models — Daybreak Blue for general defensive tasks and Daybreak Red for advanced technical research — which help analyse legacy source code, detect suspicious activity, prioritise risks, and develop patches. A pilot with the Multi-State Information Sharing and Analysis Center (MS-ISAC)Dealroom has a profile for this one. Try Dealroom → will train public bodies and water authorities on incident response, while more than 35 partner security products integrate the models via the Daybreak Defense Network.
Why now? AI-driven cyberattacks are rising in frequency and sophistication. Large firms have big budgets and dedicated security teams, but critical services often run on outdated systems with tight budgets.
OpenAI frames the moment as a "defender's window" — a temporary opening to close security gaps before attackers exploit them. "Defenders must act now: test systems, find weaknesses, and strengthen defences before attackers do," it said (translated from Dutch).
From emergency aid to structure: The investment builds on earlier acute responses. After recent cyberattacks on US water systems, OpenAI offered free API credits and technical support to affected utilities, letting operators find and fix leaks while keeping supplies running.
The product: This week OpenAI also published the architecture of its Defense Factory — an automated, AI-agent-based method that continuously finds and validates vulnerabilities, then prepares ready-made patches for human review. "The goal is not only to find more vulnerabilities, but to close them faster," the company said (translated from Dutch).
The signal: OpenAI is positioning AI as a defensive tool for the underfunded operators that keep water, power, and finance running — a bet that the same technology fuelling new attacks can help outpace them.
Image credit: crystallinelamp
