OpenAI allocates $1bn to widen cyber defence access

OpenAI has committed $1 billion to subsidise access to its Daybreak cybersecurity programme, targeting essential-service operators and other under-resourced defenders as artificial intelligence rapidly reshapes both cyber attacks and cyber defence.

The company said the funding will support subsidised access to Daybreak models and products, training, technical assistance and partnerships, with the initial focus on the United States. OpenAI said it expects the commitment to be consumed over the next six months before the model is expanded to partner countries.

Priority recipients under the “Daybreak for America” programme include water and wastewater systems, electric grid operators, state and local governments, community and regional banks, non-profit organisations and open-source software maintainers. OpenAI said many of those organisations oversee complex, ageing systems while operating with smaller security teams and budgets than major enterprises.

Daybreak is designed to let approved defenders use OpenAI’s advanced models for authorised cybersecurity work such as reviewing legacy code, analysing suspicious activity, identifying and validating vulnerabilities, prioritising serious risks and developing and testing fixes. The programme includes Daybreak Blue, which supports common defensive tasks using general-purpose models, and Daybreak Red, which gives approved organisations access to specialised cyber models for more sensitive work.

OpenAI said thousands of defenders across about 2,000 approved organisations and workspaces are already using Daybreak, including cybersecurity companies, defence bodies and law-enforcement organisations.

The company also announced a pilot with the Multi-State Information Sharing and Analysis Center, or MS-ISAC, focused on public-sector and water-system defenders. The pilot will combine Daybreak access with guided training and hands-on assistance for an initial group drawn from state, local, tribal and territorial organisations.

MS-ISAC provides threat intelligence, incident-response support, real-time information sharing and shared defensive services to thousands of public-sector bodies, including utilities, public hospitals, schools and law-enforcement agencies. OpenAI said the pilot is intended to help participating teams validate and prioritise findings, coordinate remediation and establish a repeatable approach that could later be extended more widely.

OpenAI said its second gathering of utility companies this week included participants representing 40 states and the District of Columbia, collectively providing essential services to more than half the US population. The company said the meetings are intended to identify practical needs and help teams integrate the tools without replacing human oversight, particularly where operational technology and services cannot tolerate disruptive security testing.

OpenAI is also expanding access through its Daybreak Defense Network, where partners are introducing more than 35 enterprise products and partner-operated services incorporating the company’s cyber models into existing security tools, services and workflows.

The initiative follows OpenAI’s broader push to accelerate defensive use of frontier AI as cyber-capable models become more powerful. The company and more than 150 organisations from technology, cybersecurity, finance, critical infrastructure and other sectors have backed a call for faster collective action, arguing that under-resourced operators need greater access to AI-enabled defensive capabilities.

OpenAI has warned that increasingly capable models are likely to make cyber attacks faster and more sophisticated, while giving defenders the same ability to automate vulnerability discovery, code review, incident analysis and remediation. Its stated aim is to use that period of accelerating capability to strengthen systems before attackers can exploit long-standing weaknesses.

The company said it had already provided assistance after attacks on US water systems, offering affected states and utilities up to $1 million in no-cost API credits, Daybreak access and technical support. According to OpenAI, teams used that support to review code and system configurations, validate findings, develop patches and confirm fixes while services remained operational.

Daybreak is part of a broader OpenAI cybersecurity platform that combines frontier models, Codex Security workflows, partner services and controlled-access safeguards. Access is restricted to verified users and organisations carrying out authorised defensive work, with identity checks, account-security requirements, monitoring and use restrictions.