OpenAI to allocate $1 bln for AI-powered cybersecurity of critical infrastructure
OpenAI has announced a global program worth $1 billion to help operators of critical services use advanced artificial intelligence models for cybersecurity. The support will include subsidized access to AI, training, technical assistance and partnership programs.
The new initiative is called Daybreak for Frontline Defenders. The company explained that they primarily want to help organizations responsible for critical systems but lacking cybersecurity budgets and teams of the largest corporations.
An important clarification: it is not about allocating $1 billion in direct cash grants. OpenAI estimates in this amount the subsidized access to its cybersecurity models and products, training, technical support, and partnerships.
The company plans to provide the main part of this resource within the next six months.
Who will get access
At the first stage, the program will focus on the United States. Priority will be given to operators of water and wastewater systems, power grids, state and local government bodies, regional and community banks, non-profit organizations and open-source software developers.
OpenAI plans to extend a similar support model to partner countries in the coming weeks.
With Daybreak, such organizations will be able to analyze legacy software code, investigate suspicious activity, search for and verify vulnerabilities, identify the most dangerous problems, and develop and test fixes.
The company emphasizes that organizations supporting old and complex information systems but with limited financial and human resources remain particularly vulnerable.
OpenAI expects a sharp increase in AI cyberattacks
The company explains the program launch by the rapid development of artificial intelligence capabilities. OpenAI predicts that AI-powered cyberattacks will become significantly more widespread and sophisticated in the coming months.
Powerful models can accelerate vulnerability discovery, analyze large codebases, and combine multiple system weaknesses for complex attacks. According to OpenAI, the same capabilities should also be available to defenders.
The company calls the current period a "window for defenders" — a limited time during which organizations can leverage AI to eliminate weaknesses before attackers start using similar capabilities on a mass scale.
About 2,000 organizations already connected
OpenAI earlier launched the Daybreak platform in 2026. It gives vetted public and private organizations access to advanced models for sanctioned cybersecurity operations.
Daybreak Blue is designed for defenders' everyday tasks and uses OpenAI's core models. Daybreak Red provides approved organizations with access to specialized cyber models for more complex and sensitive work.
According to OpenAI, the system is already used by thousands of professionals in approximately 2,000 approved organizations and work environments, including cybersecurity companies, defense and law enforcement structures.
In addition, through the Daybreak Defense Network, OpenAI partners are launching over 35 enterprise products and services that integrate the company's cyber models into existing tools and workflows.
After attacks on water infrastructure, assistance was already provided
OpenAI has already tested similar support after recent attacks on water systems in the United States. The company offered affected states and utilities up to $1 million in free API credits, Daybreak access, and technical assistance.
According to the company, specialists used AI to check code and system configurations, confirm issues, create patches, and verify fixes without stopping water systems.
OpenAI is also launching a pilot program with the Multi-State Information Sharing and Analysis Center to help local government and water utility specialists learn to use Daybreak for finding and fixing vulnerabilities.
The new program comes amid OpenAI's increased attention to cyber risks associated with increasingly powerful models. In August, the company already warned that the rapid growth of AI capabilities is shrinking the time during which defenders have an advantage over potential malicious actors.
