OpenAI Launches Project Daybreak, Pledging $1 Billion for AI Cybersecurity Defense to Face Anthropic Mythos Competition

TradingKey - On September 3, Eastern Time, OpenAI announced the launch of its "Daybreak for Frontline Defenders" global initiative, aiming to provide frontline cybersecurity defenders with cutting-edge AI technology support to help protect critical infrastructure in the U.S. and around the world.

[Source: OpenAI]

OpenAI noted in a statement that as AI model capabilities advance rapidly, AI-driven cyberattacks could become more prevalent and sophisticated. Defenders need to identify vulnerabilities and harden systems before attackers do, and frontier AI technology can help defenders complete these tasks more quickly.

In the field of AI cybersecurity, OpenAI and Anthropic have entered into competition. On April 7, 2026, Anthropic released the Claude Mythos Preview model and simultaneously launched Project Glasswing. The model is accessible exclusively to Project Glasswing partners, with an initial cohort of around 50 organizations gaining access, including tech companies such as AWS, Apple (AAPL), Microsoft (MSFT), Google (GOOGL), and Nvidia (NVDA).

Within weeks of the project's launch, partners had identified more than 10,000 high-risk or critical security vulnerabilities. Evaluations by the UK AI Safety Institute showed that Mythos achieved a 73% success rate in expert-level CTF tasks and became the first AI model to complete a 32-step enterprise cyberattack simulation from start to finish. In June 2026, the project was further expanded to about 150 new organizations across more than 15 countries, covering multiple critical infrastructure sectors.

OpenAI officially launched Daybreak-related cybersecurity tools on June 22, 2026, and further expanded Daybreak on August 10 with the rollout of GPT-5.6-Cyber.

The "Daybreak for Frontline Defenders" initiative launched in September builds on the existing platform to further expand support for frontline cybersecurity defenders.

Currently, Daybreak features two access pathways, Blue and Red: Blue is designed for most defensive cybersecurity work, utilizing GPT-5.6 Sol; Red is aimed at more sensitive, advanced authorized security testing, offering GPT-5.6-Cyber and requiring additional approvals.

This initiative includes three core measures:

1. A commitment to provide $1 billion in Daybreak subsidy support over the next six months to offer access, training, and technical support for resource-constrained cybersecurity defenders;

2. The launch of the "Daybreak for America" program, consolidating OpenAI's efforts in the U.S. with a focus on supporting critical systems such as water and wastewater systems, power grid operators, state and local governments, and community and regional banks, alongside initiating a pilot program with the Multi-State Information Sharing and Analysis Center (MS-ISAC);

3. Expanding Daybreak capabilities to more than 35 enterprise products and partner-operated services through the Daybreak Defense Network.

OpenAI stated that no single company can secure the critical systems that people rely on alone. This initiative will prioritize institutions with limited security resources, including water and wastewater systems, power grid operators, state and local governments, community and regional banks, non-profit organizations, and open-source maintainers. These teams often need to protect complex, legacy systems while lacking the budgets and expertise enjoyed by large corporations.

In terms of deployment progress, thousands of defenders from 2,000 approved organizations and workspaces are currently using Daybreak, spanning cybersecurity firms, defense organizations, and law enforcement agencies.

Previously, following attacks on U.S. water systems, OpenAI provided up to $1 million in free API credits, Daybreak access, and technical assistance to affected states and utilities.

Meanwhile, OpenAI has partnered with the MS-ISAC to launch a pilot program, initially providing a cohort of public sector and water system defenders with Daybreak access, training, and hands-on support. The MS-ISAC provides cyber threat intelligence and incident response support to thousands of public sector entities.

OpenAI emphasized that discovering vulnerabilities is only the first step; more critical is shortening the time from vulnerability discovery to remediation. Eligible state and local governments, critical infrastructure operators, non-profits, and open-source maintainers can visit the Daybreak website to learn about access, technical support, and training arrangements.