OpenAI gives NATO and allied cyber agencies access to Daybreak AI
OpenAI is providing NATO and allied cyber agencies with access to its Daybreak cybersecurity models, supporting work that includes vulnerability discovery, suspicious-activity investigations and software fixes. NATO has used up to 300 billion raw input and output tokens per month as part of the collaboration, according to OpenAI.
- NATO has used up to 300 billion raw input and output tokens per month in the work.
- OpenAI supported the 2026 Crossed Swords exercise with its Daybreak tool.
- ENISA used OpenAI models to identify four vulnerabilities, with fixes published for all four.
- CERT Polska used OpenAI models to help identify six MikroTik RouterOS vulnerabilities.
- Researchers and software maintainers remained involved in testing and validating findings.
Daybreak expands into NATO cyber defence
OpenAI is working with NATO’s Communications and Information Agency, known as NCIA, to give NATO’s core institutions access to advanced cybersecurity models. The company says the tools are intended for authorized security work, including reviewing software, investigating suspicious activity, validating vulnerabilities, and developing and testing fixes.
According to OpenAI, NATO has used up to 300 billion raw input and output tokens per month as part of the work. The company says the collaboration is helping multinational teams responsible for protecting Allied networks around the clock strengthen their cyber expertise and develop lessons that could benefit defenders across Europe and the broader Alliance.
The deployment is part of OpenAI’s Daybreak initiative, which puts advanced AI tools in the hands of trusted defenders. OpenAI says its work with NATO supports the Alliance’s mission of safeguarding the freedom and security of its member states while maintaining a focus on safe and secure adoption.
AI agents tested in NATO exercise
OpenAI also supported the 2026 Crossed Swords exercise, organized by the NATO Cooperative Cyber Defence Centre of Excellence, or CCDCOE. The exercise included a new training stream in which participants built, deployed and evaluated AI agents for cyber operations in a technical environment using OpenAI Daybreak.
The training placed participants alongside AI agents as operators, with an emphasis on maintaining human oversight. That approach reflects the source’s description of AI as a tool for cyber defenders rather than a replacement for the people directing, testing and evaluating security work.
Dr Dylan Browne, NCIA General Manager, said in the OpenAI announcement that AI-powered cyber defence models “accelerate vulnerability discovery, detect emerging threats at machine speed and enable defenders to stay ahead of adversaries.” Browne added that NATO would build a deeper understanding of token consumption, usage patterns and the underlying economics of AI models as adoption scales.
European agencies report vulnerabilities found with AI
OpenAI says it has also provided cyber models to defenders in Europe, including the EU’s ENISA and national cyber agencies in France, Germany and Poland. Ukraine’s Ministry of Digital Transformation is using Daybreak to help strengthen the cyber defence of civilian infrastructure, with the stated aim of identifying software vulnerabilities and developing and testing fixes more quickly.
ENISA used OpenAI’s models to identify four vulnerabilities in software used across EU institutions. OpenAI says fixes were published for all four. A November joint report from ENISA and CERT-EU reviewed 27 software projects using frontier models, including OpenAI models accessed through Daybreak, as well as tools including the Codex security plug-in.
Across the models tested in that review, software maintainers confirmed 95% of the 88 potential issues assessed as genuine. OpenAI presents the result as evidence of the value of combining AI with human expertise.
CERT Polska used OpenAI models to help identify six vulnerabilities in MikroTik RouterOS and coordinated their disclosure with the vendor. Researchers directed the work, tested findings on real systems and verified their impact. OpenAI says MikroTik released fixes, which CERT Polska confirmed prevented the attacks it had observed.
What the announcement means for cyber defenders
The announcement places Daybreak within operational cybersecurity work carried out by NATO, EU institutions and national agencies. The examples described by OpenAI focus on helping defenders find weaknesses, assess their impact and move more quickly toward fixes, while people remain responsible for directing the work and validating its results.
Sasha Baker, OpenAI’s Head of National Security Policy, said the company was “putting more capable tools in their hands to help them find and fix vulnerabilities before attackers can exploit them.” OpenAI said it will continue working with governments, national cyber agencies and international partners to support stronger, safer and more resilient systems.
OpenAI: Strengthening NATO and allied cyber defenses with AI


