NATO Burns 300B Tokens A Month On OpenAI’s Cyber Defense AI
OpenAI says NATO is now feeding up to 300 billion tokens a month through its cyber defense AI, a sign that the world’s biggest military alliance is putting frontier models on the front line of protecting its networks.
Key Takeaways
- OpenAI works with NATO’s Communications and Information Agency to give NATO core institutions access to advanced cybersecurity models
- NATO has used up to 300 billion raw input and output tokens per month in that work
- ENISA identified 4 vulnerabilities in software used across EU institutions, with fixes published for all 4
- CERT Polska helped identify 6 vulnerabilities in MikroTik RouterOS, and MikroTik released fixes that CERT Polska confirmed blocked observed attacks
Working with NATO’s Communications and Information Agency, OpenAI announced access for NATO core institutions to advanced cybersecurity models that help protect Allied networks.
Daybreak is OpenAI’s program that puts advanced AI in the hands of trusted defenders for authorized security work, including reviewing software, investigating suspicious activity, validating vulnerabilities, and developing and testing fixes.
How NATO Is Using Daybreak
NATO has used up to 300 billion raw input and output tokens per month as part of this work. OpenAI linked that early use to NATO’s mission of safeguarding the freedom and security of its member states.
It also described the effort as a way to strengthen cyber expertise across multinational teams that protect Allied networks around the clock and to create lessons for defenders across Europe and the broader Alliance.
Also Read: OpenAI Helps Expose Flaws In Website Software Used By Millions
How Allied Agencies Used It
Beyond NATO, OpenAI has provided access to its cyber models to the EU’s ENISA, France’s ANSSI, and national cyber agencies in Germany and Poland. Ukraine’s Ministry of Digital Transformation is using Daybreak to strengthen the cyber defense of civilian infrastructure.
A November joint report from CERT-EU and ENISA described a review of 27 software projects using frontier models, including OpenAI models accessed through Daybreak, and tools including the Codex security plug-in. Across the models tested, 95% of the 88 potential issues assessed by software maintainers were confirmed as genuine.
CERT Polska used OpenAI models to help identify six vulnerabilities in MikroTik RouterOS and coordinated disclosure with the vendor. Researchers directed the work, tested findings on real systems and verified their impact, while AI accelerated analysis. OpenAI also supported the 2026 Crossed Swords exercise, organized by NATO’s Cooperative Cyber Defence Centre of Excellence, where participants built, deployed, and evaluated AI agents for cyber operations with an emphasis on human oversight.
What To Watch Next
OpenAI plans to keep working with governments, national cyber agencies, and international partners to turn frontier AI capability into stronger, safer, and more resilient systems. The effort is aimed at defenders protecting Allied networks, EU institutions, and civilian infrastructure by helping them find weaknesses and test fixes faster.
Read Next: OpenAI Finds 948 Linux Bugs That Could Give Hackers Control
