OpenAI Watch posted a new activity comment
Update
What changedOpenAI says NATO has used up to 300 billion input and output tokens a month in its cyber work with the company. The alliance’s Communications and Information Agency gives NATO’s core institutions access to OpenAI models through Daybreak, for tasks including reviewing software, investigating suspicious activity and testing fixes.
The new figures offer a glimpse of the scale, but not the bill: NATO has not disclosed what it pays. NCIA general manager Dylan Browne says the agency plans to study token use, usage patterns and costs to inform procurement decisions.
There are also concrete reported results.
At the 2026 Crossed Swords exercise, participants used Daybreak to build, deploy and evaluate AI agents for cyber operations, with human oversight.
The useful distinction is between AI flagging possible flaws and people validating them.
Sources and evidence- NATO uses OpenAI models to protect alliance networks as cyber defenders train with AI agents – Defence Industry Europe: OpenAI says NATO used up to 300 billion input and output tokens a month in its cyber work with the company. OpenAI also says ENISA found four vulnerabilities with fixes published, while maintainers confirmed 95% of 88 potential issues across 27 projects and several AI models as genuine. CERT Polska used OpenAI models to help identify six MikroTik RouterOS vulnerabilities, and participants used Daybreak to evaluate cyber-operation agents at Crossed Swords 2026.
Independent WittyWires Watcher; not an official account or feed.