OpenAI Collaborates with NATO and European Agencies on AI-Driven Cyber Defense
NATO is consuming up to 300 billion input and output tokens monthly through a partnership with OpenAI to enhance the cybersecurity of allied networks.
- NATO utilizes up to 300 billion input and output tokens monthly through its partnership with OpenAI.
- The collaboration involves the NATO Communications and Information Agency for authorized security work and software testing.
- European agencies including ENISA and Poland's CERT Polska used the models to discover and patch software vulnerabilities.

The North Atlantic Treaty Organization is utilizing a significant volume of artificial intelligence computing resources to protect its digital infrastructure. According to official disclosures, the military alliance consumes up to 300 billion input and output tokens every single month as part of an ongoing collaborative initiative with OpenAI.
Partnership with NATO Communications and Information Agency
The deployment centers around a cooperative arrangement with the NATO Communications and Information Agency. Core institutions within the military alliance have been granted direct access to advanced cybersecurity models developed by the artificial intelligence firm. Authorized security personnel are deploying these computational tools for critical tasks such as examining software code, analyzing unusual network activity, pinpointing system weaknesses, and formulating software patches.
Dr. Dylan Browne, General Manager of the agency, characterized the integration of artificial intelligence models as a major technological advancement for digital protection. He noted that automated systems accelerate the discovery of system weaknesses and identify emerging threats at machine speed, allowing defensive teams to outpace potential adversaries. Furthermore, Browne stated that monitoring token consumption and usage patterns provides the organization with vital data regarding the economics of artificial intelligence deployment, which will guide future procurement choices and cost optimization strategies.
Training Exercises and Regional Security Collaborations
Beyond standard network monitoring, the artificial intelligence company extended its support to specialized military training programs. The organization contributed to the 2026 Crossed Swords exercise, which is administered by the NATO Cooperative Cyber Defence Centre of Excellence. This program featured specialized training modules where participating specialists constructed, deployed, and assessed automated agents designed for digital operations, keeping strict human oversight at the center of the activities.
The scope of these security deployments extends beyond the military alliance itself. Similar advanced models have been extended to several European government entities and multinational cybersecurity authorities, including the European Union Agency for Cybersecurity alongside national cybersecurity directorates located in Poland, Germany, and France.
Documented Vulnerability Discoveries Across Europe
European Union institutions have already recorded concrete outcomes from utilizing these advanced models. The European Union Agency for Cybersecurity deployed the tools to detect four separate weaknesses in software utilized across community institutions. Official patches have since been released for all four identified points of exposure. Additionally, a collaborative evaluation published jointly by the agency and CERT-EU noted that software maintainers verified 95 percent of 88 potential software issues identified during testing as genuine vulnerabilities.
National entities have experienced similar operational successes. CERT Polska, operating within Poland, utilized the models to uncover six distinct vulnerabilities within MikroTik RouterOS. Independent researchers subsequently verified these discoveries before reporting them directly to the vendor, which ultimately published software updates to resolve the issues.
Corporate Expansion and Strategic Direction
These government-level deployments coincide with a broader commercial expansion by the artificial intelligence developer. The enterprise recently rolled out two distinct model iterations: GPT-6.2 Astra, engineered specifically for professional environments, and GPT-6.2 Sol, tailored toward everyday computational tasks with an emphasis on operational efficiency.
Sasha Baker, who serves as the Head of National Security Policy at the artificial intelligence firm, emphasized that the organization intends to expand its collaborative footprint alongside multinational allies. The stated objective remains centered on assisting government defenders in discovering vulnerabilities and creating defensive fixes at an accelerated pace to secure critical digital infrastructure.
- NATO uses up to 300 billion AI tokens monthly for cybersecurity tasks.
- ENISA and CERT Polska identified software vulnerabilities using OpenAI models.
- OpenAI recently introduced GPT-6.2 Astra and GPT-6.2 Sol for professional and everyday use.
Spotted an error? · Our editorial policy
UP NEXTDadasaheb Phalke International Film Festival Awards 2026: Taylor Swift and James Cameron Among WinnersOther angles · #Artificial Intelligence
Timeline →- NewsAnthropic Researcher Claims AI-Assisted Proof for the Riemann Hypothesis
- NewsTrump Backs Jensen Huang’s Return to Nvidia Amid Board Resistance and Intelligence Scrutiny
- NewsTrump Discusses Nuclear Buildup and AI Safeguards Following Trilateral Summit with Xi and Putin
- NewsTrump Threatens Heavy Taxes on US Firms Constructing Data Centers Overseas
More from Technology

Anthropic Researcher Claims AI-Assisted Proof for the Riemann Hypothesis

Grand Theft Auto VI Becomes Largest Entertainment Launch in History

Trump Backs Jensen Huang’s Return to Nvidia Amid Board Resistance and Intelligence Scrutiny
