Google DeepMind has launched the Fairwind Program, a limited access initiative for governments, enterprise partners, and trusted cybersecurity organizations to use advanced AI-driven cyber defense capabilities. The program provides early access to Gemini 3.8 Flash Cyber, Google’s most advanced cyber model, combined with the Code Mender harness to autonomously find, verify, and fix software vulnerabilities at scale. Defenders can now generate verified, deployment-ready patches in minutes instead of weeks, all within their secure cloud environments.

The program targets governments, national cyber authorities, critical infrastructure operators (healthcare, energy, telecom, finance), and core technology platforms to harden systems before exploitation. Participants must adhere to strict operational standards, including internal access controls and multi-factor authentication. Over 650 global partners are currently enrolled, with feedback indicating real-world impact in vulnerability remediation.

While Fairwind prioritizes access for qualifying entities, Google Cloud customers can also use Code Mender with publicly available models via the Gemini Enterprise Agent Platform alongside AI Threat Defense solutions. The initiative aims to scale proactive defense across ecosystems, leveraging Google’s zero-trust architecture and AI innovations to protect critical infrastructure and national security.