CrowdStrike Launches AI Security Models With Nvidia

CrowdStrike unveiled SafeMind, a family of security-focused AI models built with Nvidia on the Nemotron architecture, trained on 15 years of incident response data and Falcon sensor telemetry. Two models anchor the system: Red Tempest simulates AI-driven attacks while Blue Solano handles defense, running in a closed loop to continuously improve each other. The company also launched the Cyber Superintelligence Lab to house its AI researchers and security operators. CrowdStrike claims SafeMind delivers 29% higher detection rates, six times faster remediation, and 99% lower costs compared to unnamed baseline models. The Falcon platform was also extended across Google Cloud's enterprise AI ecosystem.