NVIDIA and CrowdStrike Strengthen Agentic Cybersecurity Frontier

CrowdStrike and NVIDIA have announced a new agentic cybersecurity system called SafeMind, which aims to address the growing threat of automated attacks. At CrowdStrike's annual conference, NVIDIA CEO Jensen Huang and CrowdStrike CEO George Kurtz unveiled SafeMind, which combines CrowdStrike's "beyond frontier-capable" models with customized agentic harnesses, powered by NVIDIA's Nemotron technology.

SafeMind operates in a continuous coevolution loop, where offensive and defensive AI models continuously challenge and improve each other, hardening the customer's security environment. SafeMind's defensive model was post-trained with CrowdStrike's cyber experience and threat data using NVIDIA Nemotron open models, resulting in higher accuracy rates than leading frontier models at a lower cost.

The system can be used independently or in combination with customers' own models, providing flexibility for different environments. With AI-powered attacks on the rise, CrowdStrike emphasizes the need for human-speed response, stating that AI-enabled defense is necessary to keep up with the increasing pace of eCrime.