NVIDIA and CrowdStrike Strengthen Agentic Cybersecurity Frontier

At CrowdStrike’s Fal.Con 2026, NVIDIA CEO Jensen Huang and CrowdStrike CEO George Kurtz announced CrowdStrike SafeMind, an agentic cybersecurity system developed by the CrowdStrike Cyber Superintelligence Lab. SafeMind integrates purpose-built models with NVIDIA Nemotron to create a continuous coevolution loop where offensive and defensive AI challenge and improve each other. The system is designed to harden customer environments by adapting to emerging threats in real time.

CrowdStrike also introduced CrowdStrike Falcon IQ, which uses agentic workload automation to operationalize Project QuiltWorks, and expanded its CrowdStrike Guardian AI safety solution. SafeMind’s defensive models, built on NVIDIA Nemotron open models post-trained with CrowdStrike’s threat data, are paired with proprietary cybersecurity harnesses optimized for an agentic stack. The system ships natively within the CrowdStrike Falcon platform.

NVIDIA Nemotron 3 Ultra orchestrates the defensive agent harness, while a fine-tuned Nemotron 3 Super powers SafeMind’s rule-generation sub-agent. CrowdStrike’s internal evaluations showed the Blue Solano model, based on Nemotron 3 Super, achieved higher accuracy rates than leading frontier models at 99% lower cost. The models can operate independently or be paired with custom harnesses for flexibility in customer environments.

CrowdStrike and NVIDIA tested SafeMind in a high-fidelity cyber agent environment simulating NVIDIA’s network, using a digital twin of its infrastructure. The testing framework involves adversarial coevolution, where offensive red-team agents identify exploits and defensive blue-team agents close them, generating actionable detections. The approach is designed to secure environments through continuous learning and adaptation.