CrowdStrike, a US cybersecurity company, said at its September 1 Fal.Con 2026 event that it launched SafeMind, a family of AI security models and harnesses built with Nvidia for use in its Falcon platform.

It plans to provide trusted access to standalone models and harnesses through Project QuiltWorks.

The release includes Red Tempest for offensive testing and Blue Solano for defense.

CrowdStrike said it trained SafeMind on Falcon telemetry, threat intelligence, Falcon Complete managed detection and response event annotations, and incident-response fieldwork.

CoreWeave, a US cloud infrastructure provider focused on AI workloads, supplied cloud infrastructure for training and inference.

CrowdStrike said its evaluations against leading frontier models and open-source baselines showed a 29% higher detection rate, 6x faster end-to-end remediation, and 99% lower detection and remediation costs.

The launch adds to the AI-driven security operations center market, which has already emerged as a multi-vendor category.

It also comes as enterprises weigh data-security risks around AI systems that handle sensitive operational records.

🔗 Source: CrowdStrike

Recent CrowdStrike developments