NVIDIA and CrowdStrike partner at Fal.Con 2026 to launch SafeMind agentic defense
The partnership introduces specialized models like Blue Solano and a 'continuous coevolution loop' where offensive and defensive AI agents train each other in digital twins.
Original cover art, generated for this story. THE VISSION does not republish third-party press imagery.
- NVIDIA and CrowdStrike announced a major expansion of their partnership at CrowdStrike's Fal.Con 2026 conference to counter automated AI threats, according to NVIDIA's official blog.
- The collaboration features CrowdStrike SafeMind, an agentic security system utilizing NVIDIA's Nemotron 3 Ultra and Nemotron 3 Super models to automate complex assessment and patching.
- A new domain-specific model, Blue Solano, was trained on 15 years of threat intelligence to generate security rules, achieving higher accuracy than general frontier models at a 99% lower cost.
At CrowdStrike's Fal.Con 2026 conference in Las Vegas, NVIDIA CEO Jensen Huang and CrowdStrike CEO George Kurtz announced a major expansion of their partnership to launch an agentic security framework, according to NVIDIA's official blog. The initiative aims to deploy autonomous AI agents to counter increasingly rapid, automated cyberattacks, noting that eCrime breakout times have dropped to a record low of 27 seconds.
The centerpieces of the announcement are CrowdStrike SafeMind and Falcon IQ. SafeMind is described as the industry’s first complete agentic cybersecurity system, built natively into the Falcon platform. The system operates on a 'continuous coevolution loop' where autonomous offensive and defensive agents challenge and harden each other within digital twin environments. It utilizes NVIDIA's Nemotron 3 Ultra model to orchestrate defensive agent workforces and automate complex assessment and remediation across enterprise systems.
The partnership also introduced Blue Solano, a specialized security model based on the NVIDIA Nemotron 3 Super architecture. Post-trained on CrowdStrike's 15 years of proprietary threat intelligence, Blue Solano is designed to generate highly accurate security rules. NVIDIA claims the domain-specific model achieves superior accuracy on security tasks compared to general-purpose frontier models, while operating at a 99% lower cost. Additionally, the partnership launched Charlotte AI AgentWorks, a no-code development platform for creating agentic security workflows, and CrowdStrike Guardian for AI safety deployment.
As hackers adopt automated AI tools to mount lightning-fast attacks, human security teams can no longer react quickly enough to contain them. SafeMind represents a shift from static passive detection to active, autonomous agentic defense. By running continuous, adversarial training loops in digital twins and offering highly optimized, low-cost specialized models like Blue Solano, NVIDIA and CrowdStrike are establishing a blueprint for how enterprises can deploy agentic AI to protect their critical infrastructure at machine speed.