NVIDIA and CrowdStrike Strengthen Agentic Cybersecurity Frontier
Built using NVIDIA Nemotron open models, the platform pairs defensive AI with CrowdStrike threat data to counter automated cyber threats. Internal testing indicates that the system offers superior accuracy while significantly reducing costs. At the Fal.Con 2026 event in Las Vegas, NVIDIA CEO Jensen Huang and CrowdStrike CEO George Kurtz unveiled CrowdStrike SafeMind, an agentic cybersecurity system designed by the CrowdStrike Cyber Superintelligence Lab.
Key Takeaways
- Built on NVIDIA Nemotron open models, SafeMind combines customized defensive harnesses with cybersecurity threat data.
The platform operates a continuous coevolution loop that matches offensive and defensive AI against one another to harden customer security environments against automated attacks.
- According to internal evaluations, CrowdStrike's Blue Solano model achieved higher accuracy rates than leading frontier models while operating at 99% lower cost.
Furthermore, CrowdStrike introduced Falcon IQ for workload automation and expanded its Guardian AI safety solution to address an 89% surge in AI-enabled attacks over the past year.
- Internal evaluations showed the Blue Solano model achieved higher accuracy than leading frontier models at 99% lower cost.
SafeMind utilizes a continuous coevolution loop where offensive and defensive AI models repeatedly test and strengthen each other.
- NVIDIA Nemotron 3 Ultra coordinates the defensive harness, while a fine-tuned Nemotron 3 Super powers rule-generation sub-agents.
- The technical stack uses NVIDIA Nemotron 3 Ultra to orchestrate the defensive agent harness, alongside a fine-tuned Nemotron 3 Super model that runs rule-generation sub-agents.
Stats & Key Facts
- #According to internal evaluations, CrowdStrike's Blue Solano model achieved higher accuracy rates than leading frontier models while operating at 99% lower cost.
- #Furthermore, CrowdStrike introduced Falcon IQ for workload automation and expanded its Guardian AI safety solution to address an 89% surge in AI-enabled attacks over the past year.
- #Internal evaluations showed the Blue Solano model achieved higher accuracy than leading frontier models at 99% lower cost.
At the Fal.Con 2026 event in Las Vegas, NVIDIA CEO Jensen Huang and CrowdStrike CEO George Kurtz unveiled CrowdStrike SafeMind, an agentic cybersecurity system designed by the CrowdStrike Cyber Superintelligence Lab. Built on NVIDIA Nemotron open models, SafeMind combines customized defensive harnesses with cybersecurity threat data. The platform operates a continuous coevolution loop that matches offensive and defensive AI against one another to harden customer security environments against automated attacks.
The technical stack uses NVIDIA Nemotron 3 Ultra to orchestrate the defensive agent harness, alongside a fine-tuned Nemotron 3 Super model that runs rule-generation sub-agents. According to internal evaluations, CrowdStrike's Blue Solano model achieved higher accuracy rates than leading frontier models while operating at 99% lower cost. Furthermore, CrowdStrike introduced Falcon IQ for workload automation and expanded its Guardian AI safety solution to address an 89% surge in AI-enabled attacks over the past year.
Internal evaluations showed the Blue Solano model achieved higher accuracy than leading frontier models at 99% lower cost. SafeMind utilizes a continuous coevolution loop where offensive and defensive AI models repeatedly test and strengthen each other.
For more details please read the original article at NVIDIA Blog.
Continue Learning
Comments
Comments appear only after moderation. Your email identifies your submission to the moderator and is never displayed here.
No approved comments yet.