CrowdStrike와 NVIDIA가 프론티어 AI 모드를 사용하여 방어를 자동화하는 에이전트 기반 사이버보안 시스템 SafeMind를 발표했습니다.
At CrowdStrike's Fal.Con 2026 conference in Las Vegas on Tuesday, NVIDIA founder and CEO Jensen Huang told a sold-out crowd of 10,000 security professionals that cybersecurity has reached an inflection point. Attacks are now automated, which means defense must be automated too. Huang joined CrowdStrike CEO and founder George Kurtz to announce SafeMind, an agentic cybersecurity system developed by the CrowdStrike Cyber Superintelligence Lab.
SafeMind combines CrowdStrike's purpose-built, beyond frontier-capable models and customized agentic harnesses with defensive models built on NVIDIA Nemotron in a continuous coevolution loop where offense and defense repeatedly challenge and improve each other. CrowdStrike also announced CrowdStrike Falcon IQ to operationalize Project QuiltWorks through agentic workload automation and expanded its CrowdStrike Guardian AI safety solution.
Huang emphasized the community advantage, telling the crowd: "We have asymmetric advantages because we have a large community of cybersecurity experts who want to work with each other and keep the world safe." Kurtz added that the real gap was that attackers had frontier AI while defenders did not. "And that changes now," he said.
CrowdStrike built SafeMind's defensive model using NVIDIA Nemotron open models, post-trained with CrowdStrike's cyber experience and threat data. The models are paired with proprietary cybersecurity harnesses optimized to work as an agentic stack and ship natively in the CrowdStrike Falcon platform. The system brings offensive and defensive AI together where each side adapts to and strengthens the other, continuously hardening customer environments until attacks are unsuccessful. Huang told Kurtz: "Your decade and a half of security data that we can train on — we can take a frontier model and make it essentially a super AGI that is incredibly good at cybersecurity."
NVIDIA Nemotron 3 Ultra orchestrates the defensive agent harness, while a fine-tuned Nemotron 3 Super powers SafeMind's rule-generation sub-agent. By post-training Nemotron with CrowdStrike data, CrowdStrike's internal evaluations showed that the Blue Solano model, based on Nemotron 3 Super, delivered higher accuracy rates than leading frontier models at 99% lower cost. While SafeMind can operate as a complete system, the models can be used independently, and security experts can pair their own models with CrowdStrike's custom harnesses for the flexibility to use the right capabilities for their environment.
Data shows that AI-enabled attacks rose 89 percent in the past year, and the fastest eCrime breakout time has reached 27 seconds. Human-speed response is no longer defense but documentation. Huang explained that Nemotron was created for post-training and fine-tuning, completely free and incredibly fast, giving organizations the ability to have an asymmetric advantage. With open Nemotron as the base, CrowdStrike's security teams post-trained on their own threat data without sending it to an outside provider, customizing the AI to their specific environment. That's not possible with a closed frontier model, and in security, the ability to inspect what's defending matters.
NVIDIA announced its work testing the CrowdStrike SafeMind models and harnesses in a high-fidelity cyber agent environment running as a simulation of the NVIDIA network. The testing runs SafeMind in an offensive-defensive loop for adversarial coevolution. An offensive red-team agent finds exploits, a blue-team defensive agent closes them, and the findings become actionable detections to block attacks. The red-agent harness runs Recon, Assault and Compromise sub-agents executing attack paths inside the cyber agent environment, while the blue-agent harness monitors via Falcon sensors, generates detection candidates, validates them and promotes them. CrowdStrike built the test environment with NVIDIA as a digital twin of NVIDIA's own accelerated computing infrastructure, validated against NVIDIA's real threat landscape.
Huang noted that the basic framework of SafeMind, an adversarial model acting on a digital twin of the environment with a defender model in a continuous cat-and-mouse loop eventually learning how to secure itself, applies to robotics, edge computing, enterprise computing and just about everything.
CrowdStrike also announced Falcon IQ, where NVIDIA Nemotron models power the agentic engine at the heart of Charlotte AI AgentWorks, CrowdStrike's no-code agent development platform. Falcon IQ uses more than 50 agents working together as a unified agentic workforce to automate the most time-intensive workflows in assessment, prioritization and remediation. Partners use Falcon IQ to deliver customized findings, recommendations and executive outputs to customers, and Charlotte AI AgentWorks enables every Falcon user to build their own agentic security workforce.
CrowdStrike processes trillions of daily security events from thousands of customer organizations. With NVIDIA's full-stack accelerated computing platform, the collaboration runs from the chips up through the models to the harnesses acting on what those models find. Kurtz emphasized that the crowd in CrowdStrike is the asymmetry that puts defenders in a unique position to defeat the adversary.