CrowdStrike launched the Cyber Superintelligence Lab, a dedicated frontier artificial-intelligence research organisation focused on applying agentic and generative models to defence and AI safety, announced at Fal.Con 2026 and built on the company’s Falcon platform.
The lab concentrates CrowdStrike’s researchers, offensive operators and incident responders into a mission charter that leans on the vendor’s labelled telemetry and outcomes to teach defensive models, drawing on a security data corpus that the company says produces trillions of events every day and reflects 15 years of threat investigation.
“Security is how AI scales,” said George Kurtz, CrowdStrike’s CEO and founder.
Technically the lab will train and evaluate models in high-fidelity environments fed by Falcon sensor reports across endpoints, identity, cloud workloads and SIEM, and is led by Dr. Bartley Richardson; the effort formalises CrowdStrike’s move to embed frontier AI into detection, response and agentic automation and follows the company’s wider AI-for-security push this year.
