Gray Swan

AI security research company that stress-tests and hardens frontier models.

Gray Swan AI runs adversarial research on advanced AI systems. It operates large public red-teaming competitions, builds automated attack tooling, and develops defenses that make models more resistant to jailbreaks and misuse. Frontier labs and enterprises use its findings and benchmarks to gauge how robust their models really are.

Category

AI Model Security & Red Teaming

Founded

2024

Funding

$40M raised (Series A, 2026)

Website

https://www.grayswan.ai

Seen at Black Hat USA 2026 (AI Zone sponsor, booth 5407).

Outbound clicks are tokenized for directory analytics.