Gray Swan AI operates in the field of AI safety and security, offering automated adversarial testing, continuous red teaming, and runtime defense solutions. The company was founded by a team that emerged from Carnegie Mellon University's pioneering AI safety research group.
Its product suite comprises three main components. Shade provides automated adversarial red-teaming to identify vulnerabilities before they can be exploited. Cygnal delivers real-time runtime defense, blocking adversarial inputs and unsafe outputs. Arena, described as the world's largest adversarial AI red-teaming community, generates threat intelligence that underpins both products.
Gray Swan AI has raised $40 million in Series A funding. Its customer base includes Anthropic, OpenAI, Meta, Google DeepMind, and Snowflake, spanning both frontier AI laboratories and enterprise clients.
