New York Startup Outerlimit Raises $16M to Tame Rogue AI Agents

Outerlimit, a New York-based cybersecurity startup, has secured $16 million in pre-seed funding. The round was led by AlbionVC, Evolution Equity Partners, and Crane Venture Partners, with angel investors from Mistral AI and Goldman Sachs participating.

The company's founders argue that traditional identity and access management systems are inadequate for the current reality of AI agents, which can spawn sub-agents, chain together tool calls, and adapt their behavior without human oversight. Outerlimit's platform is designed to prevent these rogue agents from causing harm by detecting, monitoring, and enforcing policies in real-time.

The platform uses a distributed architecture with cryptographic enforcement, fragmenting sensitive access credentials throughout the ecosystem. These fragments are reassembled only when three conditions are simultaneously verified: the identity of the agent requesting access, the policy governing what that agent is allowed to do, and the specific execution context of the action being attempted.

The founding team brings credibility to the bet, with Tony Pepper and Neil Larkins having established track records in the tech industry, while Dr. Peter Vincent brings expertise in theoretical neuroscience.