Outerlimit Raises $16 Million to Secure What AI Agents Actually Do

London and New York-based Outerlimit emerged from stealth with a $16 million pre-seed round from AlbionVC, Evolution Equity Partners and Crane Venture Partners, one of the largest pre-seed rounds ever raised in cybersecurity specifically. The company was founded by Tony Pepper, Neil Larkins and Dr. Peter Vincent, combining enterprise cybersecurity experience with research in computational neuroscience.

Pepper and Larkins previously led Egress Software, a cybersecurity company that was acquired by KnowBe4 in 2024, giving the founding team a track record of building and successfully exiting a security business before starting this one. Outerlimit's core premise is that existing security architecture, built around Zero Trust principles that govern who a human user is and what systems they can access, was never designed for a world where autonomous AI agents take actions on their own inside enterprise systems, applications, data and APIs, often faster and at a higher volume than any human user ever could.

What Extending Zero Trust to Agent Actions Actually Means

Traditional Zero Trust security answers the question of identity: is this person who they claim to be, and should they have access to this system. Outerlimit is building what it describes as a decentralized security and authorization layer for the individual actions an AI agent performs once it already has access, a distinct and arguably harder problem, since an agent that has been correctly authenticated can still take an unauthorized or unintended action once inside a system, something identity-layer security alone was never built to catch. As enterprises increasingly deploy AI agents to interact directly with production systems and customer data, the gap between "this agent is allowed to be here" and "this specific action the agent is about to take is actually authorized" is exactly the gap Outerlimit is positioning itself to close.

What This Means for Founders

Outerlimit's raise is a useful case study for a founder building in any security category adjacent to a fast-moving new technology wave. The founding team did not attempt to retrofit their prior company's identity-security product to cover AI agents, they built a new company around a problem that only exists because agentic AI adoption created it, a distinction worth studying for any founder deciding whether an adjacent new risk is better addressed as a feature of an existing product or as a standalone company. The size of this pre-seed round, unusually large for the stage, also reflects how founder pedigree from a successful prior exit can compress the normal pre-seed fundraising timeline: investors underwriting Pepper and Larkins were underwriting a repeat team with a proven security exit, not a first-time founding team with an unproven idea alone.

Founders building cybersecurity, AI infrastructure, or enterprise trust and safety tooling looking for investors with an active thesis in the space can browse cybersecurity investors on AngelLinx, or explore active capital on AngelLinx @ angellinx.ai/register.

AngelLinx Intelligence | angellinx.ai