Outerlimit, a newly launched decentralised security and authorisation company built for agentic AI, has come out of stealth with $16m in pre-seed backing.
AlbionVC, Evolution Equity Partners and Crane Venture Partners supplied the capital. A group of strategic angels also joined, including Alan co-founder and CTO Charles Gorintin, who is also a co-founding advisor at Mistral, ReliaQuest founder and CEO Brian Murphy, A-lign founder and CEO Scott Price, and Goldman Sachs global head of client coverage for Global Banking and Markets Sam Morgan.
They are joined by FTV Capital partner Kyle Griswold, Twin Track Ventures partner Nicola Sinclair, Garrison founder and CEO David Garfield, whose business was bought by Everfox, and GP Bullhound co-founder and managing partner Manish Madhvani.
The raise lands as businesses hand AI agents growing access to their systems, data and APIs. According to Outerlimit, today’s tools for identity and access management, runtime enforcement and AI governance fall short of delivering the verifiable oversight and control needed for organisations to expand agent deployments without taking on unmanaged risk.
The company’s answer is to push Zero Trust principles down to the level of individual agent actions, something it says has not been done before. Its distributed design applies cryptographic enforcement at the point a tool is called.
Credentials, keys and secrets are never held in one place; instead they are split into fragments spread across the agent environment, and they can only be pieced back together when a tool is triggered. Decryption happens solely once identity, policy and execution context have all been confirmed, meaning each action is cleared before it runs.
The result, Outerlimit argues, is a model that goes further than checking who an actor is and what it may reach, giving deterministic control over precisely which actions an agent can carry out.
To help customers adopt this model gradually, Outerlimit sets out a three-stage route. The first stage uncovers agents, tools, MCP servers and shadow AI. The second delivers verifiable visibility while keeping multi-hop chains intact. The third places deterministic controls on every agent action.
Behind Outerlimit is a founding trio of Tony Pepper and Neil Larkins, the security entrepreneurs who steered Egress Software to its 2024 sale to KnowBe4, part of the Vista Equity Partners portfolio, and Dr Peter Vincent, a theoretical neuroscientist who earned his PhD through UCL’s Sainsbury Wellcome Centre and Gatsby Computational Neuroscience Unit.
The team is rethinking, from the ground up, how security should be architected for AI agents, with a stated mission of tackling the core changes needed to protect and govern agents as computing enters a new generation.
Outerlimit founder and CTO Dr Peter Vincent said, “We’ve spent decades building security systems around human qualities such as loyalty, trust, compassion, and even fear which drive predictable decision-making. In this new agentic era, where agents use reasoning models with access to tools that can impact the world, they simply don’t operate within these human constructs.
“Agents can change their behavior based on what they read, or how they interact with other agents, while acting at machine speed. Harnessing this powerful intelligence requires a fundamentally new security architecture – one that binds identity, authorization, and action into a single operation at the moment of execution.”
Outerlimit CEO Tony Pepper said, “The demand to deploy agentic AI is unparalleled, but security teams are being asked to sign off on risks they cannot adequately control. Blocking adoption isn’t a strategy, it simply drives the use of unsanctioned AI outside of enterprise oversight. Today’s launch of Outerlimit represents a paradigm shift to securing agentic AI, creating a world where AI agents can be trusted implicitly because every action is provably observed, controlled, and compliant without limiting their expressiveness.”
Vincent added, “The next phase of enterprise AI should not be a race to build the most agents. Instead, success should be measured by the ability of an organization to safely harness the full power of their agent deployments. If intelligence is to become commoditized, trust will be the limiting factor. As we accelerate into a new era of co-intelligence, getting this right is a fundamental obligation for the sake of individuals, organizations, and international security.”
Copyright © 2026 FinTech Global
