$4m · Pre-Seed · Cybersecurity · London, UK
CodeWall has emerged from stealth with a $4m pre-seed round led by Osney Capital, with participation from InReach Ventures, Antler and angel investors. The London cybersecurity company says the capital will widen the scope of its autonomous attack platform and fund hiring across offensive security, engineering, customer success and go-to-market.
From periodic pentest to persistent attacker
Traditional penetration tests usually examine a bounded environment at a point in time. CodeWall is building software agents intended to keep probing as systems change. According to the company, its agents start outside a target without credentials, source code or privileged access, then work in parallel and share discoveries through a persistent investigation graph. That structure is designed to help the system chain weaknesses into attack paths, preserve context between tests and retest fixes inside an authorised scope.
The commercial shift is from buying an occasional assessment to maintaining an ongoing adversarial-testing workflow. CodeWall says it is already working with banks and other financial institutions. Axos Financial chief information security officer John Kraft said the system shows how vulnerabilities can be combined into attack paths and supplies evidence to help teams prioritise remediation.
Proof and control are the buying test
CodeWall points to research in which its agents found vulnerabilities in platforms operated by McKinsey, BCG and Bain. Those results are company-reported, including its account of obtaining access to a database connected to McKinsey's Lilli platform within two hours; affected organisations were notified and the company says remediation followed.
For enterprise buyers, finding a weakness is only part of the job. Autonomous testing also has to demonstrate clear scope controls, reproducible evidence, useful prioritisation and reliable retesting without creating operational risk. The round gives CodeWall room to broaden coverage and connect discovery more closely with verification and remediation. Execution will depend on turning striking demonstrations into a dependable security process that customers can govern.
Explore this theme
When AI attacks, can AI defend? — Ahmed Achchak, Qevlar AI, Réza Malekzadeh & Partech, hosted on EUVC · 17 July 2026. The discussion shows how AI changes attack economics and why enterprise trust depends on transparent controls, extending CodeWall's challenge of making autonomous testing both useful and governable.
Sources checked:
CodeWall announcement · Tech.eu Funding Explorer
