CodeWall, a London-based offensive security startup, officially launched with $4 million in funding, SiliconANGLE's Duncan Riley reported on October 14. Osney Capital led the pre-seed round, with InReach Ventures, Antler Global and angel investors participating.
What CodeWall does
It provides autonomous AI agents that attack enterprises, such as banks, the way an outside hacker would. According to the company, engagements begin with no credentials or inside knowledge of the target, and several agents, each built for a different task, work at the same time. Findings are logged to a shared investigation graph, so a weakness in one application can become a route into connected systems and data. Customers see each completed attack path step by step with evidence attached, and agents retest when new systems come online or fixes are applied while staying within the customer's authorized scope. CodeWall positions this ongoing testing as a replacement for penetration tests that enterprises commission once or twice a year.
Earlier research
The company says its agents targeted the Big Three consulting firms earlier in 2026. At McKinsey, an agent with no credentials reportedly gained read and write access to the production database behind the Lilli AI platform within two hours, and research also reported exposed databases at Boston Consulting Group and Bain. Each firm fixed its flaws within days. A McKinsey spokesperson told The Register that a third-party forensics investigation found no evidence that CodeWall or any other unauthorized party accessed client data. These are company claims.
What executives said
Founder and CEO Paul Price said enterprises “shouldn't have to wait for a real attacker to find out whether their defenses hold,” and that each customer gets “an attacker that keeps looking for a way in.” Osney Capital partner Paul Wilkes said having an attacker working for an enterprise will one day be “as fundamental as having defenses in place.”
Plans and context
Most of the money will expand the parts of an enterprise environment the agents can reach, add workflows to verify and remediate findings, and fund hiring in London. The article notes that Sysdig documented what it described as the first ransomware operation run start to finish by an autonomous AI agent.
Source: SiliconANGLE. This is an AI Business Watch summary of the original report.