London-based CodeWall emerged from stealth on October 14, 2026, with $4 million in pre-seed funding for autonomous agents that test enterprise systems from an attacker's perspective. Osney Capital led the round, joined by InReach Ventures, Antler, and angel investors. The company aims to make adversarial security testing an ongoing operational capability rather than an occasional exercise.
Funding and Market Positioning
CodeWall is developing software to identify exploitable weaknesses across interconnected systems. Its approach reflects a growing challenge for corporate security teams as artificial intelligence makes it easier to automate sophisticated attack sequences. The company wants enterprises to repeatedly test their exposure using simulated attacks alongside defensive monitoring.
Founder and CEO Paul Price argues that companies should uncover potential compromises before malicious actors do. CodeWall proposes a persistent offensive capability that tests changing infrastructure without requiring specialists for every assessment. The company did not disclose its valuation.
How the Platform Works
Its agents examine externally accessible systems without internal credentials or source-code access. Specialized agents investigate in parallel, sharing findings through a persistent graph of systems and weaknesses. This allows the software to examine whether an initially limited vulnerability could enable access to other applications, accounts, or data.
Rather than simply presenting a list of possible vulnerabilities, CodeWall seeks to establish which attack paths are actually exploitable. The system traces how weaknesses can be combined and produces supporting evidence that security teams can use to prioritize fixes. Testing is conducted within authorized boundaries, with the platform designed to revisit targets after infrastructure changes or remediation work.
Early Deployments and Security Research
CodeWall says it already works with banks and financial institutions to validate exploitable attack paths. The company has also published research describing vulnerabilities it identified in platforms associated with McKinsey & Company, Boston Consulting Group, and Bain & Company. These examples illustrate the kinds of interconnected application risks the startup is seeking to address.
In one investigation, CodeWall says its agents obtained read and write access to the production database supporting McKinsey's Lilli AI platform within two hours. The finding highlights the potential consequences of weaknesses that extend beyond a single application into sensitive backend infrastructure. Such reported demonstrations provide a view of the technology's capabilities, although the funding announcement does not disclose independent performance benchmarks.
Continuous Testing and Enterprise Security
Penetration testing is often commissioned periodically, leaving gaps as applications and configurations change. CodeWall is positioning its product as a complement to existing defenses, with recurring investigations intended to reveal new weaknesses and verify whether previous fixes are effective. The objective is to give security teams timely, actionable evidence rather than treating testing as a periodic compliance exercise.
Paul Wilkes, a partner at lead investor Osney Capital, said the investment reflects the possibility of shifting some of the offensive advantage toward defenders. His view is that organizations need the capacity to probe their own environments as persistently as potential attackers might. The concept also raises an important operational requirement: autonomous testing must remain controlled and aligned with each customer's authorization and security policies.
Expansion Plans
CodeWall plans to use its new capital to extend the range of enterprise environments its agents can assess and strengthen testing, verification, and remediation workflows. It also intends to expand its London-based teams across offensive security, engineering, customer success, and commercial functions. It did not specify hiring targets or development timelines.
CodeWall's emergence from stealth adds another entrant to the market for AI-driven cybersecurity tools, with an emphasis on verifying real-world attack paths. Its $4 million pre-seed round will support product development and team expansion as it seeks to move offensive testing into day-to-day enterprise security operations. The company's progress will depend on how effectively its agents identify actionable risks while operating within enterprise safeguards.