London-based Maze secures €22.5M to beat hackers at their own game through an AI agent

Maze, an AI-native cloud security platform, has secured $25M (approximately €22.5M) in Series A funding led by Theory Ventures to advance autonomous vulnerability investigation and resolution.

Founded by Harry Wetherald, Adrian Jozwik, and Santiago Castineira in 2024, the startup has now raised a total of $31M (approximately €27.9M), coming less than a year after the company’s founding and 9 months after a $6M seed round.

The funding will support continued team growth and expansion into new use cases for its AI-native security platform. Maze plans to scale its comprehensive security solution beyond vulnerability management into additional cloud security applications.

The round attracted participation from existing investors Cherry Ventures and Tapestry VC, with the original seed round led by Cherry Ventures. Maze has deployed its AI agents in production with over ten organisations, including two of the Fortune 200 and a number of high-growth technology companies.

Addressing 40 per cent CVE growth with an AI-native approach

Maze’s AI agents are designed to operate and think like humans, deploying thousands of agents to investigate customer cloud data and accurately identify exploitable vulnerabilities.

The platform automatically resolves critical security issues, targeting the tiny number of vulnerabilities likely to cause breaches.

In 2024, vulnerability exploitation increased by 34 per cent, overtaking phishing amongst the top causes of security incidents, whilst the number of known vulnerabilities (CVEs) increased by nearly 40 per cent. Most enterprises can only fix about 10 per cent of their open vulnerabilities each month, creating a significant security gap.

From Tessian, Elastic, and Amazon to cloud security innovation

The founding team previously led product, design, and engineering teams at Tessian, Elastic, and Amazon, working alongside veteran engineers from Meta, Adyen, and Nvidia. Theory Ventures partner Andy Triedman joins the Maze board, bringing expertise in AI-native enterprise solutions.

“Before starting Maze, we talked to hundreds of security professionals and heard the same complaint again and again – dealing with vulnerabilities is a nightmare,” says Harry Wetherald, Maze CEO and co-founder. “Security teams are already struggling today, and as attackers get their hands on AI, things are only going to get more difficult. While AI gives us the ability to build better security tools, it is also enabling bad actors to deliver complex attacks at scale. Our goal is to give security teams the tools they need to fight back.”

“To keep pace with attackers, security tools will need to be re-written to be AI-native,” said Andy Triedman, Partner at Theory Ventures and latest addition to the Maze Board. “While today’s vulnerability management tools are built to aggregate data for human analysts, Maze is designed from the ground up for autonomous investigations and triage.”

€27.9M trajectory: from seed to Series A in nine months

The company’s rapid funding progression from €5.4M seed to €22.5M Series A demonstrates strong market validation for AI-native security solutions. Maze built its first AI agent by replicating the workflow of expert human analysts, differentiating from traditional vulnerability management tools that rely on pre-defined logic and rules.

“Maze’s product is almost too good to be true. It is a truly novel use of Agentic AI that goes well beyond automating manual workflows,” said Mandy Andress, CISO at Elastic.