London-based Maze, which uses AI agents to stop cloud security breaches, launches with $31M in total funding, including a $25M Series A led by Theory Ventures
We're launching Maze and announcing our $25M Series A led by Theory Ventures, with participation from Cherry Ventures and Tapestry VC. … Santiago Castiñeira : Today, we're launching Maze and announcing our $25M Series A led by Theory Ventures, with participation from our seed investors Cherry Ventures and Tapestry VC. …
Context & Ripple Effects
Maze’s launch brings a newly funded cloud-security entrant into an agentic-security field where [[a:883327|Crogl raised capital for an AI agent that helps researchers triage network alerts and fix incidents]]. Maze is positioned earlier in the incident chain: stopping cloud breaches rather than analyzing the alert flow afterward.
The $25M Series A, part of $31M total funding, gives the London-based company backing from Theory Ventures alongside existing investors Cherry Ventures and Tapestry VC as it enters that adjacent market.
First-order effects
- Maze gains $25M in new Series A capital and launches with $31M total funding, giving it resources to build and commercialize its AI-agent approach to cloud-breach prevention.
- Theory Ventures becomes the round’s lead investor, while Cherry Ventures and Tapestry VC deepen their involvement in Maze.
Second-order effects
- Security buyers evaluating AI-assisted operations will be able to compare a prevention-focused cloud agent with tools such as Crogl’s alert-analysis agent, increasing pressure on vendors to define where their automation sits in the security workflow.
- Competing agentic-security startups will need to distinguish whether they automate design, prevention, detection, investigation, or remediation as funding and buyer attention spread across those layers.
Third-order effects
- If adoption follows, cloud security is likely to shift from point tools that surface risk toward agent-driven systems expected to take more operational action; the key market test will be whether customers trust those actions in production.
- The category may segment by workflow rather than by the broad “AI security” label, favoring vendors that can show a clear handoff between prevention and incident response.
The trend: Maze is part of the move to apply AI agents to discrete, high-stakes security workflows rather than treating cybersecurity as a single undifferentiated automation market.