Fleuret AI Secures €4M Pre-Seed to Replace Static Penetration Tests with Autonomous Hackers

The traditional penetration testing model—where companies pay thousands of dollars for a one-off security audit—is increasingly obsolete in an era of continuous software deployment. To address this gap, French cybersecurity startup Fleuret AI has secured a €4 million pre-seed round led by RAISE Ventures. The capital will be used to scale its platform, which deploys autonomous agents to constantly probe enterprise networks for vulnerabilities just as a human hacker would.

Rather than relying on a single large language model to execute tests sequentially, Fleuret utilizes a multi-agent architecture. Its primary agents, dubbed Emile and Champollion, divide the labor: one maps the infrastructure and APIs, while another attempts to exploit weaknesses and generate proof of compromise. This approach contrasts with US-based rivals like XBOW, which recently raised $155 million, by offering a localized, sovereign alternative hosted entirely on European cloud provider Scaleway.

By automating the labor-intensive aspects of offensive security, Fleuret can reduce the turnaround time for a penetration report from a month to under 72 hours, cutting costs by more than half. The investment highlights a broader shift in the cybersecurity industry toward "continuous red-teaming," where AI is used not just to defend perimeters, but to actively and relentlessly attack them before malicious actors can.