Alias Robotics and the EIC Accelerator: From Robots to Hacking AI

Alias Robotics, incorporated in Vitoria-Gasteiz in October 2017 and led by founder Víctor Mayoral Vilches, set out to be the world's reference firm in robot cybersecurity with a "Robot Immune System" to protect industrial robots from attack. The EIC Accelerator backed it to build a next-generation, AI-powered version of that platform. CORDIS now lists the project as "Revolutionising cybersecurity with the next generation AI-powered security platform", without the word robot, with an EU contribution of €2,499,875 towards a total cost of €3,571,250, running from July 2024 to June 2027, according to CORDIS.

The project's first report says Alias delivered six deliverables and two milestones, a first prototype of edge hardware, its own language model, and CAI, an open-source AI agent for security testing, engaging stakeholders including Denso, ABB, Mercedes and Telefónica, according to CORDIS. CAI, released on GitHub in March 2025, had 9,838 stars and 1,485 forks by September 2026, according to GitHub. The company's website now calls it the "EU Leading Cybersecurity AI Lab", selling on-premise cybersecurity language models to defence and intelligence users, with robotics services further down the menu, according to Alias Robotics.

Where Alias Robotics Stands Now

In June 2026 Alias published the CAI Dataset, 18.07 terabytes of what it calls real-world offensive and defensive hacker trajectories collected through CAI, to train private, on-premise models, according to Alias Robotics. In March 2026 it published research using CAI to find vulnerabilities in consumer robots, and the EIC selected it for CES Las Vegas in January 2026. Its latest filed accounts, for 2024, show annual turnover below €500,000, according to Empresia.

The Technology: An AI That Hacks

CAI is a framework of AI agents that automate penetration testing and other security tasks. The company says its agent topped the 2025 capture-the-flag competition circuit and that its alias2-mini model runs air-gapped for military users. Robots remain a use case: the first EIC report describes CAI as able to automate robot cybersecurity and to deceive attackers with simulated robots, and the platform was aligned with the industrial security standard IEC 62443.

What the EIC Award Promised, and What the Record Shows

The award was for a robot-specific, AI-powered security platform. The project runs until June 2027, so it is too early to judge it. The record shows strong technical output and a product that found a larger market in general cybersecurity. Whether the robot platform itself reaches customers is less visible.

1. AI security engineDeliveredCAI released.
2. Edge hardwarePrototypeFirst report.
3. Robot OEM deploymentsUnverifiedStakeholders only.
4. Revenue growthUnverifiedUnder €0.5M in 2024.

The Money

In 2024 the Basque robotics network Robotekin reported that besides the €2.5 million grant, Alias planned a €5 million investment round in early 2025 with the European Investment Bank committing half. No closing of that round has been announced.

What Cannot Be Checked

Why Europe Should Care

Alias argues that security teams leak sensitive data when they use commercial AI assistants, and offers European models that run on-premise instead. That matters for governments and militaries. Alias built it with EIC money meant for robots, which is not what was funded.

The Verdict

Alias Robotics and the EIC have not been asked for comment for this article. It is based entirely on public records and published statements, linked throughout.