The Barcelona Supercomputing Center spin-off, founded eighteen months ago, uses AI to generate realistic test scenarios that expose failures, hallucinations, bias, and security risks before enterprise AI agents go live. 42CAP led the round; Mozilla Ventures participated.

The gap between an AI agent that works in a demo and one that works in production is, increasingly, a testing problem. Building a generative AI product is no longer the hard part, the models exist, the APIs are accessible, and the developer tooling has matured.

What has not kept pace is the infrastructure for knowing whether the thing you built will behave reliably once real users start interacting with it at scale.

Galtea, a Barcelona-based startup spun out of the Barcelona Supercomputing Center in October 2024, is building that infrastructure. It has raised $3.2 million in seed funding to continue the work.

The round was led by 42CAP, the Munich-based early-stage technology fund, with participation from Mozilla Ventures, whose investment thesis centres on trustworthy AI, and existing investors JME Ventures, Masia, and ABAC Nest Ventures. Total funding is now $4.1 million. The new capital will go towards expanding Galtea’s engineering and commercial teams and developing the platform further.

Galtea’s core product generates test cases and synthetic user simulations from descriptions of how an AI agent is intended to behave. The idea is to create adversarial and edge-case scenarios automatically, at scale, without requiring engineering teams to write them by hand, a process that is time-consuming, expensive, and, in practice, rarely complete.

The platform evaluates models across hallucination rates, bias, security vulnerabilities, and toxicity, and outputs structured metrics that developers and compliance teams can use to make deployment decisions. This week, the company also launched a self-service tier with a free trial, broadening access beyond its existing enterprise customer base.

The founders are unusual for this space. Jorge Palomar, CEO, worked at Amazon and within BSC’s Language Technologies research group before founding Galtea; his co-founder Baybars Külebi, CTO, holds a PhD in Astrophysics, has co-founded several earlier language and audio technology projects, and spent years as a machine learning expert at BSC.

The technology itself was originally developed at BSC to evaluate large language models for internal research purposes, running on MareNostrum 5, one of Europe’s most powerful supercomputers. That provenance gives Galtea an unusual degree of scientific depth for a company that is eighteen months old.

The commercial timing is structured. The EU AI Act is now requiring companies deploying AI in high-risk applications to document and validate their models’ safety and compliance, with fines of up to €35 million for violations.

For European enterprises that have been building AI products without systematic testing infrastructure, the regulation has created urgency. Galtea’s platform sits directly in that gap: helping development and legal teams produce the evidence of compliance that the regulation demands, without rebuilding workflows from scratch.

The Mozilla Ventures involvement in particular signals that the round was framed partly around the trustworthy-AI narrative that has defined the fund’s portfolio thesis since its launch in 2022.