Galtea, a Barcelona-based AI evaluation platform focused on helping enterprises and developers launch reliable AI agents in production, has raised €2.7mn in Seed funding.
The round was led by German fund 42CAP, with participation from US-based Mozilla Ventures and existing investors JME Ventures, Masia, and ABAC Nest Ventures.
In 2024, Galtea raised nearly €750k in a pre-Seed round. Total funding now stands at €3.5 mn.
Galtea emerged from the pioneering work at the Barcelona Supercomputing Center, harnessing the power of the MareNostrum 5 supercomputer.
Our mission is to guide companies in safely and responsibly deploying generative AI, ensuring trust, compliance, and long-term success.
Jorge Palomar, Galtea’s co-founder and chief executive, said limited access to affordable testing data is slowing AI deployment across the market. He said many teams cut corners or fall back on weak generic benchmarks because proper testing costs too much.
Thorough testing and simulations for accuracy, performance, behaviour, and security are the only way developers can know how their platform will perform in a real-world setting.
Jorge Palomar, Galtea’s co-founder and chief executive
“Galtea is providing generative AI developer teams with infrastructure to overcome barriers to AI adoption and deploy systems reliably, scalably, and efficiently”.
He argued that testing and simulation across accuracy, performance, behaviour, and security remain the only way developers can understand how a platform will perform in real conditions.
In his view, Galtea gives generative AI teams the infrastructure needed to move past adoption barriers and deploy systems with more consistency, at scale, and with less waste.
Founded by Jorge Palomar and Baybars Kulebi, Galtea was spun out of the Barcelona Supercomputing Centre in 2024. The company gives enterprise AI teams continuous test scenarios and tailored metrics to evaluate agents and guide decision-making.
Galtea says one of the main bottlenecks in AI is the lack of accessible, trustworthy test data.
The company estimates AI testing costs businesses across Europe and the US around €11.2 bn a year. Big number, and a painful one for teams under pressure to ship fast.
The startup says it wants to remove that barrier by offering test scenarios built to be trusted, faster to run, and cheaper to access. It is also launching a self-service model, including a two-week free trial.
The company argues that AI teams moving fast without thorough checks on accuracy, performance, behaviour, and security risk falling behind. Speed helps, sure. Weak validation catches up later.
Galtea says most teams still ship AI agents with a small set of hand-written test cases and then hope nothing breaks in production.
Its platform automatically creates hundreds of use-case-specific test cases from product specifications and evaluates AI agents against structured metrics, turning manual testing into the exception rather than the process.
The company has a team of 12. Its customers include Telefonica, ABANCA, and several other startups and enterprises.
Galtea says clients have reported an average 71% drop in AI validation costs, leading to an estimated 10x return on investment.
In one case involving a Tier 1 institution, a customer support agent failed 2,164 evaluations linked to seven critical vulnerabilities.
Galtea said that failure rate was 12 times higher than internal testing had shown. Its platform automatically created more than 6,000 test scenarios for the agent, saving around 600 hours of manual work.
With the new funding, Galtea plans to expand its commercial and engineering teams and keep building its platform for high-quality, use-case-specific test scenarios for generative AI agents.
