Published on September 2, 2026 · The mitza.es team
For the first time, the Spanish State has taken a direct stake in a Spanish artificial intelligence company. It did so through SETT (Sociedad Española para la Transformación Tecnológica), the public tech-investment vehicle linked to Spain's Ministry for Digital Transformation and the Civil Service, informally nicknamed the "digital SEPI." The company in question is Sherpa.ai, and the news broke in early September 2026.
Who is Sherpa.ai
Sherpa.ai is a Spanish company specialized in artificial intelligence and data privacy, founded in 2012 in Erandio (Vizcaya, Basque Country) by Xabi Uribe-Etxebarria. It's one of the most experienced Spanish companies in AI, having operated in the space for over a decade. The company recently closed an $18 million funding round, announced around July 2026, with capital that includes Silicon Valley investors. SETT participated in that round through the US fund Forgepoint Capital, marking the first time the Spanish State has entered the shareholding of an AI company.
What "data sovereignty" in AI actually means
Sherpa.ai's core focus is what's known as data sovereignty: the ability to train and use large language models without the underlying data ever leaving the client's own servers or infrastructure. In practice, this means companies handling especially sensitive information — healthcare, finance, public administration — can use AI capabilities without sending their data to third-party cloud providers, typically US-based ones.
Translated into terms any SME can relate to: when you use a chatbot or AI tool to handle customer support, write content, or analyze internal data, that information travels somewhere. With most mainstream AI tools, that "somewhere" is a server in the United States, governed by US law. Sovereign AI infrastructure offers an alternative: keeping that data under European or Spanish control at all times.
What this is not: there's no "Spanish ChatGPT"
It's worth being honest about the real scope of this news. It does not mean Spain now has its own national foundation model comparable to GPT, Gemini or Claude. What it represents is public backing for AI infrastructure that is controlled from Europe and Spain, aimed at organizations that handle especially sensitive information. It's a more modest step than some headlines suggest, but a real one: for the first time, the government is putting public money behind a Spanish company building these capabilities.
What a Spanish SME should actually take away from this
Beyond the headline, there's a practical lesson for any business evaluating AI tools: it's worth asking where your data is processed and stored, especially if you handle customer personal data under GDPR and Spain's LOPDGDD. This isn't an abstract geopolitics debate — it's a compliance question that can have real consequences if your company manages sensitive information. It's the same kind of thinking we already covered when discussing NIS2 cybersecurity for SMEs: knowing where your information lives, and under which jurisdiction, is basic digital risk management for any business.
This move also fits the broader trend we already covered on AI data center investment in Spain: more and more AI infrastructure is being built and controlled on Spanish and European soil, which over time should give Spanish businesses more options beyond the major US providers.
How we see it at Mitza
You don't need to be a bank or a hospital for this to matter to you: any SME using free or low-cost AI tools should be clear on each tool's data policy before uploading customer information. If you want to review how your website and digital tools handle the data they collect, we can help with an audit that looks at both regulatory compliance and technical performance.