€107 million for Basque quantum tech: Sovereignty or business venture?

€107 million for Basque quantum tech: Sovereignty or venture capital

One hundred and seven million euros of public funds are earmarked for a Basque company specializing in quantum computing and artificial intelligence. The figure is on the table, the recipient has a name—Multiverse Computing, based in San Sebastián—and the official objective is to reinforce the country's technological sovereignty. The amount itself is not the problem. The issue lies in how the money is delivered and who benefits from the appreciation if the invention succeeds.

The operation is channeled through SETT, the public vehicle that appears as a partial funder of the company. They cite thousands of devices, clients, and savings in inference regarding the technology. Auditable metrics? None. Some summarize it in two words: “it smells strange.”

What is known about the €107 million and SETT

The disbursement exceeds one hundred million, and the calculations suggest 107, with SETT acting as the channel and holding institutional representation of the Government on its board. The company sells a hybrid of quantum computing and language models. They are not the same discipline, they do not share a roadmap, nor are they financed with the same logic. That cocktail is the first technical alarm bell.

The CEO's background and indirect beneficiaries

The chief executive is a mathematician from the University of Barcelona, an IT engineer from UNED, and a doctor from the same university, with a master's from the University of Navarra and an MBA from IESE. Previously, he served as deputy CEO of Unnim Bank, with over twenty years in banking and finance. The uncomfortable reading is that the financial background outweighs the laboratory experience.

The most frequent suspicion does not point to the shareholders, but to the periphery: suppliers and banks that refinance debt which would now be paid with public funds, as in Plus Ultra, where the business lay in the margins.

Grant without milestones or equity investment with penalties

Putting public money into deep technology is not absurd by definition. What is debatable is the format: a grant without verifiable milestones, without external technical oversight, without mandatory presence, and with no return for the taxpayer. If the State risks 107 million, the minimum requirement is to enter as a serious investor: clear participation, anchored intellectual property, penalties if the company relocates, and auditing of linked suppliers.

And the usual argument: if the business were that good, private capital would line up. One hundred million public and nine hundred million private sounds credible; this current situation, not yet.

95% of companies investing in AI achieve nothing

The context doesn't help either. The market allocates over 500 billion dollars annually to artificial intelligence infrastructure and data centers, yet 95% of investing companies fail to achieve tangible or scalable results. In Europe, the dominant model is requesting grant after grant, promising results in the next call.

It's also unwise to laugh at the country. Here machines were made with their own label—Marconi, Azcoyten, Standard Eléctrica, Payma, Invicta, Danobat, Goratu, Maher, Zayer—and the greatest success of national software remains PC Fútbol and Panda Bicho. The black turtleneck uniform imitating Steve Jobs stopped being a good idea after the Theranos case.

---

With this design, the reasonable prediction is that the money will be spent, the company will announce calendar milestones, and if the product works, the headquarters will end up where incentives are better. This may not happen. The 3-5% margin managed as standard in public works and grants does not invite betting on a miracle.