Openchip is four years old, employs about 300 people in Barcelona, and has not yet sold a chip at volume. On Monday the Spanish government paid €115 million to own a piece of it, and the Catalan government, which had already lent the company €35 million across the spring, converted those loans into shares to take a stake of its own.

The cheque came from SETT, the Sociedad Española para la Transformación Tecnológica, a state company that exists to put public money into semiconductors, microelectronics, and the other technologies Madrid has decided it cannot afford to import wholesale. The investment runs through SETT’s Next Tech facility, financed by the EU Recovery, Transformation and Resilience Plan, and was authorised by the Council of Ministers the same day it was announced. Add the Generalitat’s converted loans to Madrid’s injection and the two governments end up holding roughly 16.5% of a company valued at about €700 million. There is no private venture fund in this round. The state is the round.

The cheque is European, and so is the anxiety behind it

The thing Openchip actually makes is a family of accelerators and systems-on-chip built on RISC-V, the open instruction-set architecture that anyone can license without paying or asking a foreign owner. That last clause is the whole pitch. The dominant architectures in the data centre belong to American and British companies, and the manufacturing that turns a design into silicon sits mostly in Taiwan. A European chip designed on an open standard, optimised for performance per watt, and aimed at the supercomputers and AI workloads the continent is now building, is a way of owning a layer of the stack that Europe currently rents.

Brussels has already blessed the framing. The European Commission named Openchip an Important Project of Common European Interest, the label it reserves for industrial bets it considers strategic enough to justify state aid that would otherwise raise eyebrows. The company’s own language leans into it without embarrassment, talking about “European digital sovereignty” the way a SaaS startup talks about retention. The cheque-writers talk the same way. SETT was created precisely to take what it calls a proactive co-investor role, a deliberate break from the older idea that the state should regulate and subsidise but never own.

This is the same instinct now driving a wave of European chip money, and it is worth being honest that the instinct is reactive. Europe spent the first AI-compute cycle as a customer, watching the value accrue to a handful of American firms and one Taiwanese foundry, and the policy class has decided not to do it twice. The €115 million is less a vote on Openchip’s product than a vote against repeating the last decade.

A chip company valued on its blueprints

Strip away the sovereignty language and the financial picture is plain: a largely pre-revenue company, four years old, has been handed €115 million and a €700 million valuation on the strength of what it intends to ship. Openchip runs fabless, so it carries the design risk without the capital sink of a fab, but it also has to prove that an open-architecture accelerator can compete on performance and on the unglamorous things that decide data-centre purchases, software support, supply, and the confidence that the next generation will arrive on schedule.

The same logic underwrites the rest of Europe’s chip spending right now, and Openchip is not the only company being valued on a roadmap. The bet that Europe can build its own silicon base, rather than design around someone else’s, is also what put €115 million into Quobly’s silicon-qubit gamble in Grenoble this month, a near-identical sum chasing a near-identical thesis one layer further out in physics. In both cases the investors are buying a plan to ride the existing semiconductor industry rather than replace it, and in both cases the cheque arrived before the product did.

State capital has a mixed record at this. Public money is patient, which is exactly what a chip company needs and exactly what makes it easy to keep funding a design that never quite clears the market. The discipline that a commercial venture round imposes, the threat that the next cheque does not come, is the discipline a wholly state-owned cap table is structurally bad at applying. Openchip’s defenders would say that is the point, that some capabilities are worth owning even at a loss, and they would not be obviously wrong. They would just be making an industrial-policy argument wearing a venture-capital costume.

What the money cannot buy

None of this is a reason to be sour about it. Europe genuinely lacks a domestic high-performance compute architecture it controls end to end, the gap is real, and €115 million of patient public capital aimed at closing it is a more serious answer than another strategy paper. Openchip has the European Commission’s strategic blessing, a credible team, partnerships with the likes of SUSE to get enterprise software running on its chips, and a clear, defensible reason to exist. The company that emerges from this could matter.

The question the cheque cannot answer is whether public ownership produces a chip the market wants, or a chip the state keeps because it paid for it. Openchip now has the money, the mandate, and Brussels watching. The next few years will show whether sovereignty is a product strategy or only a procurement one.