As Brussels admits its fab only chip policy was too narrow, Spain is building a complementary model on light based chips (photonics), quantum hardware, advanced chip packaging, and applied research.
Spain is not trying to build the next TSMC. It is trying to prove there is a different way to matter in chips.
Where Germany, France, and the Netherlands have spent the last decade chasing mega-fab capacity, Spain has assembled a distributed, specialized semiconductor portfolio built on photonics, quantum hardware, chip design, advanced packaging, and applied research. Those are the categories AI-era compute depends on but flagship fabs do not, on their own, supply. Two regions carry the wager. Catalonia has organized a research-to-startup coordination machine around its flagship institutes. The Canary Islands have turned astronomy-derived optics into semiconductor metrology tools. Together they form a working counter-thesis to the assumption that a country must anchor a billion-dollar fab to be strategically relevant in chips.
The state has put institutional weight behind the bet. Spain's Council of Ministers approved PERTE Chip on May 24, 2022 as the national semiconductor arm of the Recovery Plan, and the PERTE de microelectrónica y semiconductores program has since channeled public funding into design, packaging, and pilot-line infrastructure. The institutional vehicle that gives the strategy teeth, the Sociedad Española para la Transformación Tecnológica (SETT), was created by Real Decreto 676/2024 and saw its estatutos published in the BOE on June 12, 2026, turning SETT from a decree into a working entity with a mandate over technology-transfer programs including chips (per Wikipedia summary). On top of that, the government adjudicated 19 million euros in March 2026 for semiconductor and quantum chip research.
Catalonia is the half of the model that looks like a cluster. The EE Times' six-part Spain overview frames the Barcelona ecosystem, ICFO for photonics, IMB-CNM for micro- and nano-electronics, the Barcelona Supercomputing Center, and Eurecat, as a candidate distributed research-and-technology organization: a network of labs that could coordinate with industry the way a single RTO would. The industrial hook is PIXEurope, the photonic integrated circuit pilot line coordinated by ICFO under Europe's effort to industrialize integrated photonics for AI-era compute. Out of that ecosystem have come commercial bets: Quside, an ICFO spin-off shipping quantum random number generation hardware for security to global customers, and Qilimanjaro Quantum Tech, which is positioning analog quantum computing as infrastructure for future AI and HPC workloads. Both are early revenue, not research demos.
The Canary Islands model is structurally different. There is no cluster office. Instead, Wooptix has migrated wavefront-sensing technology originally built for telescope optics into semiconductor metrology for advanced packaging and AI chips. The bet is that a tool invented to peer through atmospheric turbulence can read the lithography and packaging tolerances of the next generation of chips. It is a single-company technology-transfer story rather than a coordinated regional plan, which is part of why the Spanish wager reads as more than a geography story. Catalonia shows what organized coordination can produce. The Canary Islands show what a strong research spillover can produce on its own.
The European Commission has now aligned with that diagnosis. The EU Chips Act Industry Advisory Group report published March 26, 2026 argues that the original Chips Act, Regulation (EU) 2023/1781, focused too narrowly on manufacturing and left demand, design, and startup scale underfunded. The Commission's implementation dialogue on the Chips Act is the policy venue where the rethink is being turned into Chips Act 2.0. Spanish industry voices including Maria Marced and Francesc Guim are inside that conversation, pushing Brussels toward a model that rewards specialization, not just gigafabs.
Two caveats matter. Specialization is not a substitute for fab capacity; it is a complement, and the EU still needs the flagship projects. Diamond Foundry's $850 million D Foundry II in Trujillo, Extremadura is a synthetic-diamond plant, not a silicon fab, and should not be counted toward logic or memory chip capacity.
Spain has now run the test the rest of Europe is still writing. The next read is whether SETT's first year of operating statutes produces measurable design wins, pilot-line output, and startup scale, or whether the institutional structure exists on paper while the commercial layers stay thin.