Partnership strategy: who Europe should invite in

After July 2025 a realistic conclusion must be accepted: the U.S. will not be a source of new capital for European fabrication in 2026–2028. The U.S. remains a critical partner in regulatory cooperation (TTC), defence technologies, research, x86 architecture, AI models, and cybersecurity. That is not changing and Europe should not want it to change. But the investment capital for new manufacturing capacity in Europe must come from elsewhere — if it comes at all.

The CHIPS Act ($52 bn) systematically pulls investment into the U.S.: TSMC Arizona over $165 bn , Samsung Taylor TX, Micron Idaho. The EU has no comparable fiscal instrument. Hoping for transatlantic investment flows in the coming three years is not strategy.

This chapter sketches the partnership axis that follows from this constraint. Each candidate gets a short profile in its own chapter:

• South Korea — the most underrated partner, with structural strength in HBM (53–62 % world market share via SK hynix) and advanced packaging
• Japan — the structurally most natural ally, sharing Europe's position of strength in materials and equipment, weakness in advanced logic
• Taiwan — already inside via ESMC Dresden, but at non-leading-edge nodes; consolidate, do not expect more
• United States — a critical partner, but not for new European fab capital
• India — a long-term option for 2028+

Six directions for EU partnership policy

Drawing the partnership map together yields six concrete directions for the EU's partnership policy in 2026–2028:

1. Move the Korea-EU Digital Partnership into the investment phase. Specific target: by end of 2027, sign a memorandum of understanding with Samsung Foundry or SK hynix on a pilot investment in the EU.
2. Structure a deep industrial partnership with Rapidus. Next step: preferential access for European AI and HPC customers to its 2 nm capacity, in exchange for European capital participation.
3. Consolidate the ESMC and TSMC relationship without expecting more. Multi-year preferential allocation contracts for European customers and contractual continuity guarantees.
4. Protect Codasip for the European ecosystem. A possible European consortium buyout (with EuroHPC, national funds, and Bosch/Infineon/NXP participation).
5. Differentiate the U.S. partnership. Regulatory cooperation yes, investment expectations no. Use U.S. firms as key customers of the European supply chain.
6. Open a second partnership circle with India, with a realistic 2028+ horizon.

These directions are not strategic dogma — they are working hypotheses, each testable against forthcoming announcements. What they have in common: none assumes that Europe can rebuild its industrial position alone, and none assumes that any single partner can substitute for the U.S. as the dominant source of new capital. Partnership strategy in 2026 is a portfolio strategy.