Intel Returns to Europe: Ireland's Fab 34 Introduces 3nm Node, Adding Variables to European Semiconductor Manufacturing

Background: From Germany to Ireland, a Shift in Intel's European Strategy

In March 2025, Intel announced an investment of approximately €5 billion (about $5.7 billion) in its Fab 34 wafer fabrication plant in Leixlip, Ireland, to upgrade to the Intel 3 process and expand capacity. This investment marks a major turning point in Intel's semiconductor manufacturing layout in Europe—previously, under Pat Gelsinger's leadership, the company planned to build a new cutting-edge fab in Magdeburg, Germany, but canceled the project in 2024 due to financial difficulties. The expansion in Ireland not only retains Intel's manufacturing footprint in the EU but also hints at a shift in strategy from "building new large factories" to "expanding existing capacity."

Fab 34 previously adopted the Intel 4 process (4nm). This upgrade to Intel 3 (classified as a 3nm node) will be directly used to manufacture the Xeon 6 series (Granite Rapids and Sierra Forest) and subsequent unnamed Xeon products. Notably, Intel had sold a 49% stake in Fab 34 to private equity firm Apollo Global Management in 2024, but bought it back for $14.2 billion in early 2025, regaining full control—this cleared the ownership obstacle for this investment.

Technology Impact: Technical Positioning and Roadmap Significance of Intel 3

Intel 3 is an enhanced 3nm process introduced by Intel after Intel 4, primarily targeting high-performance computing and server chips. Compared to Intel 4, Intel 3 offers improvements in performance, power efficiency, and transistor density. It is currently Intel's second most advanced process after Intel 18A (1.8nm). Although there is a marketing component in node naming, from a technical perspective, Intel 3 competes in the same generation as TSMC's N3 series.

The introduction of Intel 3 in Ireland means that Intel is deploying its most advanced volume production capable process (ahead of Intel 20A) within the EU. This is crucial for Intel's product line: Xeon server chips are its core profit business, and chips manufactured using Intel 3 will directly compete with AMD's EPYC (based on TSMC's 3nm/5nm). Furthermore, whether Fab 34 will be upgraded to Intel 18A in the future remains unclear, but if realized, it would provide 1.8nm-level manufacturing capability in Europe—a global top-tier process.

In terms of technical barriers, Intel 3's maturity is already higher than its initial stage, but yield improvement still takes time. Intel has cumulative investments of over €3 billion and approximately 4,900 employees in Ireland, with the accumulation of technical talent being a relative advantage. However, compared to TSMC's 3nm fabs in Arizona and Japan, Intel's single fab is smaller in scale, limiting economies of scale.## Supply Chain Impact: Beneficiaries and Risks in the Industry Chain

Upstream Equipment and Materials - Equipment Manufacturers: The €5 billion investment in Fab expansion will inevitably drive procurement of key equipment such as lithography, etching, and deposition tools. ASML (extreme ultraviolet lithography), Applied Materials, Lam Research, etc., will receive orders. Since the Ireland project is within the EU, equipment transportation and installation may benefit from European local procurement policies, which is particularly favorable for ASML (Netherlands). - Material Suppliers: Demand for silicon wafers (Shin-Etsu, SUMCO, GlobalWafers), photoresist (JSR, TOK), specialty gases (Linde, Air Liquide), etc., will increase. European local material suppliers (such as Merck, INEOS) may gain a larger share.

Midstream Manufacturing and Packaging/Testing - Intel Itself: The upgrade of Fab 34 directly enhances its own chip production capacity, reducing reliance on TSMC (Intel has currently outsourced some Xeon and GPU products to TSMC). Meanwhile, if Intel Foundry Services (IFS) can provide Intel 3 capacity manufactured in the EU to external customers, it will attract potential orders from European chip design companies (such as SiPearl, Imagination) and automotive chip manufacturers (such as NXP, Infineon). - Packaging and Testing: If Intel's packaging and testing capabilities in Ireland are expanded simultaneously, it will reduce cross-continental chip transportation and shorten the supply chain cycle. However, the current announcement does not specify packaging and testing investments.

Downstream Applications - Cloud and Server Market: Xeon chips manufactured in Europe can meet the EU's demands for "digital sovereignty" and "data security", for example, serving customers in critical infrastructure such as government, finance, and telecommunications. This may drive an increase in Xeon's market share in the EU. - AI Infrastructure: Although Intel 3 is not the ideal process for the most advanced AI training chips, Xeon servers, as AI inference and infrastructure support chips, still benefit from the wave of AI investment in Europe.

Risk Factors - Reliance on a Single Factory: Fab 34 is Intel's only advanced process factory in the EU. If natural disasters or geopolitical events occur, supply will be severely impacted. - Cost Pressure: The €5 billion investment is much smaller compared to the German factory (originally planned at around €33 billion), but Intel is still in a financial restructuring phase, and tight capital expenditure may affect subsequent upgrades. - Export Controls: As an advanced process, Intel 3 may be subject to export restrictions to China, potentially affecting the scope of services this factory can provide to third-party customers.

Competitive Landscape: Subtle Changes in the Foundry Market

• Intel's return to investment in Europe has the most direct impact on the competitive dynamics of TSMC and Samsung in the EU market.- TSMC: TSMC is building its first European fab in Dresden, Germany (expected to mass-produce in 2027 using 12-28nm mature processes) and is advancing leading-edge nodes in Japan and Arizona. Intel’s Intel 3 fab will be the first to mass-produce advanced nodes in the EU (announced in 2025, expected to mass-produce in 2026), potentially coming online ahead of TSMC’s German facility. However, TSMC’s German fab focuses on automotive/industrial chips, with a different positioning. If Intel IFC can attract European AI chip clients (such as SiPearl, European CPU startups) to adopt Intel 3, it could weaken TSMC’s monopoly in advanced processes.
• Samsung: Samsung currently has no advanced fab in Europe, and its 3nm yield issues in Texas limit its competitiveness. Intel’s Ireland fab poses a smaller threat to Samsung.
• European IDMs: Existing European fabs like STMicroelectronics and GlobalFoundries primarily offer mature processes. Intel’s entry may not directly compete, but could compete for talent and policy subsidy resources.

In terms of market share adjustments, if Intel achieves high-yield mass production of Intel 3 in the EU, it will enhance the appeal of its foundry services (IFS), especially to clients valuing the "Made in Europe" label. In the short term, however, TSMC will still dominate the global 3nm market (expected to hold over 90% market share in 2025).However, Intel's Irish fab only produces the Intel 3 node generation, and is dedicated to the Xeon series, rather than being primarily for open foundry. If the EU hopes for true autonomous manufacturing, it still needs diversified foundries and a local chip design ecosystem.

Investment Perspective: Capital Market Re-evaluates Intel's Transformation

After Intel announced the investment, its stock price fluctuated slightly, reflecting the market's cautious attitude towards its financial sustainability. The previous behavior of selling and then repurchasing equity in the Irish fab exposed its cash pressure. Although the €5 billion investment is smaller than that for the German fab, it will still consume cash flow. However, Intel has made it clear that the Irish fab will be used for profitable Xeon products rather than indiscriminate capacity expansion, which investors may see as a positive signal.

The long-term value lies in: if Intel IFS can secure external orders (especially EU security-related chips) through this fab, it may open up new revenue sources. Additionally, the scarcity of semiconductor manufacturing in Europe may allow this fab to receive government subsidies, reducing net costs.

Long-Term Outlook: Trends for the Next 3-10 Years

• Within 3 years: Fab 34 completes Intel 3 upgrade and mass production, Xeon 6 series is launched, and yield gradually improves. Intel may continue investing in packaging and testing, but will not upgrade to 18A.
• Within 5 years: The effects of the EU Chips Act become apparent, and TSMC's German fab enters mass production, forming a dual-track pattern of "TSMC's mature processes + Intel's advanced processes." If Intel IFS cannot secure significant external orders, the European fab may be limited to internal use.
• Within 10 years: If Intel 18A succeeds, the Irish fab may receive investment in back-end processes (such as advanced packaging), but the probability of reintroducing 18A core wafer manufacturing is low due to high capital intensity. Europe's semiconductor autonomy still depends on cooperation among multiple manufacturers.

Conclusion

Intel's 3nm expansion of Fab 34 in Ireland is a substantial improvement in Europe's semiconductor manufacturing capability, but it also exposes its limitations: limited scale, a single node, and concentrated customers. The biggest industry judgment is that global foundry competition has shifted from "pursuing the most advanced nodes" to "regional capacity deployment," and Intel's strategy of using existing bases for rapid response is more pragmatic than building new fabs. For the EU, this is not so much a milestone of "autonomous manufacturing" as it is an exposure to the risk of "strategically relying on a single American company." In the future, Europe should simultaneously cultivate a local design, equipment, and materials ecosystem, rather than relying solely on foreign investment.