Intel vs TSMC: High-NA EUV Adoption Strategies

Intel and TSMC want the same thing: smaller, faster, more efficient chips. But they are taking different paths in advanced lithography. Intel is investing early in High-NA EUV lithography. TSMC is extending the life of its conventional EUV scanners. The split reflects different long-term plans, manufacturing priorities, and views on cost and risk.

How High-NA EUV Differs From Standard EUV Technology

Conventional EUV uses 13.5 nm light and 0.33 numerical-aperture optics. It is essential for leading-edge logic because it prints finer features than deep-ultraviolet tools. High-NA EUV keeps the 13.5 nm wavelength but raises the numerical aperture to 0.55. That improves resolution by about 1.7×. It can also print some critical layers in one exposure instead of multi-patterning.

Intel’s Early High-NA EUV Rollout Plan

Intel was first to receive ASML’s commercial High-NA system, the TWINSCAN EXE:5000. It is installed at Intel’s R&D site in Hillsboro, Oregon. Intel is using it to refine processes, materials, masks, and design rules. The company plans to introduce High-NA EUV lithography in its Intel 14A process, after Intel 18A. Intel will still use conventional EUV where it is more cost-effective.

The early bet supports Intel’s push to regain process leadership and expand its foundry business. Starting now helps Intel build expertise before High-NA becomes mainstream. It can also simplify production for key layers. Fewer patterning steps mean fewer masks, shorter cycle times, and fewer overlay errors. That creates both a technical advantage and a stronger pitch for Intel Foundry.

TSMC’s Cautious Approach To Next-Gen Lithography

TSMC is taking a more conservative approach. It believes it can ship A16 and A14-class nodes without High-NA in high-volume manufacturing. Instead, it is improving its 0.33-NA EUV platform. The focus is on better masks, photoresists, overlay control, computational lithography, and process tuning. It is also leaning on other innovations, such as nanosheet transistors and backside power delivery. These can boost performance and density without relying only on higher lithography resolution.

Why Economics Are Core To TSMC’s Decision

For TSMC, the economics are decisive. High-NA tools cost much more than conventional EUV scanners. They also require new supporting infrastructure. Anamorphic optics cut the exposure field to about half of standard EUV. That can complicate large dies and may require stitching. High-NA also faces open issues in depth of focus, resist performance, masks, inspection, metrology, and yield.

TSMC runs conventional EUV at huge scale and knows how to maximize uptime and throughput. Staying on a proven platform lowers execution risk. It also improves returns on existing equipment. For a high-volume foundry, stable yields often matter more than adopting the newest tool first.

High-NA EUV Is Still In Both Firms’ Future Roadmaps

TSMC has not ruled out High-NA EUV. It has already bought High-NA tools for R&D. It is developing processes for future production. TSMC is also ASML’s largest customer, and CEO C.C. Wei has said the two companies are working closely on High-NA.

TSMC says timing will depend on measurable benefits, maturity, and cost competitiveness. Intel also plans selective deployment. It will use High-NA only on layers where the payoff is clear.

Final Takeaway: The Gap Is All About Timing

The main difference is timing. Intel is paying more now to learn faster and possibly lead sooner. TSMC is waiting for clearer high-volume economics. Over time, both are likely to adopt High-NA broadly. For now, their roadmaps reflect two rational strategies for next-generation scaling.

Related Articles

• ASML High-NA EUV is Not Ready for High-Volume Production
• Crossing the Yield Cliff: IDP V6 and the Future of Manufacturing Forecasting
• Beyond Moore’s Law: High NA EUV Lithography Redefines Advanced Chip Manufacturing