Intel tells Taiwan's chip suppliers to build service capacity early

Intel's demand that chip equipment and material suppliers in Taiwan, China make regional service redundancy a hard prerequisite for orders is not a procurement tweak—it is a structural response to the single-point-of-failure vulnerabilities exposed since 2020. At the 18A node, a 24-hour EUV or high-AR etch tool outage costs millions in lost wafer starts; spare-parts turnover radius and on-site engineer density now directly gate the yield-ramp curve. This forces mid-tier suppliers like Tokyo Electron and KLA to hold 2-to-3x safety stock locally, fundamentally rewriting their capex models. In the competitive arena, TSMC and Samsung are almost certainly applying identical pressure, while AMD—routed through TSMC's foundry stack—bypasses this constraint entirely, deepening Intel's structural disadvantage in equipment negotiation. Over the next 12 to 24 months, 'service as capacity' will displace 'equipment as capacity' as the industry's new pricing anchor, and SEMI may for the first time embed regional SLA clauses into equipment acceptance standards. The strategic signal is unambiguous: the post-pandemic era of just-in-time semiconductor logistics is over, and the cost of resilience is being socialized onto the supplier base.