Apple Seals Preliminary Foundry Pact With Intel, Nudging—Not Toppling—TSMC

Apple’s move echoes its 2006 switch from IBM’s PowerPC to Intel x86 chips—then likewise driven by supply, thermals, and strategic leverage. Today, supply-chain resiliency and Washington’s re-industrialisation push (reminiscent of the 1987-funded SEMATECH consortium) intersect: the U.S. now owns 10 % of Intel and is prodding domestic production just as talk of a cross-strait crisis looms over Taiwan. Over the past decade, TSMC’s near-monopoly on sub-5 nm nodes created a single point of failure; the scramble for AI silicon and capacity shortages (Nvidia, Microsoft et al. plan $725 billion cap-ex) are forcing even Apple to cultivate a “plan B.” Still, history cautions that catching the process-yield leader is hard: GlobalFoundries exited the bleeding edge in 2018, and Intel itself lost it after 2014. If Intel proves it can ship 18A at scale, this could mark the first genuine multi-foundry era since the 1990s; if not, the episode may be remembered as a politically-driven footnote. On a 100-year arc, the event illustrates how geopolitical power, not just Moore’s Law, steers where and by whom the world’s critical compute engines are made.