Analysts map Apple’s decade‑long chip shift from Samsung to TSMC as a sovereignty play

Apple’s chip story used to look simple. Design more parts in-house, ship better devices, rely on TSMC to build them. But the new wrinkle is not about chip design at all — it’s about where those chips get made, and what happens if one foundry becomes too important. That is why the latest reports about Apple talking with Intel and Samsung matter. They suggest the company’s decade-long silicon push is widening from product control into supply-chain control. ### What changed this month? The immediate news is that Apple has held exploratory talks with Intel and has evaluated Samsung as possible U.S.-based manufacturing partners for some future main device processors, instead of relying entirely on TSMC for leading-edge production. The reporting still describes this as early-stage, not a signed production shift. But even that is notable, because Apple has spent years concentrating its most advanced chips at TSMC. (9to5mac.com) ### Why does TSMC concentration matter now? Because TSMC is no longer just serving smartphone and PC chips. It is also the key factory for the AI boom. Nvidia, AMD, and other accelerators are pulling huge amounts of advanced packaging and wafer capacity, which tightens the whole system. Apple may still be one of TSMC’s most important customers, but “important customer” is not the same thing as “infinite guaranteed capacity.” That makes a second source valuable even if Apple never moves most volume there. (9to5mac.com) ### Wasn’t Apple already doing sovereignty? Yes — but mostly on the design side. Apple started by bringing core chip architecture closer to itself, then made the leap in Macs. In June 2020, Apple said it would move the Mac from Intel processors to Apple silicon over about two years. By June 2023, that transition was effectively finished with the Apple-silicon Mac Pro replacing the last Intel holdout. That was sovereignty over the brain of the device. (cnbc.com) ### Where do modems fit in? Modems are the next layer. Apple’s iPhone 16e, announced in February 2025, introduced the C1 — Apple’s first self-designed cellular modem. That matters because modems were one of the biggest remaining external dependencies in the iPhone. Apple is not fully free of Qualcomm yet, though. Qualcomm said in September 2023 that it would supply Apple modem systems for smartphone launches in 2024, 2025, and 2026. So the C1 is a beginning, not the end state. (apple.com) ### So what is the real strategy? Basically, Apple has been removing dependencies one layer at a time. First, reduce dependence on Samsung in application processors by moving advanced manufacturing to TSMC. Then replace Intel in Macs with Apple silicon. Then start replacing Qualcomm in cellular. Now, if the recent talks turn into real foundry deals, Apple would be attacking a different dependency — overreliance on one manufacturing partner, even a very good one. (apple.com) ### What’s the catch? More sovereignty can mean more complexity. A single foundry is risky, but it is operationally clean. Multiple foundries mean more process tuning, more validation, more testing, and more software and hardware coordination across slightly different manufacturing outcomes. Chips are not like PDFs — you do not just send the same file to three printers and get identical results back. The gain is resilience. The cost is engineering overhead. (9to5mac.com) This tradeoff is an inference from how semiconductor platform qualification works and from the reported diversification push. ### Why Intel specifically? Intel offers something Apple does not get from TSMC alone — geographic and political diversification, plus a potential U.S. manufacturing option for advanced chips if Intel’s foundry execution improves. For Intel, landing Apple would be a credibility event. For Apple, even a limited deal could work as capacity insurance and negotiating leverage. That does not mean Intel becomes Apple’s new primary factory. It means optionality starts to have real value. (9to5mac.com) ### Bottom line? The big idea is simple. Apple’s silicon project started as a performance play, then became a control play, and now looks more like a resilience play too. If the Intel and Samsung talks go anywhere, the next phase of Apple sovereignty will not be about designing one more chip — it will be about making sure no single bottleneck can hold the whole roadmap hostage. (9to5mac.com) (cnbc.com)