Apple has held early-stage discussions with Intel and Samsung about producing the main processors for its devices in the United States, Bloomberg reported. The talks include site visits by Apple executives to a Samsung plant under construction in Texas and a preliminary agreement with Intel that has not yet produced volume orders.
TSMC remains Apple's lead supplier and is unlikely to lose that position in the near term. Both Intel and Samsung relationships are described as preliminary.
The interesting part of this story isn't what got reported. It's what made it possible to be reported.
Apple's chip strategy has been one of the cleanest case studies in supply-chain monogamy in the industry. M-series, A-series, the security coprocessors — TSMC, exclusively, on whatever leading-edge node was available. The relationship was tight enough that Apple's quarterly volume effectively defined which node TSMC ramped first.
That stops working when node demand exceeds what TSMC can ship. The constraint isn't political; it's physical. AI data-center build-out has consumed advanced-node capacity faster than the foundry industry can add fabs. The unanticipated knock-on — Macs running local AI models against the leading-edge process — has tightened Apple's own queue. When you can't get all of what you want from one supplier, you start opening other doors. Even if you still prefer the original.
What Intel and Samsung can actually deliver is a different question. Intel Foundry's 18A is supposed to be production-ready in 2026 but has slipped before. Samsung's leading-edge yield problems on past nodes are well documented. Neither has TSMC's volume profile. Apple isn't going to ship a flagship phone on either yet.
What this announcement signals is the future shape of Apple's supply chain: dual-source on advanced nodes by 2028, even at a yield and cost penalty. The penalty is the price of not having a single point of failure in Taiwan.
- Apple's diversification follows a broader pattern: every hyperscaler has been pushing on Intel Foundry's roadmap as a TSMC alternative
- Samsung's Texas plant is part of the CHIPS Act-era US manufacturing build-out — government subsidy is partially what makes this economic
- "Preliminary agreement" framing usually means engineering-sample work, not committed volume
- See 9to5Mac's coverage for the diversification framing
If you depend on TSMC's leading edge for your hardware roadmap, watch Apple — they're the leading indicator. Diversification at the top of the stack means yields and pricing on advanced nodes are about to matter to product decisions for everyone else.