Apple eyes Intel 18A-P node
- Apple is reportedly evaluating Intel’s 18A-P process for future M-series chips, reviving a rumor that Intel could finally win a top-tier external foundry customer. (trendforce.com) - The hook is the node itself: Intel says 18A-P can deliver over 9% more performance or 18% lower power than baseline 18A. (vlsisymposium.org) - If it happens, this is less about Apple dumping TSMC and more about supply-chain hedging while Intel tries to prove 18A is real. (intel.com)
Apple and Intel are back in the same sentence — but not in the old “Macs use Intel CPUs” way. This time the question is manufacturing. A fresh supply-chain report says Apple i(trendforce.com)hip supplier. That matters because Apple has leaned heavily on TSMC for its most advanced chips, and because Intel’s whole foundry comeback depends on proving big outside customers will actually trust its leading-edge nodes. (trendforce.com) #(intel.com)facturing node. The base 18A platform is Intel’s big transistor reset — RibbonFET gate-all-around transistors plus PowerVia backside power delivery. The “P” version is the tuned-up follow-on, meant to squeeze more speed or lower power out of the same basic platform rather than deliver a full density shrink. Intel’s VLSI 2026 materials say 18A-P is design-compatible with 18A, which is important because it lowers the switching cost for any customer already exploring the base node. (vlsisymposium.org)just want raw speed — it wants laptops that stay cool, tablets that sip power, and desktops that still scale up. Intel’s own technical summary for 18A-P says the node can deliver over 9% higher performance at the same power, or over 18% lower power at the same performance, versus plain 18A. Those are the kinds of gains that get Apple’s attention even before you ask whether Intel can manufacture at scale. (vlsisymposium.org) ### Is this a don(vlsisymposium.org)Earlier reports tied the idea to lower-end M-series chips and a possible 2027 window, with Apple waiting for more mature design kits and better evidence on yields. Evaluation is real progress, but it is still a long way from a production commitment. (techpowerup.com) ### Why not just stay with TSMC? Apple probably still will, for most of its leading-edge volume. The point here is optionality. If Apple can qualify a(vlsisymposium.org)cally, even if Intel only wins a slice of future Apple silicon, that could still matter a lot. (intel.com) ### Why is this such a big test for Intel? Because Intel Foundry has spent years promising a comeback and now needs proof. Intel says Panther Lake is already in production on 18A, making that node the first real stress test for its (techpowerup.com)it can sell advanced process technology to the hardest customers in the industry. (intc.com) ### And what’s Google doing here? A parallel report says Google may use Intel’s EMIB advanced packaging for a future TPU generation. That is a different part of the stack — (intel.com)me direction. Intel is trying to win business piece by piece: process where it can, packaging where it can, and credibility everywhere. (trendforce.com) ### What’s the catch? Manufacturing claims are easier than manufacturing volume. Intel can show attractive technical numbers, but Apple wou(intc.com)is story is interesting now: the rumor is no longer just “Intel has a roadmap.” It is “Intel may finally have something Apple thinks is worth testing.” (vlsisymposium.org) ### Bottom line? This is a credibility story disguised as a process-node story. Apple evaluating Intel 18A-P would not mean TSMC is out. It would mean Intel has gotten far enough back into the game that Apple has to take the meeting seriously. (trendforce.com)
