18A-P: Intel eyes Apple, beacon of hope or last cartridge?

When Intel and Apple sit down at a table, it’s not about small change. According to well-known analyst Ming-Chi Kuo, Apple is on the verge of using Intel’s new 18A-P process for future MacBook and iPad chips and this would not just be a technical statement, but a first-class geopolitical one. In an industry that is currently dominated by TSMC like Blockbuster used to dominate VHS tapes, a deal between Apple and Intel would be tantamount to an earthquake. And one with a political package insert: “Made in the USA”, strategic independence, diversification. But before you open the champagne: it hasn’t been signed (yet). And the conditions are anything but simple.

Intel 18A-P: The beast in the shadows

What is this 18A-P anyway? Officially presented at “Direct Connect 2025”, it is a derivative of the upcoming 18A node, refined for power optimization and with support for Foveros Direct, Intel’s variant of 3D hybrid bonding, where chiplets with less than 5 micron pitch can be stacked. With through-silicon vias (TSVs), the whole thing should also be efficiently interconnected. This sounds like music to Apple’s ears. Cupertino loves compact, efficient designs with high integration density. This makes the combination of 18A-P and Foveros particularly exciting for entry-level M-series SoCs, which could be installed in cheaper iPads and MacBooks.

According to Kuo, Apple has already signed an NDA with Intel, a rare move for a company that normally handles everything through TSMC. Apple is said to have secured access to PDK 0.9.1GA – an early design kit for the process. PDK 1.0/1.1 is expected for Q1 2026. Only then will it be decided whether the flirtation will turn into a serious relationship. Production would be scheduled for 2H 2027 at the earliest, with an estimated 15 to 20 million units for entry-level Macs and iPads. For comparison: TSMC manufactures Apple’s high-end chips (such as M3 or A18) in significantly larger quantities. Apple wants to diversify. Not to leave TSMC, but to cushion geopolitical risks. Taiwan remains a powder keg. And US-based production, especially at a company like Apple, is a diplomatic pawn.

Why Apple is (perhaps) joining Intel

• Technologically appealing: 18A-P offers exactly what Apple needs for its low-end designs, power density, power saving optimization, 3D stacking.
• Strategically clever: US production offers geopolitical security and PR points (“Made in America”).
• Diversification with a perspective: TSMC remains a given, but anyone who manufactures products worth billions needs a plan B. And plan C.
• Exclusivity: According to leaks, Apple is said to have exclusive access to 18A-P, which smacks of special conditions and control.

Conclusion: A beacon of hope with an expiration date?

Nothing is final yet. The Intel Foundry has promised more than it has delivered in the past. Gelsinger himself has admitted it: no discipline, no product on time, delay festival at the 18A node. Even now, HVM (High Volume Manufacturing) for 18A is planned for Q4 2025 at the earliest. Will yield, defect density and packaging be mature enough for Apple by then? Doubtful. And Apple is notorious for imposing surgically precise quality requirements on manufacturing partners. If Intel delivers even 3% below TSMC in one area, that’s it for the deal.

Intel 18A-P is a promising attempt to finally become a first-class foundry customer again, with a customer that stands for prestige, volume and political backing. But Apple is no Samaritan. If 18A-P does not deliver on time and with high performance, Cupertino will move on to the next supplier. And TSMC is not dead, but the undisputed leader. But if Intel delivers, it would probably be the biggest foundry coup in years and a liberating blow for a company that has long struggled more with itself than with the market.

Source: Wccftech