Apple signed an agreement with Intel in December 2025. The M7 is reportedly heading to Intel 18A-P. Here is what that actually means.
There is a certain irony in this story that is pretty hard to ignore. Apple spent the better part of a decade using Intel processors in its Mac lineup, growing increasingly frustrated with Intel’s inability to meet its performance and power targets, and ultimately replaced the entire architecture with its own Apple Silicon in 2020. That transition turned out to be one of the more consequential moves in consumer computing in recent memory. And now, just six years later, Apple is reportedly going back to Intel. Not for the x86 processors it spent so long moving away from, but to use Intel’s semiconductor manufacturing to make its own chips.
According to notes from a GFHK Monthly Call shared by leaker Jukan on X on May 13th, Apple and Intel signed a formal agreement back in December 2025. The M7, which would power Apple’s next MacBook lineup, is expected to use Intel’s 18A-P process technology and enter mass production by the end of 2027. A future smartphone chip is also reportedly in the pipeline on Intel’s 14A node, with mass production targeting the end of 2028.
This is, at this stage, a leaker report rather than anything officially confirmed by either Apple or Intel, and that caveat matters. But the direction of travel is consistent with a number of previous reports, including analysis from Ming-Chi Kuo and GF Securities analysts Jeff Pu and Evan Lee, who have been pointing at an Apple-Intel foundry relationship for some time. The sourcing is credible enough to take seriously, and it is worth going through what it actually means in some detail.
Why Apple Is Doing This
The straightforward answer is supply. Apple’s former CEO Tim Cook explicitly cited the inability to meet high demand for Apple products as a direct consequence of insufficient chip capacity. The AI boom has made TSMC’s leading-edge capacity considerably more contested than it was even two years ago. Every major AI company is essentially competing for N2 and N3P wafer starts, and TSMC’s ability to prioritise any single customer, even one as historically significant as Apple, has been constrained by that demand. Apple’s annual investment at TSMC grew from around $2 billion in 2014 to $24 billion in 2025, which gives some sense of how dependent that relationship had become. Diversifying away from a single foundry at that scale is not a quick process, but it is clearly one Apple has decided is necessary.
There is also the geopolitical dimension. The US government’s push to onshore semiconductor manufacturing has created real incentives for American companies to move at least some of their production to domestic fabs. Reports suggest Washington played a fairly active role in encouraging Apple to consider alternatives to TSMC. Intel’s Fab 52 in Arizona, where 18A is currently in volume production for Panther Lake, is precisely the kind of domestic advanced manufacturing capacity that policy is designed to promote. Apple aligning some of its production with Intel’s Arizona fabs makes a lot of sense in that context.
What Intel 18A-P Actually Is
Intel’s 18A process is the company’s 1.8nm-class node, the one that Panther Lake, which is the Core Ultra Series 3 processor that launched at CES 2026, is currently manufactured on. It combines RibbonFET gate-all-around transistors with PowerVia backside power delivery, and this delivers up to 25% higher performance or 36% lower power consumption compared to Intel’s previous Intel 3 process, with around 30% improved transistor density.
The 18A-P variant, which is what the M7 is reportedly targeting, is essentially a refined version of the base 18A node. Ben Bajarin of Creative Strategies described the base 18A as somewhat rough, with 18A-P being the version that addresses those issues, with PDK versions 1.0 and 1.1 having arrived in Q1 2026. That refinement matters considerably for Apple, which has famously demanding requirements around yield and process consistency. The fact that the M7 is reportedly targeting 18A-P rather than base 18A suggests Apple’s engineers looked at the initial node characterisation and simply decided to wait for the cleaner version, which is arguably the right call.
For the M7 specifically, Apple may require some of Intel’s advanced packaging techniques to hit its performance targets, potentially involving combinations from the Foveros family alongside EMIB. Apple’s own packaging work through TSMC has been sophisticated for some time, so the expectation that Intel’s packaging needs to match that level is a pretty reasonable one.
The Smartphone Side of Things
The iPhone angle is somewhat less clear. The first mobile candidate may not be the A21 for the iPhone 19 in 2027. Current reports instead point toward an unnamed SoC targeting mass production on Intel’s 14A node by late 2028, which suggests Apple will continue relying on TSMC for its smartphone chips for the immediate future, at least for the higher-end models. Whether this smartphone chip is a non-Pro model, leaving the Pro lineup on TSMC, or something else entirely, is not yet determined.
Intel’s 14A is the node beyond 18A. CEO Lip-Bu Tan has been fairly explicit that external customers need to commit before Intel ramps 14A production, which makes commercial sense for a company trying to justify the capital expenditure of building out an entirely new node. Apple is likely to wait for the final 14A PDK before beginning trial production, given that the node is still roughly two years from mass production readiness.
The Competitive Irony
It is worth pausing on what this arrangement actually means competitively. Intel will be manufacturing chips for Apple’s Mac lineup, which is in direct competition with Intel’s own Core Ultra PC platform. Intel will be producing chips for a product that directly rivals its own PC business. That is a genuinely unusual dynamic and one that Intel’s foundry leadership has presumably thought carefully about. The logic, from Intel’s perspective, is that foundry revenue is foundry revenue, and winning Apple as a customer gives the foundry division a credibility that no internal chip launch can simply provide on its own.
Apple is one of the most demanding chip customers in the world, and winning any Apple production gives Intel Foundry a proof point that other potential customers, including Amazon, Qualcomm and Broadcom, will pay close attention to. Every M7 that meets Apple’s performance targets is essentially a data point for every other company evaluating Intel Foundry as an alternative to TSMC.
What to Watch
Intel’s 18A-P yield rates through the rest of 2026 are arguably the most underweighted signal in this whole story. If yields disappoint, the 2027 M7 timeline moves right with them, and the whole framework shifts accordingly. Any announcement of a lower-cost MacBook variant in Apple’s 2027 product roadmap would be consistent with the M7 on 18A-P story, since entry-level Mac hardware is reportedly the likely initial target for Intel-fabbed chips rather than the Pro lineup.
The December 2025 agreement date is also worth noting in itself. That is five months ago, which means both companies have presumably been working through the technical details of the arrangement for some time. At that stage of engagement, a formal agreement suggests this is considerably further along than a typical exploratory conversation.
The Bigger Picture
If this plays out as reported, what it represents is something fairly significant for the foundry industry as a whole. TSMC has effectively held a monopoly on the world’s most advanced consumer chip manufacturing for several years. Apple has been its most prominent customer and its most visible proof point. If Apple begins splitting its most advanced chip production between TSMC and Intel, it changes the competitive dynamics of the foundry business in a way that matters well beyond these two specific companies.
It is also, in a somewhat strange way, a validation of everything Intel has been building toward since Lip-Bu Tan took over. Whether 18A-P delivers what Apple needs and whether 14A ramps cleanly enough to power iPhone silicon by 2028 are questions that will take years to answer. But the fact that Apple is reportedly at the table at all says quite a lot about where Intel Foundry currently stands.
All details in this piece are based on a leaker report from Jukan sharing notes from a GFHK Monthly Call, published May 13th 2026. Neither Apple nor Intel has officially confirmed anything referenced above.
