Apple Strikes Preliminary Deal For Intel To Make iPhone And Mac Chips

Apple has reached a preliminary deal with Intel to manufacture some of the chips inside future iPhones, iPads, and Macs, ending nearly a decade of single-foundry dependence on Taiwan Semiconductor Manufacturing Company. The agreement, first reported by The Wall Street Journal on May 8, 2026, follows more than a year of talks pushed by the Trump administration and triggered by an A19 chip shortage that throttled iPhone 17 sales last quarter. Intel shares closed at an all-time high of $108.18 the same day, lifting the value of the U.S. government’s 9.9% stake to roughly $56.5 billion.

The Cupertino company is not abandoning TSMC. It is buying a hedge.

What the Apple-Intel Deal Actually Covers

Neither company has officially confirmed the agreement. Intel declined to comment. Apple did not respond. But the contours, drawn from a year of analyst leaks and the WSJ scoop, are unusually clear for a deal at this stage.

Intel will fabricate Apple’s lowest-end M-class processor first, on Intel’s 18A-P node at Intel’s 18A process page in Fab 52, Arizona. The chip targets the MacBook Air, entry-level MacBook Pro, and iPad Pro. According to Ming-Chi Kuo’s industry survey on X, Apple already signed an NDA, took delivery of the 18AP PDK 0.9.1GA, and is waiting on PDK 1.0 or 1.1 to lock the design. First commercial shipment is set for the second to third quarter of 2027.

The Pro and Max variants stay with TSMC. So do the A-series chips powering iPhone 17, iPhone 18, and the upcoming 2nm parts. Jeff Pu of GF Securities told clients the deal could later extend to non-Pro iPhone silicon on Intel’s 14A node in 2028, but that is a forecast, not a contract.

Volume, Yield, And Why Intel Says Yes To The Cheapest Chip

Industry estimates put the initial volume at 15 to 20 million chips a year. That is small against Apple’s 200-million-iPhone yearly run, but it is the largest external customer Intel Foundry has ever signed.

Yield is where the math gets interesting. Intel 18A yields at Fab 52 are tracking between 65% and 75% on representative dies. SemiAnalysis’s deep dive on the Apple-TSMC partnership pegs TSMC’s N3 at over 80% on a 150 mm² die. The gap is real. Apple is choosing the lowest-stakes silicon in its lineup precisely because a worse-than-expected wafer can be absorbed without bricking a flagship.

The Trillion-Dollar Backstory: A19 Chips Apple Couldn’t Get

This deal exists because of a single sentence Tim Cook said on Apple’s most recent earnings call. iPhone 17 demand was “off the charts.” Supply was not.

Cook conceded that iPhone 17 and iPhone 17 Pro shipments were held back by TSMC’s advanced-node capacity, with storage and memory also tight. Apple’s fiscal Q2 still posted $111.2 billion in revenue and roughly $57 billion in iPhone sales, with guidance for 14% to 17% growth next quarter. The constraint, in other words, was not weak demand. It was AI.

Nvidia’s Rubin GPU mass production has pulled enormous volumes of N3 wafers off the consumer queue. Apple, once TSMC’s largest customer by a wide margin, no longer holds that pricing or scheduling whip. The era when a Cupertino purchase order moved every other line item is over.

The constraint, as I had mentioned, is due to the advanced node capacity, and it is really a result of growing so well in Q1 with by 23% and having less flexibility, partly due to that in the process, to increase it as much as we would like to increase it.

That was Cook on the same call, quoted by analysts who attended. The unspoken half of the sentence: he needed a second supplier and he needed one in the United States.

Washington’s $47 Billion Win

The deal’s geopolitics outweigh its silicon. In August 2025, the Trump administration converted unpaid CHIPS Act grants into an Intel equity position tied to the $600 billion Apple manufacturing pledge, buying 433.3 million shares at $20.47 apiece. Total cost to taxpayers: $8.9 billion.

By the close of trading on May 8, 2026, that 9.9% stake was worth roughly $56.5 billion. The Treasury sits on $47.6 billion in unrealized gains in nine months. Commerce Secretary Howard Lutnick, who orchestrated both the equity conversion and the Apple introduction, met repeatedly with Cook, Elon Musk, and Jensen Huang to herd capital toward Intel’s order book. Lutnick’s August 2025 announcement of the U.S. Intel stake on X framed the original investment as national-security policy. Nine months later it looks like the most profitable trade the federal government has ever made on a single equity.

Apple’s own incentive runs through the same channel. Cook stood at the White House last August to announce the Apple American Manufacturing Program announcement raising the U.S. commitment to $600 billion. An Intel order for chips fabbed in Arizona is the most visible deliverable that program has produced.

The Intel CEO Who Closed It

Lip-Bu Tan, who replaced Pat Gelsinger in March 2025, has spent his first 14 months stitching customers onto Intel’s foundry roadmap. On Intel’s Q1 2026 prepared remarks, Tan told analysts the company is now arguing about how fast it can scale, not whether it can survive.

“A year ago, the conversation about Intel was about whether we could survive,” Tan said on the April 23 call. “Today it is about how quickly we can add manufacturing capacity and scale our supply to meet enormous demand.” Intel posted Q1 revenue of $13.6 billion, $1.4 billion above its own midpoint, with foundry revenue of $5.4 billion, up 20% sequentially. The catch: external foundry revenue was just $174 million. Apple, even at 15 to 20 million low-end M chips a year, multiplies that figure many times over.

Why Intel 18A-P, Not 18A Or 14A

The node selection is a tell. Intel 18A is the company’s first node with PowerVia backside power delivery and RibbonFET gate-all-around transistors, both of which arrive a few quarters ahead of TSMC’s equivalent. The 18A-P variant trims that further: roughly 8% better performance per watt at the same density, plus mobile-tuned threshold voltages for battery life. That last detail matters for an iPad Pro and a fanless MacBook Air.

Apple did not pick 14A because 14A will not reach production until 2028 and is reserved for Tesla’s Terafab project in Austin, where Musk has committed to next-generation chips for AI. The 18A-P node is the earliest sub-2nm process available on American soil, and that is a phrase Cook can use in any congressional hearing for the next four years.

The Yield Question Apple Has Not Answered

Bank of America’s Vivek Arya kept his Underperform rating on Intel after Q1 earnings, citing the absence of a confirmed major external wafer customer. The Apple deal answers that critique, but it does not yet answer whether Intel can hit yields on a chip Apple will actually ship at scale. The 65% to 75% number on test dies is not the same as the 80%-plus yield Apple needs on a 150 mm² mobile SoC.

If Intel misses, the 18A-P tape-out slips and Apple either delays the M7 baseline or quietly redirects the order back to TSMC. Both companies have an interest in not letting that happen. Neither has a track record together to suggest it cannot.

What Changes For Apple Buyers

Almost nothing visible. The reader holding an iPhone today will not notice an Intel-made chip when it arrives. Apple designs the silicon. Intel only prints it. Performance, battery, and thermals are governed by Apple’s architecture, the 18A-P process specs, and the firmware tuning, not by which campus the wafer came from.

The supply story is different. If Intel delivers, Apple gains a second source for entry-level Macs and iPads in a window where TSMC capacity is being eaten by AI buyers. That should ease the kind of constraint that hurt iPhone 17 availability and could keep MacBook Air prices stable into 2027 and 2028. If Intel stumbles, Apple’s $599 entry tier feels the squeeze first.

The Samsung Wildcard Nobody Is Pricing

Bloomberg’s Bloomberg’s earlier report on Apple’s Intel and Samsung chip plans noted Apple executives also visited a Samsung fab under construction in Texas. Samsung already won an Apple contract for CMOS image sensors at its Austin facility, breaking Sony’s decade-long lock on iPhone camera silicon. Industry estimates put Samsung’s potential CIS volume at 150 million to 200 million sensors a year by 2027, worth roughly $1 billion to $1.5 billion in foundry revenue.

That is the part of Apple’s diversification mainstream coverage has underplayed. The Intel headline is loud. The Samsung sensor deal is quietly already shipping. Combined with Intel’s 18A-P, Apple is rebuilding its chip supply chain on three legs instead of one, with Texas, Arizona, and Taiwan as the geographic anchors.

Frequently Asked Questions

Will My Next iPhone Use An Intel Chip?

No, not for at least two years and not in the Pro lineup. The first Intel-made Apple silicon is targeted for the entry-level M7 in MacBook Air and iPad Pro in the second half of 2027. iPhone chips, including the A19, A20, and the 2nm A21 expected in iPhone 18, stay with TSMC. Jeff Pu floated a 14A iPhone deal for 2028, but that is unconfirmed and tied to non-Pro models only.

Does An Intel-Made Mac Chip Mean Worse Performance?

No. Apple still designs every transistor. Intel’s role is to fabricate the design Apple hands over, just as TSMC does today. Intel 18A-P offers about 8% better performance per watt than the base 18A node and is broadly comparable to TSMC’s N3P on density. The bigger risk is yield variability in the first production runs, which would affect availability, not the speed of any chip you actually receive.

How Will This Affect MacBook And iPad Prices?

The deal is more likely to stabilize prices than cut them. Adding a second foundry gives Apple bargaining leverage and reduces single-source risk, which keeps entry-tier MacBook Air and iPad Pro pricing predictable through 2028. TSMC has signaled price increases of 3% to 10% on sub-3nm wafers through 2029, so without an alternative, low-end Mac prices would have drifted up. Expect stability, not discounts.

Why Did The U.S. Government Get Involved?

Commerce Secretary Howard Lutnick converted CHIPS Act grants into a 9.9% Intel equity position in August 2025 at $20.47 a share, then personally lobbied Tim Cook, Elon Musk, and Jensen Huang to anchor demand at Intel Foundry. The Apple deal lifted the Treasury’s stake from $8.9 billion to about $56.5 billion. Washington is both Intel’s largest shareholder and its most aggressive customer-acquisition team, with around $47.6 billion in paper gains so far.

Is Apple Leaving TSMC?

No. TSMC continues to make every Pro and Max chip across iPhone, iPad, and Mac, plus all A-series silicon and the upcoming 2nm A21. The Intel deal covers the lowest-end M chip first, with possible expansion to non-Pro iPhone parts in 2028 on the 14A node. Apple’s TSMC volume in Arizona and Taiwan still dwarfs everything moving to Intel by an order of magnitude.

The story under the story is simpler than the trillion-dollar headlines suggest. Apple needed a second foundry. Washington needed an Intel customer. TSMC needed someone to take the AI overflow. One agreement solved three problems in three capitals. Whether 18A-P holds yield in 2027 is the only question that still matters, and Cupertino has roughly 14 months to find out.