For nearly a decade, Apple has trusted exactly one company to build the brains of its iPhones, iPads, and Macs: Taiwan Semiconductor Manufacturing Company. That exclusivity may be loosening. According to a May 2026 Bloomberg report, Apple is exploring whether Intel and Samsung can manufacture its primary device processors inside the United States, giving the company a domestic production path it has never had before.
The timing is hard to ignore. Microsoft has already committed to Intel’s most advanced manufacturing technology, the 18A process node, for its own custom chip designs. Now Apple appears to be evaluating the same facility. Two of the world’s most valuable companies are, independently or not, testing whether Intel can deliver silicon that rivals what TSMC produces in Taiwan.
What we actually know
Bloomberg’s report, attributed to multiple people familiar with the deliberations, describes Apple’s outreach to Intel and Samsung as early-stage. No public statement from any of the three companies has confirmed active negotiations, contract terms, or production timelines. The framing is specific: Apple is looking at these foundries as secondary options alongside TSMC, not as wholesale replacements.
Intel’s own filings add important context. In its annual 10-K report for the fiscal year ended December 28, 2024, Intel named TSMC, Samsung, GlobalFoundries, UMC, and SMIC as its major foundry competitors. The filing details the risks Intel Foundry faces in courting outside customers: technology transition execution, pricing pressure, and the challenge of proving yield and reliability on new process nodes. Intel has publicly positioned 18A as the cornerstone of its bid to win external business, though the 10-K does not name specific customers or disclose order volumes tied to any particular node.
Microsoft’s booking of Intel 18A capacity has been widely reported in trade and financial press over the past year, and Intel executives have referenced external customer wins for the node during earnings calls. Still, no SEC filing names Microsoft or any other customer in connection with 18A orders. That distinction matters: the financial record confirms Intel’s foundry ambitions but stops short of the granular detail that would let analysts model near-term revenue from a single account.
Why Apple would look beyond TSMC now
Several forces are converging. Tensions across the Taiwan Strait have not eased, and both Republican and Democratic lawmakers have pressed U.S. tech companies to reduce their exposure to a single point of failure in East Asia. The CHIPS and Science Act, signed into law in 2022, created tens of billions of dollars in subsidies and tax credits specifically designed to pull advanced chip manufacturing onto American soil. Intel has been one of the largest recipients of those incentives, with new or expanded fabs planned in Arizona, Ohio, and Oregon.
TSMC itself is building advanced fabrication capacity in Phoenix, with its first Arizona fab producing chips on the 4-nanometer node and additional phases targeting more advanced processes. That partially addresses the geographic risk without requiring Apple to switch partners. But “partially” is the operative word: a single Arizona campus does not replicate the depth of TSMC’s Taiwanese operations, and a serious disruption in Taiwan would still ripple through Apple’s supply chain.
Apple also has history here. Before its exclusive relationship with TSMC began around 2016, Samsung fabricated Apple’s A-series processors. The partnership ended amid competitive tensions (Samsung makes its own smartphones) and TSMC’s growing process lead. A return to Samsung would be a reversal, but the calculus has changed: Samsung has invested heavily in its foundry-services division and is building a major fab in Taylor, Texas, though the company has faced well-documented yield challenges on its most recent nodes.
The technical bar Apple would set
Apple’s A-series and M-series chips are fabricated on TSMC’s most advanced nodes, where transistor density, power efficiency, and manufacturing yield are tightly optimized. Any alternative foundry would need to match or closely approach that performance. Intel’s 18A node, which uses gate-all-around transistor architecture and backside power delivery, represents a genuine technological leap for the company. But 18A is still ramping toward volume production, and Intel’s track record over the past several years has included delays and revised timelines on earlier nodes.
Scaling matters as much as technology. Apple ships hundreds of millions of iPhones per year, each containing a custom processor. Even if Intel proves the core 18A technology works, producing at the volumes and consistency Apple demands would stress any new foundry relationship. The same concern applies to Samsung, whose yield issues on cutting-edge nodes have been a recurring theme in industry reporting.
There is also the question of which chips Apple might entrust to a new partner first. A cautious approach would start with a lower-volume product line, perhaps processors for accessories or secondary devices, before graduating to flagship iPhone or Mac silicon. The Bloomberg report does not specify which designs Apple is considering, and that ambiguity leaves room for outcomes ranging from a modest pilot to a meaningful share of Apple’s total production.
The economics nobody has disclosed
Leading-edge chip fabrication requires enormous capital expenditure. A single advanced fab can cost $20 billion or more to build and equip. Foundry customers typically commit to multi-year volume and pricing frameworks to justify that investment. There is no public indication of whether Apple is prepared to underwrite new capacity at Intel or Samsung, whether it would seek additional government incentives, or whether it expects the foundries to shoulder most of the financial risk.
For Intel, landing Apple as a foundry customer, even for a secondary product line, would be a landmark validation of its turnaround strategy. Intel’s 10-K flags customer concentration risk and the difficulty of winning initial orders from new accounts. Converting reported interest into contracted revenue would give investors something concrete to model, rather than trading on expectations and press reports. That confirmation has not appeared in any public filing as of June 2026.
What this signals for the chip industry
Zoom out, and the pattern is striking. The two companies that buy more cutting-edge silicon than almost anyone else on Earth are both, in different ways, testing whether Intel can serve as a credible alternative to TSMC. Microsoft has moved past the testing phase and booked capacity. Apple is still in the exploration stage. Neither move, on its own, proves that a broad migration away from TSMC is underway. TSMC’s technology leadership remains formidable, and its Arizona expansion gives U.S. customers a partial domestic option without switching foundries.
But the direction of travel is clear. The semiconductor industry spent two decades consolidating leading-edge manufacturing into a single company on a single island. The largest buyers of that manufacturing are now, quietly and methodically, looking for a second option. Whether Intel and Samsung can actually deliver one is the question that will define the next chapter of chipmaking. The answer will not come from anonymous sourcing or analyst speculation. It will come from production lines, yield data, and signed contracts, none of which are public yet.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
