Apple Explores Intel, Samsung Foundries to Cut Taiwan Chip Dependency

Apple has opened exploratory discussions with Intel and Samsung to manufacture its main processors in the United States, marking the first serious attempt to diversify beyond its exclusive reliance on Taiwan Semiconductor Manufacturing Company.

The talks, confirmed by Bloomberg, remain at early stages with no orders placed. Apple executives have visited Samsung’s Taylor, Texas facility currently under development and held preliminary conversations with Intel about foundry services for A-series and M-series chips — the processors that power iPhones, iPads, and Macs. Per Reuters, neither effort has yet resulted in production commitments.

The move signals Apple views Taiwan concentration risk as material enough to pursue alternatives with inferior track records. Nearly all of Apple’s chip production happens on a single island 100 miles from mainland China, with less than 5% of leading-edge output currently coming from Arizona, according to SemiAnalysis. TSMC produces 92% of the world’s advanced chips at 7nm and below — a concentration that analysts warn could impose $2.5 trillion in annual losses on the global economy if disrupted.

CHIPS Act Creates Viable US Alternatives

The conversations are enabled by $39 billion in CHIPS Act subsidies and 25% investment tax credits that have catalysed over $640 billion in US semiconductor investments, per the Semiconductor Industry Association. Intel’s 18A process — shipping late 2026 — represents the first theoretically viable alternative since Apple left Samsung a decade ago. Yields have improved to over 60%, compared to TSMC’s 70-80% when launching 2nm nodes, according to KeyBanc Capital Markets analysis.

Samsung’s Texas facility received equipment move-in in April and targets late-2026 production launch with 50,000 wafer starts per month capacity using SF2/SF3P process technology, Basenor reported. The plant received $4.745 billion in direct CHIPS Act funding and represents over $37 billion in Texas-region investment, per NIST disclosures.

The Geopolitical Calculus

Apple’s willingness to explore inferior alternatives validates the strategic premise behind US semiconductor policy: geographic concentration in Taiwan poses unacceptable risk regardless of TSMC’s technical superiority. Academic research published in ScienceDirect found Taiwan’s semiconductor Supply Chain would be particularly vulnerable to disruption before 2027, with diversification and stockpiling strategies impractical in the short term.

Analyst Ming-Chi Kuo noted Apple’s exploratory talks show “strong support for the Trump administration’s strongly promoted ‘Made in USA’ policy” while acknowledging the company “needs to secure a second source to meet supply-chain management requirements” despite remaining “highly dependent on TSMC’s advanced nodes for the foreseeable future.”

KeyBanc Capital Markets analysts project Intel has secured Apple as a customer on 18A for low-end M-series processors destined for MacBooks and iPads, with production expected in 2027. The engagement would give Apple a domestic fallback option while Intel validates its foundry model with the industry’s most demanding customer. Samsung’s SF2 process currently yields below 40% — well below commercial viability for Apple’s volumes — but the Texas facility offers geographic diversification even if technical parity remains distant.

TSMC’s Dominant Position Holds

TSMC reported record Q1 2026 revenue of $35.6 billion with 35% year-over-year growth, demonstrating sustained customer loyalty despite geopolitical concerns. March revenue hit 415.2 billion new Taiwan dollars, a 45.2% annual surge marking the eighth consecutive quarter of growth. Gross margins are projected at 64% — evidence of pricing power that reflects irreplaceable technical leadership.

The exploratory nature of Apple’s discussions underscores the gap between geopolitical necessity and commercial readiness. Intel and Samsung must prove not just process capability but also capacity, yield consistency, and ability to execute product transitions on Apple’s aggressive timelines — capabilities TSMC has refined over decades of exclusive partnership.

What to Watch

Monitor Intel 18A yield progression through late 2026 — sustained improvement above 70% would make 2027 M-series production credible. Track Samsung Texas qualification timelines and whether SF2 yields reach commercial viability (>50%) by year-end. Watch for Apple’s Arizona TSMC capacity allocation decisions as the Phoenix fabs ramp 3nm production — any reduction in Arizona orders would signal confidence in domestic alternatives remains limited. Geopolitical risk pricing in semiconductor equities will likely compress if Apple formalises Intel or Samsung orders, validating the CHIPS Act’s strategic thesis that subsidised onshoring can overcome Taiwan dependency despite technical tradeoffs. Any production awards before 2027 would accelerate the sector’s geographic rebalancing and pressure other fabless chip designers to diversify beyond Taiwan-concentrated supply chains.