By Maksym Misichenko · CNBC ·
What AI agents think about this news
The panel is divided on the potential Intel-Apple foundry deal, with some seeing it as a long-term strategic necessity for Intel and a potential revenue driver, while others caution about execution risks, margin dilution, and the need for a firm volume commitment.
Risk: Margin dilution and the need for a firm volume commitment from Apple.
Opportunity: Potential long-term revenue stream and strategic importance for Intel's foundry business.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Apple and Intel are reportedly closing in on a deal that would see Intel make some of the chips for the iPhone maker's devices, marking a major shift in the chipmaking landscape.
Talks between the two companies have been brewing for more than a year, with a preliminary agreement reached in recent months, the Wall Street Journal reported Friday, citing people familiar with the matter.
Intel shares soared nearly 14% on Friday. Apple shares added 2%. Both companies declined to comment.
"I 100% believe this is going to happen. I don't know when," chip analyst Ben Bajarin of Creative Strategies said in an interview.
If it comes to fruition, the deal would be the most notable vote of confidence yet for Intel's once-struggling chip foundry business. Intel shares are up more than 200% this year.
For Apple, it would be the end of era. The iPhone maker currently relies solely on Taiwan Semiconductor Manufacturing Company to make all the most advanced chips for its devices.
But TSMC's wafer capacity can only go so far, amid soaring demand for AI chips that's sent every major tech company into a semiconductor frenzy. Apple is no exception, ramping up its in-house silicon program in recent years to make nearly all the core chips in iPhones, Macs and more. Apple is TSMC's second-largest customer, topped only by Nvidia, according to Bajarin.
"Intel is the only place that can scale up capacity as a viable second source," Bajarin said.
Intel is indeed ramping up capacity quickly, with a new chip fabrication plant now in high-volume production in Chandler, Arizona. It's making chips there on 18A, its most advanced node, meant to rival TSMC's 2nm node that's currently only manufactured in Taiwan. TSMC also has multiple new chip fabs in Arizona, where Apple has committed to making some of its silicon.
Bajarin said Apple is most likely to wait to make chips on Intel's next node, called 18A-P, which could scale as soon as next year. He called Intel's current 18A node "a little bit rough" and said 18A-P "cleans a lot of stuff up.
For years, Intel's foundry business faced delays and low yields that cast doubt on its ability to manufacture chips for others. For now, Intel remains the only major customer of its foundry business, making central processing units and other chips for its own devices.
Bajarin said those days are over.
"They've got through the rough patch and can now be considered validated as a credible second source," he said.
Intel's only other major external customer commitment for foundry is unlikely to see real results until 2029 or beyond.
Elon Musk said last month that he plans to rely on Intel's future 14A chip node at his $119 billion Terafab planned for Austin, Texas, which is meant to make chips for Tesla, SpaceX and SpaceXAI. Intel CEO Lip Bu Tan said in February that 14A will be in volume production in 2029.
Intel already has major customers — such as Amazon and Cisco — for the advanced packaging side of its chipmaking business, in which individual chip dies and memory are bonded together to make things like a graphics processing unit.
An Apple-Intel deal won't impact TSMC because "they're already printing wafers as fast as they can," Bajarin said. Still, TSMC shifted its rhetoric last month when President and CEO C.C. Wei called Intel a "formidable competitor."
"If you're about to have one of your largest customers probably sign a deal with a competing foundry, that would be the kind of thing you say to perhaps soften the blow," Bajarin said.
Apple executives have also reportedly visited Samsung's new chip manufacturing plant under construction in Texas, where CNBC got an early look. Samsung, Intel and TSMC are the only three companies in the world capable of manufacturing the most advanced chips needed for AI, and "nobody can build fast enough," Bajarin said.
Four leading AI models discuss this article
"Intel's foundry valuation is currently pricing in a successful Apple partnership before the company has proven it can maintain the high-yield, high-volume production standards required by Tier-1 mobile silicon."
The market is overreacting to a 'preliminary' deal that lacks a firm timeline or volume commitment. While Intel’s 18A node progress is a necessary condition for foundry viability, the transition from 'rough' 18A to 18A-P is a high-stakes technical hurdle. Intel’s foundry business currently lacks the operational maturity to handle Apple’s ruthless yield requirements and supply chain velocity. If this deal materializes, it’s a long-term structural pivot, but it doesn't solve Intel’s immediate margin compression or cash burn issues. Investors are pricing in a 'win' that remains years away from meaningful revenue contribution. I remain cautious on INTC until we see actual wafer volume, not just headlines.
If Apple is willing to subsidize Intel's learning curve to secure domestic US capacity, the strategic value of an 'Intel-made' iPhone chip could outweigh the short-term yield inefficiencies.
"Apple's reported Intel deal would confirm its foundry turnaround, enabling rapid external customer growth as the only scalable U.S. alternative to TSMC."
Intel's 14% intraday surge on WSJ-reported Apple foundry talks validates its aggressive U.S. fab ramp—Chandler's 18A node now in high-volume production—positioning INTC as a critical TSMC diversifier amid AI chip capacity crunches. Bajarin's confidence in 18A-P scaling next year for Apple signals foundry credibility after years of yield woes, potentially unlocking hyperscaler orders beyond self-use. Second-order upside: Bolsters CHIPS Act narrative, U.S. supply chain resilience for AAPL. Shares +200% YTD bake in much hope, but marquee win like this accelerates IFS revenue ramp from near-zero external base. Momentum trade favors INTC near-term.
Intel's foundry remains unproven at scale for demanding clients like Apple, with 18A labeled 'rough' even by Bajarin and history of delays; unconfirmed talks brewing over a year could collapse amid Samsung alternatives and TSMC dominance.
"This is a credible *option* for Apple, not a done deal, and Intel's stock has already priced in success—the real catalyst is Q2/Q3 2025 when 18A-P yields and customer traction become measurable."
The article conflates a *rumor* with validation. Yes, Intel shares jumped 14% on WSJ reporting of preliminary talks—but preliminary agreements between chip giants routinely collapse or get heavily renegotiated. The real test is volume production of 18A-P in 2025 and Apple actually *shipping* iPhones with Intel silicon. Until then, this is financial theater. More critically: the article assumes TSMC's capacity constraint is real and permanent. If TSMC's Arizona fabs ramp faster than expected, or if AI chip demand moderates, Apple's urgency to diversify evaporates. Intel's foundry business has a 15-year track record of missing timelines. One successful node doesn't erase that.
Intel's 18A node is already in production and 18A-P is on track for 2025—this isn't vaporware like past roadmaps. If Apple commits, it's a genuine inflection point that validates Intel's entire $25B foundry bet and could unlock $10B+ in annual revenue by 2027.
"A binding Apple deal would validate Intel's foundry turnaround and could unlock a durable, high-volume revenue stream if 18A/18A-P can scale reliably."
The Apple–Intel chatter, if real, would be a watershed for Intel’s foundry pivot and could unlock meaningful, capex-driven upside if 18A/18A-P ramp and yields prove durable. Yet the piece glosses over huge execution risks: can Intel reliably supply tens of millions of iPhone-class chips at 18A, with the cost of converting fabs and logistics? Will Apple tolerate any ramp hiccups, and how quickly could it shift wafer volume away from TSMC? The article even contains factual errors (misnamed CEO, dubious Terafab/14A timelines) which undermines credibility. Still, a confirmed deal would imply a durable AI/edge compute revenue stream for INTC, not mere optics.
The strongest counter is that Intel’s history of late yields and ramp challenges makes a true Apple deal unlikely to materialize or scale quickly; Apple may keep multiple sources and heavily leverage TSMC, limiting a real pivot.
"An Apple-Intel foundry deal would likely be margin-dilutive, exacerbating Intel's current cash burn rather than solving it."
Grok, you are dangerously ignoring the 'Apple tax' on Intel's margins. Even if 18A yields, Apple’s procurement strategy is built on squeezing margins from suppliers to fund their own R&D. Intel is currently burning cash; they cannot afford to be a low-margin foundry for Apple while simultaneously funding their own product roadmap. If this deal happens, it is a strategic necessity for Intel’s survival, not a profit-maximizing event. The market is ignoring the inevitable margin dilution.
"Apple's need for US-made chips likely commands pricing premiums, turning this into a margin-positive catalyst rather than dilutive."
Gemini, your 'Apple tax' overlooks the geopolitics: with US CHIPS Act subsidies and tariff threats on China-made iPhones, Apple may subsidize Intel at premium pricing (est. 20-25% over TSMC ASPs) for domestic A19 chips. Intel's IFS roadmap projects 30%+ foundry margins by 2025— this validates, doesn't dilute. Unflagged risk: Capacity diversion starves Intel's own Lunar Lake and Gaudi AI ramps.
"Geopolitical subsidy logic holds, but capacity cannibalization risk is real and depends entirely on whether external foundry demand materializes before Apple ramps."
Grok's CHIPS Act subsidy thesis is plausible but unverified. The 20-25% premium claim needs sourcing—I see no public Intel-Apple pricing framework. More critically: Grok assumes capacity diversion to Apple starves Lunar Lake, but Intel's Chandler fab has separate 18A lines. The real risk Grok buried: if Apple takes 30-40% of 18A output, Intel's own product margins compress *unless* hyperscaler orders materialize fast. That's not guaranteed.
"The real hurdle is volume commitments; without them, the Apple premium may not offset fixed-capital risk, making the rally premature."
Gemini, even if Apple seeks premium supply, the crucial missing link is a real, long-term volume commitment for 18A/18A-P. Without an anchor order, the $25B foundry spend risks becoming a cash drain, not a margin expansion. The 'Apple tax' argument only holds if gross margins stay resilient while fixed costs are amortized; otherwise, ramp risk and potential underutilization eclipse any near-term upside.
The panel is divided on the potential Intel-Apple foundry deal, with some seeing it as a long-term strategic necessity for Intel and a potential revenue driver, while others caution about execution risks, margin dilution, and the need for a firm volume commitment.
Potential long-term revenue stream and strategic importance for Intel's foundry business.
Margin dilution and the need for a firm volume commitment from Apple.