Apple and Intel are reportedly close to a deal in which Intel will manufacture some of the chips for the iPhone maker’s devices, marking a major shift in the chip manufacturing landscape.
The Wall Street Journal reported Friday, citing people familiar with the matter, that the two companies have been negotiating for more than a year and reached a tentative agreement in recent months.
Intel stock soared nearly 14% on Friday. Apple stock rose 2%. Both companies declined to comment.
“I 100% believe this will happen. We don’t know when,” Ben Bajarin, a chip analyst at Creative Strategies, said in an interview.
If it goes through, the deal would be the most notable vote of confidence yet for Intel’s once-struggling chip foundry business. Intel stock is up more than 200% this year.
For Apple, it would be the end of an era. The iPhone maker currently relies solely on Taiwan Semiconductor Manufacturing Co., Ltd. to manufacture all cutting-edge chips for its devices.
But as demand for AI chips soars and every big tech company goes into a semiconductor frenzy, TSMC’s wafer production capacity is limited. Apple is no exception, ramping up its in-house silicon program in recent years to make nearly all the core chips in things like iPhones and Macs. Bajarin said Apple is TSMC’s second-largest customer, topped only by Nvidia.
“Intel is the only place we can expand capacity as a viable second source,” Bajarin said.
In fact, Intel is rapidly ramping up production capacity, with mass production currently underway at a new chip manufacturing facility in Chandler, Arizona. The company manufactures chips at 18A, its most advanced node or production process, and is intended to rival TSMC’s 2nm node, which is currently only manufactured in Taiwan. TSMC also has several new chip factories in Arizona, where Apple is working to make some of its silicon.
Bajarin said Apple will most likely wait to manufacture chips on Intel’s next node, 18A-P, and could scale up as early as next year. He said Intel’s current 18A node is “a little rough,” and that 18A-P is “cleaning up a lot of things.”
Intel’s foundry business has long faced delays and low yields that have called into question its ability to make chips for other companies. For now, Intel is the only major customer for the foundry business, making central processing units and other chips for its devices.
Bajarin said those days are over.
“We believe they have overcome the challenges and are now verified as a reliable secondary source of information,” he said.
Intel’s only major external customer commitment for foundries is unlikely to see real results until 2029 or later.
Elon Musk said last month that his $119 billion Terrafab in Austin, Texas, will rely on Intel’s future 14A chip nodes. The Terrafab is intended to make chips for Tesla, SpaceX, and SpaceXAI. Intel CEO Lip-Bu Tan said in February that 14A is expected to go into mass production in 2029.
Intel already has major customers such as Amazon and Cisco, and the advanced packaging side of its chip manufacturing business involves gluing individual chip dies and memory together to make things like graphics processing units.
The Apple-Intel partnership won’t affect TSMC, Bajarin said, because “they’re already printing wafers as quickly as possible.” Still, TSMC changed its word last month, with President and CEO CC Wei calling Intel a “formidable competitor.”
“If you’re trying to get one of your biggest customers to sign a deal with a foundry that’s probably a competitor, you’re probably going to say something like that to soften the blow,” Bajarin said.
Apple executives also reportedly visited Samsung’s new chip manufacturing factory under construction in Texas, and CNBC got an early tour there. There are only three companies in the world capable of manufacturing the cutting-edge chips needed for AI: Samsung, Intel and TSMC, and “nobody can do it fast enough,” Bajarin said.
Watch: How Samsung became the world’s second-largest advanced chipmaker
