Intel splits itself in two to aid CEO Pat Gelsinger’s turnaround plans—and links up with archrival Arm

“This is a bit of ‘strange bedfellows,’” said Arm CEO Rene Haas yesterday … onstage at an Intel event. No kidding.

Once the undisputed leader in computer processors that weren’t for mobile handsets or embedded devices, Intel’s x86 architecture is increasingly being challenged by Arm’s more efficient architecture everywhere from laptops to data centers (even Nvidia, whose stock continues to soar on the back of AI mania, is using Arm’s architecture in its latest AI data center processors). So, over the past few years, Intel CEO Pat Gelsinger has tried to pivot his company toward a foundry model, i.e., making chips for others, like industry leader TSMC does. And yesterday, Intel took a huge step in that direction by effectively splitting itself in two.

Under the dramatic reorganization and rebranding, we now have Intel Foundry and Intel Product—separate legal entities that remain parts of the same overall company but even have their own sales forces and back-end business systems. Intel Foundry will obviously make the processors that Intel Product develops but via arm’s-length transactions. Intel Foundry’s other customers get strict confidentiality, and everyone’s welcome. “If we’re going to be the Western foundry at scale, we can’t be discriminating in who’s participating in that,” Gelsinger said yesterday. Of course, this wouldn’t be a 2024 chip story without copious lashings of AI, so he described Intel Foundry as “the world’s first systems foundry for the AI era.”

Despite early customer wins a couple years back in the form of Qualcomm and AWS, Intel’s self-prioritization has so far largely held back the company’s foundry hopes. With this clear separation of church and state, Gelsinger hopes everyone will now flock to Intel. The ambition here is stark: He wants Intel to be No. 2 to TSMC by 2030, which would mean leapfrogging the likes of China’s SMIC and South Korea’s Samsung. And that requires linking up with Arm, whose architecture is being used by many potential Foundry customers.

The details on Intel and Arm’s vaguely titled “Emerging Business Initiative” remain scarce, but the gist is that Intel will be working closely with startups developing their own Arm-based system-on-chip processors, “offering essential IP, manufacturing support, and financial assistance,” according to a statement given to Fierce Electronics.

Describing Intel’s incoming chipmaking technologies as “industry-changing,” Haas said Arm had to be “a part of it … We need total transparency, seamless communication. We need to act like we’re working with Intel Foundry and not the side of the house that we might consider a competitor, and you guys have been terrific,” Haas told Intel Foundry general manager Stuart Pann, with whom he shared the stage.

Gelsinger used the Intel Foundry Services Direct Connect event to show off wafers of the company’s upcoming Clearwater Forest processors, which will be made using Intel Foundry’s online-next-year 18A manufacturing process—with which Intel hopes to overtake TSMC as maker of the world’s fastest chips. But he also announced a potentially major new customer: Microsoft, whose CEO, Satya Nadella, briefly popped up on video to lend his support to Intel in its efforts to build up its U.S. supply chain (Intel’s reportedly in line for $10 billion in CHIPS Act onshoring subsidies) and to mention that Intel will build a Microsoft chip on the 18A process—and yes, that’s about as much detail as we have on that front, though Microsoft has recently said it plans to make an AI chip and a computer processor.

Gelsinger also chatted with OpenAI CEO Sam Altman, who reiterated for the millionth time how the world needs loads more AI chips. Who knows? If Altman succeeds in getting funding for his new chip project, maybe Intel will make the things. More news below.

David Meyer

Want to send thoughts or suggestions to Data Sheet? Drop a line here.

This story was originally featured on Fortune.com