On Tuesday, Intel chief executive Lip-Bu Tan confirmed that the semiconductor giant will “work closely” with Elon Musk to back the billionaire entrepreneur’s ambitious Terafab initiative—a large-scale chip development and manufacturing venture co-developed by SpaceX and Tesla. A snapshot shared via Intel’s official X account captured the two leaders shaking hands over the weekend, standing in front of a large branded Intel sign. Musk’s Terafab, a planned network of 1-terawatt, ultra-high-performance chip manufacturing facilities that may span multiple sites, is projected to cost billions of dollars to complete.

“Terafab marks a transformative shift in how silicon logic chips, memory modules, and chip packaging will be manufactured going forward,” Tan shared in a public social media post. “Intel is proud to step in as a partner and collaborate closely with Elon on this high-priority strategic project.”

Concrete details of how Tan and Musk plan to deliver on this massive, ambitious venture remain fully undisclosed. Musk has been publicly discussing the need for a dedicated Terafab operation for months, framing the project as a solution to meet the staggering chip demand his portfolio of companies requires for electric vehicles, humanoid robots, and AI data centers. Still, many chip industry analysts remain deeply skeptical that Musk can successfully pull off such a complex, capital-heavy undertaking.

For Intel, the potential partnership comes as the company fights to stage a major turnaround after years of market stagnation. A core pillar of its current comeback strategy is marketing its advanced semiconductor manufacturing capacity to tech companies scrambling to source chips to power the global AI boom. As WIRED recently reported, securing major external clients is make-or-break for Intel’s turnaround success—and Musk would represent one of the largest, most high-impact clients the company could win.

Musk did not reply to WIRED’s requests for comment on the new partnership. An Intel spokesperson directed WIRED to the company’s existing public social media posts about the agreement and declined to share any additional details. For now, we break down five key unanswered questions that will shape how Intel’s involvement impacts Terafab’s odds of success.

How Large Is This Actual Partnership?

That is still impossible to pin down. Neither Intel nor Tesla has filed any formal disclosure with the U.S. Securities and Exchange Commission (SEC)—a requirement that typically kicks in when a new partnership or deal would materially change a public company’s capital spending or manufacturing output.

For context, when chipmaker AMD and Meta announced a multiyear, multi-generation partnership in February 2024 to deploy up to 6 gigawatts of AMD GPUs for Meta’s AI infrastructure, AMD immediately disclosed the deal in an official SEC filing. As of this publication, no similar disclosure has been submitted by Intel or Tesla. That suggests the current agreement between Tan and Musk is little more than a handshake deal and preliminary alignment at this stage. As one anonymous chip industry insider put it: “It makes for a pretty great headline for a couple of days, doesn’t it?”

What Exactly Is Intel Bringing to the Table?

Intel’s public statements about its collaboration with Musk are notoriously vague, to the point of being almost comical. The company has only confirmed that its “proven ability to design, fabricate, and package ultra-high-performance chips at massive scale” will help Terafab hit its goal of generating 1 terawatt of annual computing power to support future breakthroughs in AI and robotics.

Pat Moorhead, a veteran chip industry analyst and founder of Moor Insights & Strategy, predicts Musk will first leverage Intel’s cutting-edge advanced packaging capabilities. He points out that Tesla “doesn’t need help with chip design engineering—they already have a very strong, capable in-house team for that work.” Moorhead adds that Musk may also seek to license Intel’s core chip architecture, which Terafab could then build on and customize for its own use cases.

Starting with advanced packaging is a low-risk first step for all parties involved, Moorhead notes, because it lets the partners test their working relationship without alienating TSMC, the world’s largest and most advanced contract chip manufacturer. “If you start with packaging, you won’t anger TSMC nearly as much as you would if you moved wafer production straight to Intel,” he explains. (Tesla already has existing chip manufacturing partnerships with both TSMC and Samsung.)

Moorhead says Musk’s long-term goal is almost certainly still to control as much of the entire chip production stack as possible, from initial design through final manufacturing, while also developing new proprietary methods for wafer production. But Moorhead and other analysts remain skeptical that building an entirely new, massive-scale fab that integrates every single stage of chip development and manufacturing is even feasible.

How Much Customization Will Musk Demand?

Tesla’s past work developing custom chips suggests the answer is: a lot. Last year, Tesla signed a $16.5 billion deal with Samsung to manufacture Tesla’s next-generation A16 chip at Samsung’s facility in Texas. But Tesla handled full chip design in-house, to make sure the component was perfectly tailored for its fleet of autonomous vehicles and humanoid robots. And according to Musk, “Samsung agreed to allow Tesla to assist in maximizing manufacturing efficiency.”

“This is a critical point, as I will walk the line personally to accelerate the pace of progress. And the fab is conveniently located not far from my house,” Musk posted on X around the time the deal was announced.

Chip industry analysts expect Intel will likely agree to a similar level of customization for Musk. “Technically, as an in-house chip design team, Elon and his group can customize their chips as much as they want,” says Austin Lyons, author of the Chipstrat newsletter and semiconductor analyst at Creative Strategies. “But the real question is whether Elon will want to customize the actual manufacturing process itself, for both wafers and packaging. And knowing Elon, I have no doubt he’ll push for full end-to-end process adjustments, and extremely aggressive development timelines.”

Who Controls the Project’s Intellectual Property?

While Intel has struggled in recent years to keep up with leading market competitors, it still operates a global network of manufacturing facilities and boasts decades of hands-on industry expertise that Terafab cannot build without. For Musk to access that manufacturing know-how, he will have to license Intel’s existing intellectual property.

According to Moorhead, that structure means Intel will almost certainly retain ownership of any intellectual property developed as part of the Terafab project. Musk would be able to develop his own proprietary “recipe” for chip manufacturing, but until his companies are in a position to purchase their own advanced chip-making hardware—like cutting-edge lithography machines—he will still need to license core manufacturing processes and specialized process design kits from an outside foundry.

Who Will Actually Construct the Terafab Facilities?

Widespread skilled labor shortages add another major layer of challenge to Musk’s goal of turning the Terafab vision into reality. Musk has not yet announced any formal locations for the new manufacturing facilities, but Tesla is already building a 2-million-square-foot chip design lab on its existing automotive campus near Austin, Texas—and the state has increasingly become the central hub for Musk’s sprawling global business empire, as well as his personal home.

Both Texas and much of the broader U.S. are facing critical shortages of skilled tradespeople, including plumbers and electricians, that are required to build data centers and semiconductor fabs. Chap Thornton, business manager for the Austin-area UA Plumbers & Pipefitters Local 286 union, says the data center industry currently outbids all other sectors for available skilled labor. Data center developers “aren’t afraid to pay premium rates to get work done on their tight timelines,” he says. “Any new construction project that pops up now is going to end up in a bidding war for workers.”

Construction on Tesla’s 10-million-square-foot Austin Gigafactory, which broke ground in 2020, was marked by grueling work schedules, multiple workplace injuries, and at least one worker death, according to safety regulators and accounts from on-site workers. With so many other job options available now, many construction workers may be reluctant to take on another high-pressure project for Musk. “Every worker who wants a job already has one,” Thornton says of his union’s roughly 2,000 active members.

Intel’s involvement could actually help Musk address some of these safety and labor challenges. A few years ago, Intel was one of the first major companies in Texas to eliminate seven-day workweeks for construction crews on its projects. “Productivity falls off a cliff when you force people to work seven days a week,” Thornton says. “Intel definitely has a proven track record of prioritizing worker safety on their job sites.”