Intel Terafab Deal|14A License or Foundry Win in Musk's $119 Billion Fab?

The promise that calmed Intel

Elon Musk is recruiting chip engineers for Tesla and SpaceX's Terafab, where Intel is the only named manufacturing partner. The promise that calmed Intel investors, that Musk's companies will run the fab themselves, is what leaves Intel's role unclear. Terafab is a planned chip complex in Grimes County, Texas. Spending across all its phases could reach 119 billion dollars.

Intel's next-generation manufacturing process, called 14A, is the only technology the plant has identified. Musk's reassurance is best read as keeping Intel's 14A process inside Terafab. That is not the same as making Intel the plant's manufacturer. For a chipmaker, those are two very different jobs. The worry started with a newsletter.

Culpium, run by tech journalist Tim Culpan, reported that TSMC was exploring ways to help operate Terafab's Texas factories. Musk replied on X: “Just discussions, but something may come of it.” Intel shares fell 5.7 percent over the following Monday and Tuesday. On Wednesday, two statements arrived. Intel CEO Lip-Bu Tan told Bloomberg that Intel would keep working with Musk on Terafab.

Musk wrote that TSMC might sublease part of the site, but nothing more. Intel's shares were flat that afternoon, while the PHLX Semiconductor Index fell 1.6 percent. Invezz, an investing site, said Musk's words removed the biggest fear: Intel being sidelined as the key technology partner. Invezz also named the remaining risk, Intel's role shrinking to a minor technology supplier.

Set the two parties' own words side by side. Intel joined Terafab in April. It pointed to its ability to “design, fabricate, and package ultra-high-performance chips at scale.” Musk's October reply to talk of a TSMC-run plant sounded different. “No, we will build and run the fab. Let there be ZERO doubt about that.” Intel described a company that fabricates chips.

Musk described a plant that his own companies build and run. His words keep TSMC out of the operator's seat. They keep Intel out of it, too. Musk had drawn that line months earlier. In April, he said Terafab would use Intel's 14A process, and that SpaceX would handle high-volume manufacturing. Intel has never publicly detailed the commercial terms.

Terafab could license the process, get manufacturing help from Intel, or both. One trade report described a general understanding that Intel will license its manufacturing technology. Whether Intel also helps ramp up production is unclear. So the Wednesday reassurance protected a place for Intel's process. It did not reveal a manufacturing contract, because none has been described.

That narrows what Intel is actually defending.

A valuable seat that isn't exclusive

A process seat can sound modest. Investors did not treat it that way. Intel joined Terafab on April 7, and its stock more than doubled that month. TheStreet reports it was the stock's best month on record. The reason lies in Intel's own history. According to TheStreet, Terafab was the first major outside commitment for Intel's foundry business. Until then, that business had only ever made chips for Intel itself.

A foundry builds chips that other companies design, and Intel is still proving it can run one. 14A is where that proof matters most. Intel has said early design commitments for 14A are expected from the second half of 2026 into 2027. As Benzinga's analysis put it, Intel needs customers willing to put real products through its process, not merely evaluate it.

Terafab's first phase alone carries 16.8 billion dollars of committed spending. The full plan, at up to 119 billion dollars, is roughly seven times that. No one has disclosed what Terafab pays Intel. The larger value is proof that a big outside customer will build real chips on 14A. That is why talk of a second foundry cost Intel's stock more than five percent in two days.

Keeping that seat depends on how Musk buys chips. For Tesla's AI5 self-driving chip, he did not pick one foundry. Musk said in April that Tesla had taped out AI5 to both Samsung and TSMC. Tape-out is when a design is locked and handed over for manufacturing. Each foundry makes a slightly different version. In Musk's words, “they translate designs to physical form differently.”

Terafab now shows signs of the same pattern. Culpium followed its first report with a claim that Intel's own Terafab work was not making progress. Tan's answer, through Bloomberg, was that Intel would stay involved. CNBC reported that if TSMC joins, Intel's process could be dropped in favor of TSMC's own technology. Musk's sublease line keeps TSMC from running Terafab.

But eeNews Europe noted it still leaves open a separate TSMC operation on the same campus. One reading favors Intel. Musk first raised a deal with Intel in November 2025, after saying his current partners could not meet his chip needs. A GuruFocus report said there might be room for both Intel and TSMC. It did not say how the work would be split. That reading and the AI5 record lead to the same place.

Intel's seat is real, but it is not exclusive. Benzinga framed it this way: 14A may have to compete for Musk's future chip business, rather than simply supply it. So even the process seat is not settled. Intel has to keep winning it.

The customer learning the trade

Then Musk went after the skills themselves. On October 11, he posted a direct pitch to chip engineers. “If you're a world-class chip design or production engineer, consider joining SpaceX or Tesla.” He added: “How cool would it be to iterate on designs and customize production machines at lightning speed?” He quoted research on his goal of ten chip designs a year through Terafab.

The industry norm, that research said, is about one every two years. Musk is aiming for a pace twenty times faster. At a livestream near Tesla's Austin factory, Musk said his companies could eventually beat established chipmakers. He pointed to his companies' skill at designing custom manufacturing equipment as a potential edge. He also admitted he lacks direct experience in chip fabrication.

And he cautioned that getting there could take several years. 24/7 Wall St. read the post as Tesla and SpaceX moving to reduce or eliminate their reliance on outside chip suppliers. That independence has not arrived. The Financial Times reported that SpaceX is seeking 40 billion dollars in loans and bonds to buy Nvidia AI chips. For now, Musk's companies still rely heavily on outside chipmakers.

Put the pieces together. Terafab will be run by Musk's companies, and Musk himself admits he lacks chip fabrication experience. That is why an experienced process partner like Intel matters now. But every chip engineer Musk hires shifts some of that know-how to his side of the table. Intel's leverage looks strongest in the early years. After that, it is less certain.

Musk's own estimate of several years is the closest thing to a timetable. Intel's foundry story is often told through Terafab alone. 24/7 Wall St. called the project the most visible outside endorsement of Intel's foundry turnaround. On October 6, in the same week the TSMC talks hit Intel's shares, Intel expanded a different partnership.

Intel and chip equipment maker Applied Materials will work together on transistors and the wiring between them. They will also work on advanced packaging, including Intel's Foveros technology for stacking chips in 3D. Tan said the goal is to move new technology from the lab into high-volume U.S. production faster. Packaging may be the faster route to outside revenue.

An Insider Monkey writer argued that process customers take years to win, but packaging deals can come much faster. The same writer cited Chief Financial Officer Dave Zinsner. Zinsner has said packaging can reach 40 percent gross margins, with some deals worth billions of dollars a year. Industry reports have linked Intel to packaging deals with major cloud customers, but none has been publicly confirmed.

Intel's April pitch to Terafab also named packaging, alongside design and manufacturing. So Terafab is one test of Intel's foundry, not the only one. Musk's in-house push puts the most pressure on 14A, the process. Packaging is where Intel is building proof that doesn't depend on who runs Terafab.

Where Intel stands

Back to the promise that calmed Intel investors. Musk's companies will build and run Terafab, and Intel says it is staying. The evidence leans toward Intel being a process partner in a plant Musk runs, not the maker of Musk's chips. That seat's value depends on 14A staying the process while Musk's own team learns the trade. The signal to watch is Intel's own disclosure.

Intel expects early 14A design commitments from the second half of 2026 into 2027. If Intel names Terafab and says it will help ramp production, this reading is wrong. Intel would then hold a manufacturing role, not just a process seat. If Terafab's chips move to TSMC technology instead, Intel loses its most visible 14A anchor. Musk's companies would gain a second foundry to bargain against Intel.

Informational only, not investment advice. Figures and quotes come from the linked reports.