 |
| Intel is now a part of Elon Musk's Terafab initiative |
A few days ago, Elon Musk announced what is arguably his most ambitious endeavour ever: Terafab. Tesla, SpaceX and xAI want to enter the lucrative chip manufacturing business, presumably to wean off TSMC and Samsung Foundry for its chip needs. However, it will be impossible to pull off without support from existing chipmakers, and Elon may have roped in a key player: Intel.
Intel says it will produce 1 TW of chips per year – a horrible way to denote capacity. It doesn't explicitly state how many wafers that translates to, nor does it state which technology it will employ. However, it was previously revealed that Terafab aimed to reach 160,000 wafers per month as its initial target. As for the node, it will be equivalent to 2 nm. If Intel's production lines are used, it will likely employ the 14A node.
The Masterplan: From 2nm Chips to a Galactic Civilization
Musk officially unveiled the Terafab project on March 21, 2026, during a public event in Austin, Texas. The facility is designed as a vertically integrated megaplex that will combine chip design, logic fabrication, memory production and advanced packaging all under one roof.
The financial scale is staggering, with projected investment between $20 billion and $25 billion. The plan calls for two dedicated fabs: one producing edge AI chips for Tesla's autonomous vehicles and Optimus humanoid robots, the other manufacturing radiation-hardened chips for SpaceX's orbital AI data centers. "Either we build Terafab, or we have no chips," Musk said during the launch event, explaining that current global chip capacity only covers a small fraction of his companies' projected future demand. The long-term vision is to process up to 1 million wafer starts per month, equivalent to roughly 70 percent of TSMC's entire current global output from a single facility.
Intel's Unexpected Rescue Mission
That's where Intel comes in. On April 7, the struggling chip giant officially announced it would join Terafab alongside Tesla, SpaceX and xAI. Intel's official post on X confirmed the news:
"Our ability to design, fabricate and package ultra-high-performance chips at scale will help accelerate Terafab's aim to produce 1 TW/year of compute to power future advances in AI and robotics," Intel wrote. Intel CEO Lip-Bu Tan called the venture "a step change in how silicon logic, memory and packaging will get built in the future".
For Intel, this partnership is a lifeline. The company's foundry division reported an operating loss of $10.32 billion in 2025, and this deal provides a desperately needed anchor customer. Intel stock jumped as much as 4.2% on the announcement. The collaboration will likely leverage Intel's 18A process node, which entered production in late 2025, as well as its EMIB and Foveros advanced packaging technologies.
The Catch: Why This Could Still Go Wrong
Despite the bold headlines, massive obstacles remain. Musk has zero direct experience in semiconductor manufacturing, and industry analysts have been openly skeptical. Bernstein analyst Stacy Rasgon went so far as to say building a competitive chip plant "might actually be harder than sending a rocket to Mars".
The most immediate bottleneck is equipment. Terafab will need next-generation High-NA EUV lithography machines from ASML – the Dutch company that holds a virtual monopoly on the advanced tools required for 2nm-class production. As a brand-new customer, Musk could face waiting times of two years or more just to secure the necessary machinery. Industry reports indicate ASML may ship as few as four High-NA EUV units in all of 2026.
Even with Intel's help, whether Terafab will develop its infrastructure from scratch or retrofit Intel's existing facilities remains unclear. Given Musk's complete lack of field experience, it will probably be the latter, at least in its initial phases. And even if in-house development is achieved, other supply chain bottlenecks, such as EUV/high-NA EUV machines from ASML, will be another pressing issue.
The Long Game
Until Terafab becomes reality, Elon will continue to rely on TSMC and Samsung Foundry for chip needs. Tesla already has multi-billion-dollar deals in place with Samsung for its AI5 and AI6 chips, with mass production scheduled for 2026 and beyond.
Musk, characteristically, is thinking bigger than earthbound manufacturing. He envisions Terafab as the foundation for humanity to become a "galactic civilization," with 80 percent of the facility's output destined for space-based AI infrastructure rather than ground applications.
Whether that vision becomes reality or joins the long list of overhyped Musk promises remains to be seen. What's certain is that for the foreseeable future, Terafab needs Intel every bit as much as Intel needs Terafab.
Sources: Intel official announcement via X, Reuters, Bloomberg, Notebookcheck, Wikipedia, ESM China, Eweek