Intel announced on Tuesday that it has joined the Terafab semiconductor project alongside Tesla, SpaceX, and xAI — formalising a partnership that Elon Musk first floated five months ago when he told shareholders it was “probably worth having discussions with Intel.”
The chipmaker said its ability to “design, fabricate, and package ultra-high-performance chips at scale” would help accelerate Terafab’s goal of producing one terawatt of compute per year to power advances in artificial intelligence and robotics.
“It was fun hosting @elonmusk at Intel this past weekend,” the company wrote on X.
As of press time, Tesla shares were trading 2.6% lower at $343.50.
From ‘Maybe’ to Partnership
Musk first raised the idea of working with Intel at Tesla‘s annual shareholder meeting on November 6, 2025. “You know, maybe we’ll do something with Intel,” Musk told shareholders. “We haven’t signed any deal, but it’s probably worth having discussions with Intel.”
Intel’s stock jumped 4% in after-hours trading that day.
At the time, Musk was laying out the scale of Tesla‘s chip problem.
Even extrapolating the best-case scenario for production from existing suppliers — TSMC and Samsung — the output would not be enough to meet the combined needs of Tesla‘s Full Self-Driving system, Cybercab robotaxi fleet, and Optimus humanoid robot programme, he said.
“I can’t see any other way to get to the volume of chips that we’re looking for,” Musk said. “So I think we’re probably going to have to build a gigantic chip fab. It’s got to be done.”
That vision became Terafab four months later.
What Intel Brings
Intel is one of only three companies in the world — alongside TSMC and Samsung — with the capability to fabricate chips at advanced process nodes. It also possesses industry-leading packaging technologies, including Foveros 3D stacking and EMIB, which are critical for high-performance AI processors.
The company has been seeking external customers for its foundry business after years of declining revenue and an $18.76 billion loss in 2024.
The US government took a 9.9% stake in Intel last year, and the chipmaker has since secured investments from SoftBank ($2 billion) and Nvidia ($5 billion).
Before Tuesday’s announcement, Intel was already handling packaging for Tesla‘s Dojo 3 chips at its Arizona facility.
Barclays had recently noted in its Terafab analysis that Intel fits Musk’s preference for a US-based supply chain and “might have available capacity” — making it a natural partner alongside Samsung and TSMC.
What Terafab Is
Terafab is a joint venture between Tesla, SpaceX, and xAI — the AI company that SpaceX acquired in an all-stock deal in February.
The project was announced on March 21, at a livestreamed event from the defunct Seaholm Power Plant in Austin, Texas. Musk called it “the most epic chip building exercise in history by far.”
The facility is designed to consolidate every stage of semiconductor production under one roof: chip design, lithography, fabrication, memory production, advanced packaging, and testing.
A prototype “Advanced Technology Fabrication” facility will be built at the North Campus of Giga Texas, with the full-scale Terafab at a location yet to be determined.
At full capacity, Tesla claims the output would match roughly 70% of TSMC’s entire current global production.
Terafab targets 2-nanometer process technology with an initial capacity of 100,000 wafer starts per month, scaling to one million.
The project will produce two types of chips: an edge-inference processor — the AI5, which Tesla claims delivers a 50x improvement over the current AI4 — for FSD, Optimus robots, and Cybercab, and a radiation-hardened D3 variant for SpaceX’s orbital AI satellites. Musk said 80% of compute output would be directed toward space applications.
The estimated cost is $20–25 billion, which Tesla‘s CFO confirmed is not yet incorporated into the company’s 2026 capital expenditure plan, which already exceeds $20 billion.
Why It Matters for Tesla
The Optimus robot programme is the largest single driver of chip demand. Morgan Stanley analyst Andrew Percoco estimated that Giga Texas alone could require 20 million chips per year for a capacity of 10 million humanoid robots — approximately six times Tesla‘s current chip demand across its entire automotive business.
Musk has said that all current global chip fabs combined produce roughly 2% of what his companies will eventually need. “We either build the Terafab or we don’t have the chips, and we need the chips, so we build the Terafab,” Musk said at the March event.
Small-batch production of the AI5 chip is expected in late 2026, with volume production in 2027 — though Tesla had already delayed the AI5 to mid-2027 before the Terafab announcement, and the AI6 chip has been delayed roughly six months due to Samsung’s 2nm production slipping.
Analyst Reaction
Wedbush’s Daniel Ives, one of Tesla‘s most vocal bulls, reiterated an Outperform rating and a $600 price target — the highest on Wall Street — following the Terafab announcement.
He called the project “the first step to ultimately what will be Tesla and SpaceX combining forces in a merger likely in 2027,” escalating a thesis Wedbush has been building since the SpaceX-xAI all-stock merger in February.
Morgan Stanley’s Percoco called Terafab a “Herculean task” and estimated the full cost could run $35–45 billion. He projected that initial chip output would not occur until mid-2028 at the earliest under an aggressive build-out scenario.
Bernstein’s Stacy Rasgon was more sceptical, stating that “a true Terafab feels like a stretch” and projecting that the full one-terawatt ambition could require $5–13 trillion in capital.
Barclays’ Dan Levy warned that the project could require spending “many multiples” above his own $50 billion bull-case estimate, calling the capital outlook “likely well more than an order of magnitude higher” than Tesla has communicated.
Levy reiterated an Equalweight rating and $360 price target. He noted that Barclays forecasts Tesla’s 2026 free cash flow at negative $3 billion before any Terafab-related spending.
Precedents
The scepticism is informed by Tesla‘s own track record on capital-intensive manufacturing pivots.
At Battery Day in September 2020, Musk projected three terawatt-hours of annual in-house cell production by 2030.
Five and a half years later, Tesla has reached roughly 2% of that target. The 4680 battery programme took years longer than promised, and Tesla‘s own top battery supplier publicly said Musk “doesn’t know how to make battery cells.”
Tesla‘s Dojo supercomputer project, announced in 2021, was disbanded in August 2025 after Musk called it “an evolutionary dead end.”
It was revived in January 2026 as an orbital compute concept tied to SpaceX — the same infrastructure Terafab is now designed to serve.
Nvidia CEO Jensen Huang publicly cautioned Musk against underestimating the challenge, saying that matching TSMC’s semiconductor capabilities is “virtually impossible.”
Tesla is scheduled to report first-quarter earnings on April 22, where capex guidance and Terafab funding details are expected to be addressed.
