The production facility will focus on two main types of chips: energy-efficient edge-inference processors for Tesla’s Full Self-Driving systems, Cybercab, and Robotaxi; as well as high-power, radiation-hardened chips designed for SpaceX satellites and xAI’s orbital data centers. Terafab’s ambitious goal is to achieve an annual output of 1 terawatt (TW) of compute capacity, representing a substantial increase—about 50 times the current global output of AI chips.
The facility's output will be specialized, with approximately twenty percent allocated to AI architectures for autonomous driving systems, a proposed robotaxi fleet, and humanoid robots. The remaining eighty percent is designated for a radiation-hardened processor for deployment within satellite constellations, aligning with objectives for orbital data centers.
“So I think we may have to do a Tesla terafab. It’s like giga but way bigger. I can’t see any other way to get to the volume of chips that we’re looking for. So I think we’re probably going to have to build a gigantic chip fab.
Elon Musk’s lieutenants have reached out to chip industry suppliers for his envisioned Terafab, early steps in an audacious and likely arduous attempt to break into the production of cutting-edge chips. Ed Ludlow has more. ——– More on Bloomberg Television and Markets Like this video?
The news on AI5 comes as Tesla has its own ambitious plans to fabricate its own chips at its upcoming Terafab facility. Analysts and experts claim the move to create its own “fab” is highly ambitious and likely a massive engineering challenge.
That specific shrinking process requires ASML's most advanced, heavily backlogged EUV machinery. GaN operates under a different set of physical rules. It can handle significantly higher voltages, temperatures, and frequencies than silicon. Because of these inherent material advantages, highly efficient GaN chiplets can be manufactured on slightly more mature nodes using widely available Deep Ultraviolet (DUV) lithography equipment. By leaning heavily into GaN architectures, TERAFAB can side
The increase is raising new concerns ... Musk announced Terafab, a proposed $20-25 billion semiconductor fabrication plant near Giga Texas, also known as the Gigafactory, in eastern Travis County....
Teams working for Elon Musk have reached out to major chip industry suppliers including Applied Materials, Tokyo Electron, and Lam Research, as well as chip manufacturing partner Samsung Electronics, as part of Musk’s Terafab project, a proposed AI chip fabrication complex linked to SpaceX and Tesla.
When asked about Terafab, TSMC said on Thursday it would not underestimate competitors, but added there are “no shortcuts” in the industry, as it takes two to three years to build a new fabrication plant.
While expanding its own assembly capacity, TSMC will keep working with outsourced semiconductor assembly and testing (OSAT) firms to meet client demand, Wei added. Addressing Tesla and SpaceX CEO Elon Musk's TeraFab project, Wei said there is "no shortcut in the pure-play foundry business."
CPUs Fueled by Musk's TeraFab tie-in, Intel's market cap hits highest level in 25 years — tops $300 billion on CPU, AI, and foundry momentum · Tech Industry Why TSMC grew four times faster than its foundry rivals in 2025
The Terafab project — a joint initiative of SpaceX, Tesla, xAI, and the newly joined Intel — is becoming a key element of this strategy. The initiative is not just about expanding the production base, but also about developing a new industrial ...
Elon Musk is moving forward with his ambitious Terafab project, as his team has reportedly begun contacting global semiconductor equipment suppliers to support the planned AI chip manufacturing complex.
TAIPEI, April 17 (Reuters) - Tesla is seeking semiconductor engineers in Taiwan for its Terafab artificial intelligence chip complex, according to job postings on its website.
Intel entra in Terafab, il mega-fab da 25 miliardi di Tesla, SpaceX e xAI ad Austin. È la risposta strutturale alla crisi della supply chain aperta da Hormuz.
TERAFAB introduces a new building block: orbital compute nodes. ... Instead of centralized data centers, compute becomes spatially distributed, forming a planetary-scale processing layer.
Intel has announced it is joining the Terafab project alongside SpaceX, xAI, and Tesla, marking a significant collaboration aimed at advancing large-scale chip manufacturing for artificial intelligence and robotics applications.