- Musk’s Terafab aims to overproduce global chip output by fifty times
- Recursive manufacturing promises speed while leaving the technical details unclear
- New physics requirements remain undefined despite strong confidence from Musk
Elon Musk has revealed plans for a massive expansion of computing capacity through a new facility called Terafab in Austin, Texas.
The facility aims to generate one terawatt of computing power annually, eclipsing the current global output of chipmakers, which stands at around 20 gigawatts each year.
Musk stated that Tesla, SpaceX and xAI will collaborate to achieve this scale and produce chips versatile enough for various applications, including lithography masks.
The article continues below
A production system with a focus on speed and iteration
The Terafab initiative relies on a recursive manufacturing process that enables rapid chip production along with iterative redesigns for better performance.
Musk refers to “some very interesting new physics” that underpins this approach and expresses confidence in its viability, though timelines remain unclear.
“We’re going to push the limits of physics in computing and do some wild and crazy things,” Musk said during a recent presentation.
Two primary chip types emerge from this effort: one optimized for inference tasks on Earth.
The former will primarily power humanoid robots expected to reach sales volumes between one billion and ten billion units annually, and the latter is designed for space-based computers in satellites, starting at 100 kilowatts per device and scaling to megawatt levels over time.
Musk expects robots to eventually outnumber humans within a single year at maximum production rates, raising questions about feasibility given resource constraints.
In space, satellites equipped with these chips require enormous launch volumes; Musk’s calculations point to deploying 10 million tons of material annually, including computer hardware, solar panels and support structures.
Such ambitions necessitate a larger Starship variant capable of towing 200 tonnes per hour. flight – yet basic arithmetic indicates a requirement of around 50,000 launches annually, or 135 daily, to reach the terawatt target.
Such activity amplifies vulnerabilities, including helium disruptions already hampering semiconductor manufacturers by 30% due to the US-Israel war in Iran.
Elon Musk did not deliver a procurement roadmap amid these shortcomings, and his track record invites scrutiny; remember Musk’s one million self-driving Tesla taxis? It yielded only 200 test units.
SpaceX also faces delays, with Starship V3 launches postponed even as boosters prepare for action.
However, Musk counters skepticism by recalling how Tesla’s electric vehicles and SpaceX’s reusable rockets overcame early doubts about viability and cost-effectiveness.
“I think it’s important to consider the vastness of the universe and what we can do that’s much bigger than what we’ve done before, as opposed to worrying about some sort of petty quibbles on Earth,” he remarked to an enthusiastic crowd.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds. Be sure to click the Follow button!
And of course you can too follow TechRadar on TikTok for news, reviews, video unboxings, and get regular updates from us on WhatsApp also.
