Elon Musk over the weekend announced plans to build two semiconductor factories in Austin, Texas, that would supply advanced chips for SpaceX and Tesla. Musk’s effort, known as Terafab, aims to manufacture 1 trillion watts of computing power per year, he said during a live stream on his social media platform X. The factories are expected to make two kinds of chips: those designed for Tesla’s cars and humanoid robots; and high-powered ones built for space.

During the live stream, Musk underscored the lack of manufacturing capacity for chips that power artificial intelligence. He said the current global output of AI chips is roughly 20 gigawatts per year, representing about 2% of what his companies need to power the next generation of humanoid robots, reusable rockets and driverless vehicles.

“We’re very grateful to our existing supply chain — to Samsung, [Taiwan Semiconductor Manufacturing Co.], Micron and others — and we would like them to expand as quickly as they can, and we will buy all of their chips,” Musk said. “But there’s a maximum rate at which they’re comfortable expanding, and that rate is much less than we would like.”

“We need the chips, so we’re going to build the Terafab,” he added.

The massive facility is expected to have all the equipment necessary to make and improve the chips under one roof. Musk said it will be able to create lithography masks, build logic and memory chips, conduct testing, packaging and more in a single building. This will create an “incredibly fast recursive loop for improving the chip design,” he said.

As Musk invests in Austin, other U.S. chip capacity is ramping up. TSMC recently finished construction of its second fab in Phoenix, Arizona, with plans to conduct “high-volume” manufacturing in the second half of 2027. The company has also begun construction on its third fab and is developing a fourth one. Intel’s Albany, Ohio, plant is also under construction following a leadership change and multiple project delays.

About 20% of the Terafab facility’s annual output will be for “terrestrial” purposes, including chips for Tesla’s Full Self-Driving and Optimus robots, according to the presentation. The remaining capacity will be for space applications, including orbital data centers.

In late January, SpaceX filed an application with the Federal Communications Commission seeking approval for an “Orbital Data Center System.” The project would create a constellation of a million satellites that harness the sun’s energy to meet growing AI demand, PCMag reported. It would dwarf SpaceX’s existing Starlink infrastructure, which currently has over 9,600 satellites in Earth’s orbit.

“As soon as the cost to orbit drops to a low number, it immediately makes extremely compelling sense to put AI in space,” Musk said during his presentation, noting increased economies of scale and pushback from people over the environmental impacts of data centers.

“Nobody wants the thing in their backyard,” he said.

Terafab is a joint venture between SpaceX, Tesla and xAI.

By: DocMemory
Copyright © 2023 CST, Inc. All Rights Reserved