SpaceX Chief Executive Elon Musk recently shared further technical details about the company’s ambitious plans to create an orbital data center constellation. Speaking at an event on March 21 in Austin, Texas, Musk highlighted the collaboration between SpaceX, Tesla, and xAI, all companies he leads, aimed at significantly boosting the production of high-end computer chips for both terrestrial and space uses.
The initiative, known as the Terafab project, aims to produce one terawatt of processors annually. Musk emphasised that this figure is 50 times the total output of existing chip manufacturers, which are necessary for advanced applications like artificial intelligence (AI). He stressed that these processors are the key to deploying a large number of satellites to form an orbital data center.
In late January, SpaceX submitted a proposal to the Federal Communications Commission (FCC) seeking to launch a constellation of up to one million satellites for AI applications. The application lacked details about the satellites’ specifications, but Musk described an “AI Sat Mini,” which would initially operate in this constellation. Each satellite will provide 100 kilowatts of power for onboard AI processors.
Musk shared a scale illustration of the satellite next to SpaceX’s Starship V3, indicating its considerable size. He also addressed concerns about heat rejection from satellites, arguing that SpaceX is well-equipped to manage thermal issues in space.
The Terafab initiative will launch from an “Advanced Technology Fab” in Austin, with a focus on producing a specific chip designed for space. However, Musk did not specify timelines for this project or costs related to the orbital data center constellation. He concluded his presentation with a vision of future satellites being constructed on the moon.
