AI is about to take a giant leap to conquer the final frontier: The Nvidia Jetson Orin NX chip, one of the most popular AI and edge computing GPUs, will soon blast into space on SpaceX’s Transporter 11 mission.
AI has taken the world by storm in recent years, but the technology’s deployment in orbit has been slow. Outside Earth’s atmosphere, ionizing particles and cosmic rays bombard satellites, threatening their electronics. For computers to survive in space, they must be made from durable materials and designed to withstand high doses of radiation. But adapting computers for space takes years, so satellite manufacturers often have to make do with processors that are severely outdated.
Cosmic Shield Corporation (CSC), a Georgia-based spinout from Sweden’s Chalmers University of Technology, thinks it has a solution to this problem: It has developed a specialized nanocomposite shielding metamaterial, a type of polymer studded with nanoparticles (the exact composition is a trade secret) that blocks charged particles in their tracks.
CSC has previously tested the lightweight material in particle accelerators on Earth and during an eight-month experiment on the International Space Station, but the SpaceX ride-share mission Transporter 11, which could launch as soon as this Friday (August 16), will mark the first time the shield will protect an actual AI computer on a space mission.
Related: Artificial intelligence could help us search for life on Mars and other alien worlds
“This will be the fastest AI computer ever launched into space,” CSC co-founder and CEO Yanni Barghouty told Space.com. “The purpose of this mission is solely to demonstrate that AI-enabled Nvidia GPUs can operate properly in orbit with minimal errors.”
The GPU will be launched aboard a CubeSat built by Aethero, a San Francisco company that makes high-performance, space-grade computers. The GPU’s sole task during the four-month orbital mission will be to perform mathematical calculations, the results of which will be transmitted back to Earth and carefully checked.
“One of the main effects of space radiation is errors,” Barghouty said. “AI-enabled GPUs are particularly susceptible to this because they have many transistors that perform calculations in parallel. When one transistor is affected, it often affects many others as well.”
Barghouti hopes that if the test is successful, it will truly mark the satellite industry’s entry into the AI era. CSC is already seeing strong demand for its shields from existing satellite manufacturers and Earth observation providers who want to analyze and process imagery directly onboard their satellites. The arrival of AI in space is highly anticipated as it will enable a variety of new types of missions that require advanced image analysis and visual navigation, such as in-orbit servicing and manufacturing, and active removal and avoidance of space debris.
“All of these companies are interested in taking advantage of the latest AI-enabled hardware,” Barghouti said. “On Earth, AI is having a huge impact on the efficiency of electronic systems, and the space industry wants to take advantage of that. For many years, Moore’s Law only applied on Earth, but our technology allows it to be applied in orbit as well.”
(Moore’s Law describes the observation by American engineer Gordon Moore that the computing power of integrated circuits doubles approximately every two years).
The shielding would also extend the life of computers in space, extending the lifespan of missions and saving costs, Barghouti said. The same material could also be used in space suits and radiation protection for future space stations and lunar habitats.
“As a company, we’re working on a number of different areas,” Barghouti says, “and if we’re successful here, it will pave the way for other applications.”
