Google Tests Space-Based AI Data Centers with Project Suncatcher

Google is exploring the possibility of building artificial intelligence (AI) data centers in space as part of its 'Project Suncatcher' initiative. The project aims to utilize the resources available in outer space, such as near-constant sunlight, to generate up to eight times more solar power than on Earth.

Last year, Google announced Project Suncatcher, a long-term research endeavor that seeks to determine whether space can host scalable machine learning infrastructure. To test this concept, Google launched several of its Tensor Processing Units (TPUs) into space aboard SpaceX's Transporter-18 Mission from California on Thursday.

The goal is to see how these units handle space flight and the extremes of space, as well as maintain cool temperatures without overheating. However, experts have raised concerns about speed and space debris. Florida Tech Spaceport Education Center Director Don Platt questioned whether space-based connections hundreds of miles away can deliver enough speed and raised concerns about falling space debris.

Google plans to send two satellites into space in 2027 to test maintaining a high-bandwidth connection across short distances. The technology already exists, but most state-of-the-art systems are optimized for low bandwidth over large distances, whereas Google's lasers need to operate at very high bandwidth over extremely short distances.