Google Aims to Bring AI to Space: SpaceX Could Become the Decisive Partner
Google has confirmed that it is in talks with SpaceX and other aerospace companies to evaluate future launch missions related to Project Suncatcher, an experimental initiative focused on orbital data centers. The news, first reported by the Wall Street Journal and later confirmed by Alphabet, highlights how the tech sector is concretely considering space as a possible extension of cloud infrastructure dedicated to artificial intelligence.
Project Suncatcher represents Google's effort to build a distributed AI platform in low Earth orbit. The idea is to connect solar-powered satellites equipped with Tensor Processing Units (TPUs), the hardware accelerators developed in-house for AI workloads. The ultimate goal would be to create a sort of "orbital cloud" capable of processing data directly in space, reducing some energy and infrastructural limitations of terrestrial data centers.
According to reports, Google is already collaborating with Planet Labs to design and build the first experimental satellites. A debut for an orbital prototype is expected around 2027, although the project is still in an early stage of research and technological validation.
The potential involvement of SpaceX seems particularly significant, considering that Elon Musk's company is currently the global leader in commercial space launches thanks to its Falcon rockets and the Starlink network. Any agreement would also unite two companies that, at least on the AI front, are gradually entering into competition. In recent months, Musk has announced the merger between SpaceX and xAI, outlining a strategy that also aims for AI processing in orbit via a constellation of dedicated satellites.
Both Musk and Google CEO Sundar Pichai have described space data centers as a possible natural evolution of cloud infrastructure in the next decade.
However, numerous technical questions remain. Several industry experts have pointed out that performing AI inference on a large scale in space presents significant complexities. The GPUs and AI accelerators operating in orbit would be constantly exposed to cosmic radiation, with possible impacts on the reliability of calculations. Cooling also represents a major challenge: in the near-vacuum of space, heat cannot be dissipated through convection and must be radiated away slowly.
In addition to technological challenges, there are environmental and operational concerns. The prospect of putting a large number of satellites dedicated to AI processing into orbit could further increase congestion in low Earth orbits, which are already crowded with mega-constellations for telecommunications. According to some analysts, this scenario could complicate space traffic management and increase risks for future commercial and governmental missions.