Wireva

Google taps Blue Origin rocket to launch AI computing satellites in 2029

Google has chosen Jeff Bezos's Blue Origin to launch its first heavy-lift flight for Project Suncatcher, which aims to run artificial intelligence computing in orbit.

Bezos and Rob Meyerson (fifth from left) giving NASA Deputy Administrator Lori Garver (fourth from left) a tour of Blue Origin's crew capsule in 2011
Google taps Blue Origin rocket to launch AI computing satellites in 2029
commons.wikimedia.org · Wikimedia — licence per file · rights

Google has selected Blue Origin, the space company founded by Jeff Bezos, to launch a set of computing satellites designed to run artificial intelligence workloads in orbit. The launches are scheduled to take place on Blue Origin's New Glenn rocket in the first quarter of 2029, marking the first heavy-lift flight for Google's Project Suncatcher.

Each satellite will carry Google's in-house AI chips and will be roughly equivalent in computing capacity to one conventional server rack on the ground. Operating in orbit offers a key advantage: the spacecraft can receive near-continuous sunlight, giving them a steady source of solar power. The concept reflects a broader search for new ways to meet the enormous energy, land and cooling demands that artificial intelligence is placing on Earth-based data centers.

Google is not alone in exploring the idea. SpaceX and the startup Starcloud are also investigating orbital data centers, signaling that several major players in the space and technology sectors see potential in moving computing power beyond the atmosphere. For Blue Origin, the agreement secures a significant future mission for New Glenn at a time when the company has been working through delays and recovering from a explosion at its launchpad in May.

The project is still at an early stage. A first prototype was launched on a SpaceX rocket in October and has reportedly been operating normally after communication with it was established. Significant technical questions remain, however. Engineers must determine whether the chips can withstand the stresses of launch, radiation in space and extreme temperature swings. Cooling is another challenge, since the absence of Earth-style airflow means heat must be managed in other ways.

Experts have also pointed to high costs, demanding engineering requirements and limited production capacity as obstacles that could shape how far the concept can scale. Whether orbital computing becomes a practical complement to ground-based facilities will depend on how these problems are resolved over the coming years.

Same event, other desks

Story file →