The race to put data centers in space is heating up, but for all the bold announcements and billion-dollar ambitions, the reality on the ground — or rather, in orbit — has been more modest. For all the hype about data centers in space, there just aren't very many GPUs up there. As that starts to change, the near-term business of orbital compute is beginning to take shape.
And the company currently leading that charge isn't SpaceX or Blue Origin. It's a Canadian startup called Kepler Communications.
Meet the Biggest Compute Cluster in Orbit
The largest compute cluster currently in orbit was launched by Canada's Kepler Communications in January, and boasts about 40 Nvidia Orin edge processors onboard 10 operational satellites, all linked together by laser communications links.
That may sound modest compared to the GPU farms powering AI on the ground, but in space terms, it's a milestone. Kepler now has 18 customers signed on to use its orbital infrastructure — and it just announced its newest one.
Sophia Space Joins the Constellation
Kepler's newest customer is Sophia Space, a startup that will test the software for its unique orbital computer onboard Kepler's constellation. The partnership is significant because it represents the first attempt to do something utterly routine on Earth — deploying and configuring software across multiple GPUs — in the challenging environment of outer space.
In the new partnership, Sophia will upload its proprietary operating system to one of Kepler's satellites and attempt to launch and configure it across six GPUs on two spacecraft. That sort of activity is table stakes in a terrestrial data center, and this is the first time it will be attempted in orbit. Successfully pulling this off will be a critical proof point for Sophia ahead of its own satellite launch planned for late 2027.
Solving the Overheating Problem
One of the biggest technical barriers to orbital data centers isn't connectivity or processing power — it's heat. Sophia is developing passively-cooled space computers that could solve one of the key challenges for large-scale data centers in orbit: keeping powerful processors from overheating without having to build and launch heavy, expensive active-cooling systems.
On Earth, data centers rely on massive cooling infrastructure — water systems, fans, and climate-controlled facilities. In space, that's simply not feasible at scale. If Sophia's passive cooling approach works, it could remove one of the most significant obstacles standing between today's modest orbital clusters and tomorrow's full-scale space-based data centers.
Infrastructure, Not a Data Center
Kepler's CEO Mina Mitry is careful about how she frames the company's ambitions. Kepler doesn't see itself as a data center company, but as infrastructure for applications in space. It wants to be a layer that provides network services for other satellites in space, or drones and aircraft in the sky below.
The company is already demonstrating real-world demand. Satellite companies are now planning future assets around this model, pointing to the benefits of offloading processing for more power-hungry sensors like synthetic aperture radar. The U.S. military is a key customer for that kind of work as it develops a new missile defense system predicated on satellites detecting and tracking threats. Kepler has already demonstrated a space-to-air laser link in a demo for the U.S. government.
Inference Over Training — A Different Vision
Kepler's approach to orbital compute stands apart from the large-scale ambitions of SpaceX, Blue Origin, and well-funded startups like Starcloud and Aetherflux. Rather than building massive GPU clusters optimized for AI training, Kepler is betting on distributed inference running AI models at the edge, where the data is actually collected.
Mitry told TechCrunch: "Because we have the belief it's more inference than training, we want more distributed GPUs that do inference, rather than one superpower GPU that has the training workload capacity. If this thing consumes kilowatts of power and you're only running at 10% of the time, then that's not super helpful. In our case, our GPUs are running 100% of the time."
The Bigger Picture: Earth's Limits Are Pushing Compute Skyward
Experts expect that large-scale data centers like those envisioned by SpaceX or Blue Origin won't materialize until the 2030s. The first step will be processing data collected in orbit to improve the capabilities of space-based sensors used by private companies and government agencies.
But geopolitical and regulatory pressures on Earth may accelerate the timeline. Sophia CEO Rob DeMillo pointed out that Wisconsin adopted a ban on data center construction last week — something some lawmakers in Congress are also pushing. Anything that limits data centers on Earth makes the space-based alternative more attractive.
As DeMillo put it: "There's no more data centers in this country. It's gonna get weird from here."
