Sophia Space to test orbital software on Kepler’s compute cluster

Canada-based Kepler Communications, which operates the largest compute cluster currently in orbit, announced a partnership on Monday with Sophia Space. Under the agreement, Sophia Space will test its software on Kepler’s active constellation. Kepler’s orbital infrastructure, launched in January, features about 40 Nvidia Orin edge processors onboard 10 operational satellites. Edge processing refers to performing data processing directly where the data is collected, rather than sending it to a distant server. The company now has 18 customers using its network.

The partnership serves as a critical test for Sophia Space, which is developing passively cooled space computers designed to avoid heavy and expensive active-cooling systems. Sophia Space plans to upload its operating system to Kepler’s satellites to test and configure the software across six GPUs on two spacecraft. This orbital test is intended to de-risk Sophia Space’s technology ahead of its own planned satellite launch in late 2027. The initial focus of such orbital systems is edge processing for space-based sensors, such as synthetic aperture radar—an imaging technology used by satellites to map the Earth’s surface. The US military is already a key customer for processing this radar data.

While competitors like SpaceX, Blue Origin, Starcloud, and Aetherflux are targeting large-scale orbital data centers, experts expect that we won’t see these systems until the 2030s. Instead, Kepler and Sophia Space are focusing on distributed inference workloads. Kepler CEO Mina Mitry explained the efficiency of their approach: “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.” Meanwhile, terrestrial constraints are making space-based alternatives more appealing. Sophia Space CEO Rob DeMillo pointed out that last week, a Wisconsin city adopted a ban on data center construction, signaling growing regulatory hurdles for ground-based infrastructure in the US.

Why it matters

The partnership serves as a de-risking exercise for orbital computing, testing software on existing infrastructure to prove the utility of space-based data processing before larger-scale deployments.