The era of commercial data centers in space is officially underway. While the idea of moving massive servers beyond Earth’s atmosphere has generated significant buzz, the reality of orbital computing is taking a more distributed shape today. Leading this charge is Canada’s Kepler Communications, which recently launched the largest space-based computing cluster to date.
This development marks a major shift from traditional terrestrial data centers to specialized in-space processing. With a growing need for faster data analysis and increasing restrictions on building infrastructure on Earth, orbital computing is quickly transitioning from a futuristic concept into a functional market.
Kepler’s Record-Breaking Space Network
In January 2026, Kepler Communications successfully deployed a network that fundamentally changes how data is managed in space. The company’s compute cluster currently operates at an altitude of 500 kilometers. It consists of 10 active satellites that are interconnected by high-speed laser communication links.
Instead of relying on massive server farms, this cluster uses approximately 40 Nvidia Orin edge processors designed for high-performance, low-power applications. The optical inter-satellite links provide a sustained capacity of 2.5 GB/s, creating a mesh backbone that bypasses congested radio frequency channels.
Kepler has already secured 18 customers. The company focuses heavily on edge processing and distributed inference rather than centralized training workloads. By processing data directly in orbit, satellites can transmit actionable intelligence back to Earth instead of raw, bulky sensor feeds.
Mina Mitry, CEO of Kepler Communications, notes that this approach maximizes efficiency. According to Mitry, if a system consumes kilowatts of power but operates at only 10% capacity, it is inefficient. In contrast, Kepler’s processors run at 100% utilization. Kepler views itself as a foundational infrastructure layer providing network services for other satellites, aircraft, and drones.
Tackling the Cooling Challenge with Sophia Space
One of the largest hurdles for space data centers is thermal management. Powerful processors generate substantial heat, and traditional active cooling systems are too heavy, complex, and expensive to launch.
To address this, Kepler announced a strategic partnership on Monday with Sophia Space, a startup developing passively cooled space computers. Sophia Space will upload its proprietary operating system to Kepler’s network and attempt to boot and configure its software across six processors located on two different spacecraft.
While coordinating software across multiple processors is routine on Earth, this marks the first time such an operation will be attempted in orbit. Successfully completing this exercise is a critical milestone for Sophia Space before it launches its own satellite in late 2027. If successful, passively cooled technology could allow much more powerful hardware to operate in space without heavy cooling infrastructure.
Terrestrial Limits and Defense Applications
The push toward space is being accelerated by challenges on Earth. Terrestrial data centers face increasing scrutiny over their massive energy consumption.
Rob DeMillo, CEO of Sophia Space, highlighted a recent ban on new data center construction in a Wisconsin city, noting that federal lawmakers are also pushing for similar limits. DeMillo observed that with fewer places to build data centers domestically, the industry is entering a strange new phase, making space-based alternatives highly attractive.
Beyond commercial uses, the military represents a major driver for this technology. The U.S. military is a key customer, utilizing these networks to develop next-generation missile defense systems. These defense applications require satellites that can detect and track threats with minimal latency. Kepler has already proven its capabilities in this arena by conducting a successful space-to-air laser link demonstration for the U.S. government.
The Future of the Market
The momentum behind space infrastructure is accelerating rapidly. On January 11, 2026, Axiom Space also advanced the field by launching two free-flying data center nodes, continuing experiments with its AxDCU-1 prototype on the space station.
Industry analysts forecast that the market could reach between $30 billion and $40 billion in revenue by 2035. Technology giants are preparing for this shift. Nvidia plans to ship its dedicated Vera Rubin modules with fault-tolerant firmware by late 2027, and HPE has already proven that standard x86 racks can survive space radiation.
However, the industry still faces economic disagreements. While analysts point to falling launch prices, AWS executives argue that the economics of space data centers remain prohibitive until launch costs drop significantly further.
Experts anticipate that massive, large-scale orbital data centers—like those envisioned by SpaceX, Blue Origin, and startups such as Starcloud—will not become operational until the 2030s. For now, the focus remains firmly on distributed edge networks. As companies continue to launch hardware and update software in orbit, orbital computing has firmly established itself as a practical solution.
