An ambitious partnership aims to place the first solar-powered public cloud data center into Earth’s orbit, tackling the growing energy and land constraints facing the artificial intelligence industry. AI infrastructure provider Crusoe has joined with orbital computing firm Starcloud to launch a satellite-based computing platform, with the initial launch planned for late 2026 and GPU-powered AI workload capacity expected to be operational by early 2027. This venture represents a significant leap from terrestrial data centers to the furthest possible edge: space itself.
The collaboration seeks to resolve a fundamental bottleneck for the scaling of AI technology—access to abundant and clean energy. As AI models become more complex, their demand for electricity has strained terrestrial power grids and increased the environmental footprint of computing. By moving infrastructure beyond the atmosphere, the companies plan to harness uninterrupted solar power and operate independently of Earth’s resources. The project extends Crusoe’s established business model of co-locating data centers with stranded or novel energy sources, such as repurposing flared natural gas at oil fields, into the new frontier of orbital mechanics. If successful, the initiative could redefine how data centers are built and powered for the next generation of cloud computing.
An Energy-First Approach to Computing
Crusoe’s core mission is to align the future of computing with the future of the climate. On Earth, the company has deployed over 120 modular data centers that run on natural gas that would otherwise be flared, or burned off as a waste byproduct at remote oil and gas sites. This “Digital Flare Mitigation” technology converts the gas into electricity on-site, powering high-performance computing while significantly reducing methane and CO2-equivalent emissions compared to routine flaring. The company’s patented systems achieve a combustion efficiency of up to 99.9%, preventing waste and creating economic value from an underutilized energy source.
The partnership with Starcloud applies this energy-first philosophy to space, where solar power is a consistent and virtually unlimited resource. “We believe that space will ultimately matter to the future of computing because it enables new solutions to a key scaling constraint for AI infrastructure, which is sourcing abundant, consistent and clean energy,” said Cully Cavness, Crusoe’s Co-founder, President and Chief Operating Officer. By moving off-planet, the project aims to bypass the terrestrial challenges of grid capacity, land acquisition, and water usage for cooling, which increasingly limit the growth of hyperscale data centers on the ground.
Orbital Platform and Technical Design
The venture relies on Starcloud’s specialized satellite platform, designed to function as a self-contained, high-performance data center in low-Earth orbit. The satellite integrates high-efficiency solar panels for power generation, powerful processors for computation, and an innovative onboard cooling system tailored for the vacuum of space. The first commercial satellite, Starcloud-2, is scheduled for a 2026 launch and will host Crusoe’s cloud computing module, enabling customers to run demanding AI workloads including both model training and inference.
Harnessing Solar Power and Radiative Cooling
Operating in orbit provides access to uninterrupted sunlight, eliminating the intermittency of solar power on Earth’s surface and the need for massive battery storage. Starcloud’s long-term vision includes a 5-gigawatt orbital data center with vast solar and cooling panels. A key technical innovation is the cooling system. While space is often perceived as cold, the vacuum is an excellent insulator, making heat dissipation a major challenge. Starcloud’s design uses large, deployable radiators that reject waste heat directly into space through thermal radiation, a process that is highly effective without the need for water or conventional chillers used in terrestrial facilities.
High-Performance Compute in Orbit
The initial demonstration satellite, Starcloud-1, is expected to offer processing power over 100 times more powerful than any previous space-based computer. By deploying state-of-the-art GPUs, the platform will be capable of handling the intensive computations required for modern AI applications. Data will be transmitted to and from the orbital data center using high-throughput laser terminals that connect to satellite internet constellations, enabling low-latency connections as fast as 50 milliseconds. This capability is critical for serving both in-space customers, such as Earth observation satellites that generate immense volumes of data, and terrestrial users.
Solving Terrestrial Infrastructure Bottlenecks
The explosive growth of AI has created an unprecedented demand for data center capacity, leading to significant environmental and logistical challenges on Earth. These facilities consume enormous amounts of electricity, strain local power grids, and require vast quantities of water for their cooling systems. The Starcloud and Crusoe partnership is a direct response to these mounting pressures, which are increasingly seen as a primary obstacle to continued AI development.
By operating in space, the orbital data center model completely avoids many of these constraints. The platform requires no physical land, no connection to a power grid, and no water for cooling. This approach not only reduces the direct environmental impact but also allows for rapid deployment without the lengthy permitting and construction timelines associated with building new hyperscale facilities on the ground. As AI continues to scale, this off-world solution could offer a path to sustainable growth for the industry.
Future Vision and Remaining Hurdles
The initial 2026 launch is the first step toward a much larger vision of a distributed, space-based cloud network. The companies plan to scale their operations by launching multiple satellites, eventually creating an interconnected web of orbital data centers. This network could unlock new possibilities for global research, discovery, and innovation, serving a wide range of applications from climate modeling to financial analysis.
However, the project faces considerable technical and economic challenges. The high cost of launching payloads into orbit remains a significant factor, although it is steadily decreasing. Furthermore, equipment in space must be hardened to withstand cosmic radiation and protected from the risk of collisions with orbital debris. Ensuring reliable and secure high-speed data transmission between the orbiting data centers and ground stations is another critical area of focus. The long-term success of this pioneering venture will depend on overcoming these hurdles to prove that the economics and performance of orbital computing can compete with its Earth-bound counterparts.