The design resembles a leafy plant, with multiple stems holding computing hardware and branching, leaf-like solar panels. Each stem is effectively a long tether populated with identical nodes that carry computer chips, solar power systems and cooling hardware, forming a modular chain that can be extended by adding more nodes.
Tethers in orbit experience competing forces from Earth's gravity and the centrifugal effect of orbital motion, which naturally pull them taut and align them vertically with one end toward Earth and the other toward space. By distributing computing nodes along these tethers, the system can maintain a stable orientation while supporting many interconnected modules in a single structure.
The architecture is optimized for passive orientation rather than active pointing systems. Solar panels are slightly angled, and the gentle but continuous pressure of sunlight acts like wind on a weather vane, helping keep the panels and computing hardware oriented correctly without relying on motors or thrusters.
According to the researchers, a single tethered structure could extend for several or even tens of kilometers in orbit. Simulations indicate that such a system could host thousands of computing nodes and support up to 20 megawatts of computing power, comparable to a medium-sized terrestrial data center used for AI inference.
Data processed by these orbital data centers would be transmitted to and from Earth using laser-based optical links, a technology already employed in satellite communications. While the latency and throughput requirements for AI training make full training in orbit impractical, the team notes that future growth in AI usage will largely come from running already-trained models, a role that fits the proposed system.
The researchers position their approach as a middle ground between unscalable satellite constellations and impractically large rigid structures. Constellations of many small satellites would require millions of independent spacecraft to match large terrestrial data centers, while enormous assembled platforms exceed current manufacturing and deployment capabilities.
By contrast, the tether-based architecture makes use of decades of research and in-space testing of tethers. The use of repeated, modular nodes allows incremental scaling, similar to adding beads to a necklace, without fundamentally changing the structural concept as capacity increases.
The team also investigated how micrometeoroid and orbital debris impacts would affect such a large orbital system. Using computer simulations, they examined the cumulative effects of many impacts instead of focusing on isolated collisions with individual components.
Results suggest that the tethered structure is naturally resilient to these impacts. A strike may cause a brief wobble or rotation, but the disturbance travels along the tether and gradually dissipates, a behavior the researchers compare to the way motion dies down in a wind chime after it is disturbed.
In a wide range of simulated scenarios, the system deviated from its optimal orientation by only a few degrees, remaining within acceptable limits for solar power collection and stable operations. The design also includes multiple tethers supporting each node, so that if one tether is severed by an impact, the node and the larger structure can continue functioning.
Managing heat in space poses a separate challenge, because orbital systems can only reject heat by radiating it away. The design incorporates radiators to shed waste heat from sustained computing loads, and the researchers plan to refine these radiators to make them lighter and more durable.
The next step is to move beyond simulations and develop a small-scale prototype with a limited number of nodes to validate the tether-based orientation, power and thermal concepts. The team emphasizes that the dominant growth in AI usage is coming from repeated inference rather than from training new models, and they argue that offloading this inference to orbit could reduce the environmental burden of data centers on Earth.
The work, conducted at the University of Pennsylvania School of Engineering and Applied Science, highlights how existing space technologies could be adapted to support emerging AI workloads. By placing modular, solar-powered data centers in orbit, the researchers aim to create a path for scaling AI computing while easing demands on terrestrial electricity and water resources.
Research Report:Tether-Based Architecture for Solar-Powered Orbital Data Center
Related Links
University of Pennsylvania School of Engineering and Applied Science
All About Solar Energy at SolarDaily.com
| Subscribe Free To Our Daily Newsletters |
| Subscribe Free To Our Daily Newsletters |
