by Clarence Oxford
Los Angeles CA (SPX) Jan 30, 2026
Penn Engineers have proposed a solar-powered orbital information middle structure that might scale to satisfy rising demand for synthetic intelligence computing with out drawing electrical energy from terrestrial grids. The idea makes use of versatile, tether-based buildings in orbit to host hundreds of computing nodes for AI inference, counting on established house tether know-how fairly than large inflexible platforms or huge constellations of impartial satellites.
The design resembles a leafy plant, with a number of stems holding computing {hardware} and branching, leaf-like photo voltaic panels. Every stem is successfully an extended tether populated with similar nodes that carry laptop chips, solar energy programs and cooling {hardware}, forming a modular chain that may be prolonged by including extra nodes.
Tethers in orbit expertise competing forces from Earth’s gravity and the centrifugal impact of orbital movement, which naturally pull them taut and align them vertically with one finish towards Earth and the opposite towards house. By distributing computing nodes alongside these tethers, the system can keep a steady orientation whereas supporting many interconnected modules in a single construction.
The structure is optimized for passive orientation fairly than energetic pointing programs. Photo voltaic panels are barely angled, and the light however steady strain of daylight acts like wind on a climate vane, serving to preserve the panels and computing {hardware} oriented accurately with out counting on motors or thrusters.
In response to the researchers, a single tethered construction may prolong for a number of and even tens of kilometers in orbit. Simulations point out that such a system may host hundreds of computing nodes and help as much as 20 megawatts of computing energy, akin to a medium-sized terrestrial information middle used for AI inference.
Information processed by these orbital information facilities could be transmitted to and from Earth utilizing laser-based optical hyperlinks, a know-how already employed in satellite tv for pc communications. Whereas the latency and throughput necessities for AI coaching make full coaching in orbit impractical, the group notes that future development in AI utilization will largely come from working already-trained fashions, a task that matches the proposed system.
The researchers place their method as a center floor between unscalable satellite tv for pc constellations and impractically giant inflexible buildings. Constellations of many small satellites would require tens of millions of impartial spacecraft to match giant terrestrial information facilities, whereas huge assembled platforms exceed present manufacturing and deployment capabilities.
In contrast, the tether-based structure makes use of a long time of analysis and in-space testing of tethers. Using repeated, modular nodes permits incremental scaling, just like including beads to a necklace, with out basically altering the structural idea as capability will increase.
The group additionally investigated how micrometeoroid and orbital particles impacts would have an effect on such a big orbital system. Utilizing laptop simulations, they examined the cumulative results of many impacts as an alternative of specializing in remoted collisions with particular person parts.
Outcomes counsel that the tethered construction is of course resilient to those impacts. A strike might trigger a short wobble or rotation, however the disturbance travels alongside the tether and progressively dissipates, a habits the researchers evaluate to the way in which movement dies down in a wind chime after it’s disturbed.
In a variety of simulated situations, the system deviated from its optimum orientation by only some levels, remaining inside acceptable limits for solar energy assortment and steady operations. The design additionally consists of a number of tethers supporting every node, in order that if one tether is severed by an influence, the node and the bigger construction can proceed functioning.
Managing warmth in house poses a separate problem, as a result of orbital programs can solely reject warmth by radiating it away. The design incorporates radiators to shed waste warmth from sustained computing hundreds, and the researchers plan to refine these radiators to make them lighter and extra sturdy.
The subsequent step is to maneuver past simulations and develop a small-scale prototype with a restricted variety of nodes to validate the tether-based orientation, energy and thermal ideas. The group emphasizes that the dominant development in AI utilization is coming from repeated inference fairly than from coaching new fashions, and so they argue that offloading this inference to orbit may scale back the environmental burden of information facilities on Earth.
The work, carried out on the College of Pennsylvania College of Engineering and Utilized Science, highlights how current house applied sciences might be tailored to help rising AI workloads. By inserting modular, solar-powered information facilities in orbit, the researchers purpose to create a path for scaling AI computing whereas easing calls for on terrestrial electrical energy and water sources.
Analysis Report:Tether-Based Architecture for Solar-Powered Orbital Data Center
Associated Hyperlinks
University of Pennsylvania School of Engineering and Applied Science
All About Solar Energy at SolarDaily.com
