Space Premium
-Extra CAPEX above GPU hardware cost — the price of orbit: launch + arrays + radiators + bus structure.
Space Datacenter Feasibility
What if you moved a datacenter into orbit? This tool lets you explore how much extra it would cost — the "Space Premium" — by adjusting assumptions about rocket launches, solar panels, cooling, and orbital altitude. Drag the sliders and watch the numbers update in real time.
The Space Premium captures every space-specific cost: launch, solar arrays to power the chips, radiators to cool them, and satellite bus structure — everything beyond the GPU hardware itself.
All figures update live. The Space Premium (highlighted) is the headline — it measures how much more expensive a space-based GPU fleet is compared to buying the same GPU hardware for a ground datacenter.
Space Premium
-Extra CAPEX above GPU hardware cost — the price of orbit: launch + arrays + radiators + bus structure.
Fleet CAPEX
-Total cost of all satellites: GPU hardware + every space-specific cost.
GPU Hardware Cost
-What the equivalent compute would cost to buy outright for a terrestrial datacenter. Space Premium = Fleet CAPEX − this number.
Effective Launch Cost
-$/kg delivered to the target altitude, after applying the Hohmann transfer mass multiplier for orbits above the baseline altitude.
Weighted Satellite Mass
-Beta-angle-weighted average mass per satellite. Includes compute module, radiators, solar arrays, and bus structure.
Satellites Needed
-Fleet size required to deliver the target compute continuously, scaled up to compensate for eclipse downtime.
GPU hardware cost vs. the Space Premium (everything else). A thin GPU bar and tall Space Premium bar means orbit is adding a large multiple on top of the hardware price.
How the Space Premium scales with the datacenter size you want to replace. The dot marks your current selection. In this model the premium scales linearly — no economies of scale from building a larger space fleet.
Waterfall showing the four space-specific cost categories that add up to the Space Premium. Hover each bar to see the dollar value. Launch typically dominates at high altitudes; solar arrays are the main mass driver.
How launch spend breaks down by which component is being lifted. Lighter solar arrays (higher W/kg) or radiators (lower kg/m²) directly cut the launch bill proportionally.
Illustrative 2D perspective sketch of the fleet at the selected altitude. Satellites are grouped by beta-angle family (color-coded). This is for visual intuition only — orbits are not propagated mechanically.
This simplified interactive model follows the notebook's sunlight-only fleet sizing path. It computes total fleet CAPEX and isolates the Space Premium as the extra cost beyond GPU hardware that an equivalent terrestrial datacenter would also buy.
Space Premium = Fleet CAPEX − Fleet GPU Hardware Cost. The GPU hardware cost is what you'd pay to buy the same compute on the ground. Everything else — launch, solar arrays, radiators, and bus structure — is the Space Premium.
How a satellite is sized: Each satellite carries one compute module (the selected chip rack). The model then calculates how much solar array mass is needed to power it, how much radiator area is needed to reject the heat, and how much bus structure results from those masses. The total satellite mass drives launch cost.
Sunlight-only simplification: Rather than modeling batteries for eclipse periods, the model sizes the fleet assuming compute runs only in sunlight and uses more satellites to hit the target continuous MW. Sunlight fraction depends on orbital altitude and the beta-angle distribution of the constellation.
Launch cost model: A baseline $/kg to the preset reference altitude (e.g., 550 km), then a Hohmann-transfer mass multiplier for reaching higher orbits. The multiplier grows steeply above ~2,000 km. Isp and propulsion dry-mass fraction control the steepness of that curve.
What this model does not include: operations costs, on-orbit maintenance, inter-satellite links, latency penalties, regulatory/spectrum costs, battery mass for eclipse computing, multi-orbit shell architecture, or revenue potential. It is a first-order capital cost feasibility estimate.
Tooltip footnotes and anchor cards point here so users can trace major mass, solar, launch, and assumption anchors back to notebook-linked references.