Terrestrial Compute Total Cost of Ownership
What is the amortized TCO per kW_IT/year for terrestrial compute?
Answer
Terrestrial TCO ranges from 6,036 $/kW_IT/year (optimistic, 2026) to 13,240 $/kW_IT/year (conservative, 2030). The central estimate is approximately 8,806 $/kW_IT/year in 2026, remaining nearly flat through 2040 at 8,758 $/kW_IT/year. The remarkable stability of terrestrial TCO reflects the dominance of GPU hardware cost (which is time-invariant in the model) over the more volatile energy cost component.
Terrestrial TCO is composed of amortized GPU cost, amortized infrastructure cost, annual energy cost, and non-energy opex. GPU cost is the dominant component at 47-77% of total TCO across scenarios.
Inputs
| Input | Question | Answer | Page |
|---|---|---|---|
| terrestrial-infrastructure-cost | What is the all-in infrastructure cost per kW_IT? | $8,000-$20,000/kW_IT (central: $12,500) | link |
| terrestrial-energy-cost | What is the effective electricity cost for AI data centers? | $0.045-$0.110/kWh (scenario/year dependent) | link |
| terrestrial-pue | What is the PUE for modern liquid-cooled AI data centers? | 1.03-1.20 (central: 1.10) | link |
| gpu-cost-per-kw | What is the baseline GPU cost per kW_IT? | $25,000-$40,000/kW_IT (central: $32,500) | link |
| gpu-useful-life | What is the expected useful life of AI accelerator hardware? | 4-6 years (central: 5 years) | link |
Analysis
TCO Formula
Terrestrial TCO = gpu_amortized + infra_amortized + energy_annual + non_energy_opex
Where:
- gpu_amortized = gpu_cost_per_kw / gpu_useful_life
- infra_amortized = terrestrial_infra_cost / 15 years (facility life)
- energy_annual = energy_cost x 8,760 hours x PUE
- non_energy_opex = $750/kW_IT/year (staff, maintenance, property tax, insurance)
Component Breakdown
GPU cost (amortized) is the largest single component:
| Scenario | GPU Cost | Useful Life | Amortized |
|---|---|---|---|
| Optimistic | $25,000 | 6 years | 4,167 |
| Central | $32,500 | 5 years | 6,500 |
| Conservative | $40,000 | 4 years | 10,000 |
GPU cost is time-invariant (same across all years) and dominates TCO in all scenarios. In the central case, GPU cost is 6,500 $/kW_IT/year, representing 74% of total TCO. This dominance is the fundamental reason why terrestrial TCO is so stable over time -- the largest component does not change.
Infrastructure cost (amortized) over a 15-year facility life is modest:
| Scenario | Infra Cost | Amortized (15yr) |
|---|---|---|
| Optimistic | $8,000 | 533 |
| Central | $12,500 | 833 |
| Conservative | $20,000 | 1,333 |
Infrastructure amortization is a small fraction of TCO (6-10%) because the 15-year facility life spreads the cost over a long period. This is a structural advantage of terrestrial deployment: facilities outlast the GPU hardware they house, allowing multiple generations of compute to amortize the same building.
Energy cost is the only time-varying component:
| Year | Optimistic | Central | Conservative |
|---|---|---|---|
| 2026 | 586 | 723 | 946 |
| 2030 | 496 | 771 | 1,156 |
| 2035 | 451 | 723 | 1,104 |
| 2040 | 406 | 675 | 999 |
Energy cost is computed as electricity price x 8,760 hours/year x PUE. In the central case, energy is 723 $/kW_IT/year (2026), representing only 8% of total TCO. This is consistent with the widely cited industry observation that energy is "only 5-15% of total data center cost" (per SemiAnalysis and Catalyst). Even in the conservative scenario, energy peaks at 1,156 $/kW_IT/year (2030) -- still only 9% of total conservative TCO.
The low share of energy in total TCO is the single most important finding for the orbital comparison. Orbital compute's primary advantage is eliminating energy costs through free solar power, but if energy is only 8% of terrestrial TCO, eliminating it provides only an 8% savings -- far too small to offset the additional costs of launch, platform manufacturing, and orbital operations.
Non-energy opex of $750/kW_IT/year covers staffing, maintenance, property tax, and insurance. This is a modest, stable component representing about 8-12% of total TCO across scenarios.
Total TCO Trajectories
| Year | Optimistic | Central | Conservative |
|---|---|---|---|
| 2026 | 6,036 | 8,806 | 13,029 |
| 2028 | 5,991 | 8,830 | 13,135 |
| 2030 | 5,946 | 8,854 | 13,240 |
| 2035 | 5,901 | 8,806 | 13,187 |
| 2040 | 5,856 | 8,758 | 13,082 |
The trajectories are remarkably flat. The central estimate varies by only ~$50/kW_IT/year across the entire 2026-2040 period (less than 1% variation), driven by small fluctuations in energy cost. The optimistic trajectory declines modestly as energy costs fall; the conservative trajectory shows a slight hump as energy costs peak around 2030 before declining.
Why Terrestrial TCO Is Hard to Reduce
The dominance of GPU hardware cost creates a structural floor for terrestrial TCO. Even with zero energy costs and zero infrastructure costs, terrestrial TCO could not fall below:
- Optimistic: 4,167 + $750 = ~$4,917/kW_IT/year
- Central: 6,500 + $750 = ~$7,250/kW_IT/year
- Conservative: 10,000 + $750 = ~$10,750/kW_IT/year
Actual terrestrial TCO is only 15-25% above this theoretical floor, indicating that infrastructure and energy are secondary cost drivers. This makes terrestrial compute surprisingly cost-efficient relative to its theoretical minimum.
Sensitivity Analysis
The three most impactful parameters are:
GPU cost per kW_IT (highest sensitivity): A 20% change in GPU cost shifts central TCO by ~$1,300/kW_IT/year (15% of total). This dwarfs all other inputs.
GPU useful life (moderate sensitivity): Extending useful life from 5 to 6 years reduces central TCO by ~$1,083/kW_IT/year (12%). This is why hyperscalers care deeply about depreciation schedules.
Energy cost (low sensitivity): Doubling energy cost from $0.075 to $0.15/kWh increases central TCO by only ~$723/kW_IT/year (8%). Energy is a small lever.
Infrastructure cost and PUE have even lower sensitivity due to the long facility amortization period and already-low PUE values of modern liquid-cooled facilities.