Calculators › GPU TCO & $/GPU-hour

GPU TCO & $/GPU-hour

Compute all-in cost per active GPU-hour from complete-system capex, installed-to-active allocation, depreciation life, TDP, PUE, power price, productive utilization, and opex — then compare against the 2026 GPU rental price ladder.

ScenarioNVIDIA GB200 NVL72 · 56 active / 72 installed GPUs · 0.13 MW IT · $5M programguide defaults (asof 2026-07)Edit the shared scenario →

GPU TCO & cost-per-GPU-hour

Server cost$GPUs / serverGPUsDepreciation lifeyr2–3 yr economic · 5–6 yr bookGPU TDPWServer overhead× GPU TDPNon-GPU server load (CPU/mem/NIC/fans) as a multiple of GPU TDPPUEratio≥1 by definitionPower price$/kWhOwned-use utilization%Annual productive hours used to recover ownership costInstalled / active allocation×Installed GPUs ÷ active GPUs; allocates stranded rack slots to productive unitsOpex%/yrRental comparator$/GPU-hrH100 public list median ~$3.99 (Jul 2026); replace with the actual quote

$4.14 / active GPU-hr

Annual cost / active GPU$25,357

· Depreciation$18,917

· Energy$1,900

· Opex$4,540

· Installed-to-active allocation1.285714× across all three cost lines

Breakeven utilization vs rental15%

Ownership beats the rental rate only above the breakeven utilization computed from your own inputs — below it a debt-financed cluster bleeds cash → Ch 1.8.

These are transparent screening estimates, not final designs, financial advice, or project approvals. Replace every default with current project data and have the responsible project authorities validate the result. You can save, share a permalink, or export to CSV; see the full calculator suite.