Two-phase · Direct-to-chip · Waterless
ZutaCore savings vs air-cooled data centers
Estimate input-power, cooling, cost, real-estate and PUE gains from moving an air-cooled workload to ZutaCore two-phase direct-to-chip cooling.
Configuration
20
40
Air baseline: OCP IT-40air (40.4 kW/rack). DtC: OCP IT-73dtc-l2l (73.7 kW/rack).
Power Usage Effectiveness
PUE = total facility power ÷ IT power (lower is better). Both bars sit on the same IT load (1.0); the stack above it is power-distribution loss plus cooling overhead.
Power-chain loss
Air cooling overhead
ZutaCore + Munters cooling
−0.00lower PUE — 0% less cooling energy with ZutaCore
Air-baseline IT —
DtC IT —
Total electricity saved —
Cooling-cost savings
0/ year
CPU input-power savings
0kW
Server-fan / conversion overhead removed
GPU input-power savings
0kW
Cold-plate capture on accelerators
Cooling-power savings
0kW
Real-estate saved
0m²
Method. IT input power falls as ITE efficiency rises from 0.723 (air) to 0.792 (DtC) — OCP L4 figures — applied to CPU and GPU loads. Cooling overhead for ZutaCore + Munters is the measured annualized cooling pPUE at the four reference sites (1.02–1.06, no chilled water); other locales are modelled from OCP's 8,760-hour weather data. The air baseline is a modelled chilled-water counterfactual. PUE holds the IT denominator constant across both designs. Indicative figures for planning, not a guarantee.
Reference-site cooling pPUE is measured on the ZutaCore + Munters solution; other locales are modelled from the OCP climate dataset.