Layer 1 · How the heat leaves
Rear-door heat exchangers
Operators pushing an existing air-cooled hall further without replumbing it.
The shape of it
- Capital cost
- Moderate; retrofits into existing halls
- Maturity
- proven
- Density supported
- ~30-50 kW per rack
- PUE
- 1.2-1.4
Verified 2026-09-09.
The catch
A bridge, not a destination. It buys you one density step and then you are having the direct-to-chip conversation anyway.
What it actually decides
The best retrofit-in-place option. Buys time, not headroom.
What this is powering
A GB200 NVL72 rack draws 120–132 kW and a GB300 pushes 135–200 kW, against a 2026 average rack of about 27 kW. The silicon decision and the power decision are the same decision, made eighteen months apart.
Layer 2 — Silicon · GB200 NVL72 · The 150 W inference option
Others in how the heat leaves
- Air cooling — Up to ~30 kW per rack
- Direct-to-chip liquid cooling — 50-200+ kW per rack
- Single-phase immersion — 100+ kW per rack equivalent
- Two-phase immersion — Highest available
- Free and evaporative cooling — Depends on the paired system
Want this whole layer as one table?
The power and cooling options on these pages as a single CSV — lead time, capex, carbon, maturity, density supported, PUE, the catch and what each one decides — ready to sort, filter or load into a notebook.
Get the energy table — free for research and evaluation
Four short lines and the download starts at once (CC-BY-NC-4.0, built from these pages the moment you ask). What you type is used only to answer the request and is never sold. Prefer Hugging Face? Request it there.
Using it inside a product, a client report or anything you sell? That needs a commercial licence.