Deployment31 July 20262 min read

Air, hydro or immersion: matching cooling to your site

The cooling route you pick decides which machines you can even consider, how loud your site is, and how much you spend before the first miner is switched on.

ASICMiner Distribution
On this page4 sections
  1. Air
  2. Hydro
  3. Immersion
  4. Deciding in the right order

Cooling is usually treated as an afterthought, decided once the hardware is ordered. That is backwards. Cooling determines which models are available to you, how dense you can build, what your infrastructure costs, and whether you can operate at all in a residential location or a warm climate.

Air

The default and the simplest. Machines arrive ready to run on a shelf or in a rack, and the infrastructure is airflow: intake, exhaust, filtration and enough separation between hot and cold sides. Capital cost is the lowest of the three, spares are everywhere and any technician can work on it.

The tradeoffs are noise and heat density. Air cooled miners are loud enough to rule out most locations near housing, and in warm climates the machines derate exactly when ambient temperature peaks. Dust is a maintenance item you never fully solve.

Hydro

Liquid runs through cold plates inside the machine and heat leaves through a dry cooler or a water source. Hydro units run far quieter, hold stable chip temperatures in climates where air cooled hardware struggles, and allow much higher density per rack.

In exchange you build a loop: manifolds, pumps, plumbing, leak detection and usually three phase power. That is real construction, and it only pays back at a certain scale. Hydro is also less forgiving of poor workmanship than air.

Immersion

Machines sit in tanks of dielectric fluid. Operation is close to silent, temperatures are very stable, dust disappears entirely and hardware life often improves. Density is the highest of the three.

The cost is the tank infrastructure, the fluid itself, handling procedures and a maintenance routine that is genuinely different from anything a dry site does. Pulling a machine for service is messier, and not every model is built for immersion.

Deciding in the right order

  • Location first. Noise limits and ambient climate rule options in or out before any spec sheet is opened.
  • Infrastructure budget second. Hydro and immersion move spend from the machines to the building. Compare total cost to first hash, not machine price.
  • Target density third. If your power capacity is large relative to your floor area, liquid routes stop being optional.
  • Then the hardware. Only at that point does the model shortlist make sense.
Worth knowing. Some models exist in air, hydro and immersion variants with different part numbers and different efficiency figures. Buying the wrong variant for your site is an expensive and surprisingly common mistake. The cooling route also decides the building, which we cover in container, warehouse or retrofit. It also decides how loud the site is, and on that the routes are worlds apart, see how loud a mining site is.
Next articleThe first 72 hours, from pallet to stable hashrateDeployment · 3 min read
Next step

Run this against your own numbers

Tell us your power situation, your site and the scale you are aiming at. We come back with a hardware shortlist and a firm quote.

Added to inquiryView