The single most common reason people stop mining at home is not the electricity bill. It is the noise. A stock air-cooled flagship miner is roughly as loud as standing next to a busy road, and it does that all night. Here is what air, hydro and immersion cooling actually sound like, what each one costs, and how to choose.
- ✓A stock air-cooled flagship miner runs at roughly 75–85 dB. That is not a living-room sound, or a garage-next-to-a-bedroom sound.
- ✓Decibels are not a normal scale. Every 10 dB up sounds about twice as loud to a human ear.
- ✓Hydro (water) cooling typically lands around 45–50 dB, because the loud part — the fans — is mostly gone.
- ✓Immersion cooling can get to roughly 35–40 dB, quieter than a fridge, but it is the biggest build.
- ✓Quieter machines also tend to be more efficient, so cooling is not only a noise decision.
Why decibels confuse everybody
Decibels do not add up the way normal numbers do. The scale is logarithmic, which is a mathematical way of saying it grows much faster than it looks.
Two rules are enough to understand every noise number you will ever see on a miner:
- 1Every 10 dB roughly doubles how loud it seems. So 75 dB does not sound a bit louder than 65 dB. It sounds about twice as loud.
- 2Distance helps fast, then stops helping. Doubling your distance from the source drops it about 6 dB. Going from one metre to two metres helps a lot. Going from ten metres to twenty helps much less in a real room, because sound bounces off walls.
An 85 dB air-cooled miner is not “a bit louder” than a 45 dB hydro machine. It is four doublings louder — roughly sixteen times as loud to your ears. That is the difference between a conversation and a lawnmower.
What each cooling type actually sounds like
The reason air cooling is loud is simple, and it is not the chips. It is the fans. To move enough air to carry away three to five kilowatts of heat, the fans have to spin extremely fast. Fast fans move air, and moving air makes noise.
Both quieter methods work by getting rid of the fans, or most of them, and moving the heat with liquid instead. Liquid carries heat far better than air, so it can move the same heat much more slowly, and slow means quiet.
How the three methods work
Efficiency: the part people miss
Cooling is usually framed as a noise decision. It is also an efficiency decision, and that shows up on the electricity bill every month.
The measure to watch is J/TH — joules per terahash. It is simply how much electricity the machine spends to do a fixed amount of work. Lower is better. Here are real air and hydro models from the live catalogue.
| Miner | Hash rate | Power | Efficiency | Cooling | Break-even power price |
|---|---|---|---|---|---|
 Antminer S19k Pro - 110 TH/s |
110 TH/s | 2,530 W | 23.00 J/TH | Air | $0.0728 |
 Antminer S21 - 188 TH/s |
188 TH/s | 3,290 W | 17.50 J/TH | Air | $0.0957 |
 Antminer S21 Pro - 245 TH/s |
245 TH/s | 3,675 W | 15.00 J/TH | Air | $0.1116 |
 Antminer S23 - 318 TH/s |
318 TH/s | 3,498 W | 11.00 J/TH | Air | $0.1522 |
 Antminer S21+ Hyd - 395 TH/s |
395 TH/s | 5,925 W | 15.00 J/TH | Hydro | $0.1116 |
 Antminer S23 Hyd - 580 TH/s |
580 TH/s | 5,510 W | 9.50 J/TH | Hydro | $0.1762 |
What each option costs to build
These are realistic planning ranges for a small home or small-business setup, not exact quotes. Prices vary a lot by country and by how much of the work you do yourself.
| Approach | Typical noise | What you have to build | Rough extra cost | Best for |
|---|---|---|---|---|
| Stock air | 75–85 dB | Nothing. Plug it in. | None | A shed, a barn, or a site far from anyone |
| Air + enclosure | 55–65 dB | An insulated box with ducting and airflow | Low to moderate | One machine, one determined owner |
| Fan swap + tuning | 65–72 dB | Quieter fans, lower power mode | Low | Taking the edge off, not solving it |
| Hydro | 45–50 dB | Cold plates, pump, radiator or dry cooler, plumbing | Moderate to high | Serious home setups and small farms |
| Immersion | 35–40 dB | Sealed tank, dielectric fluid, heat exchanger, pump | High | The quietest result, if you have the space |
Choosing, in one page
- 1Measure before you build. A phone sound-meter app is not laboratory grade, but it will tell you what your room already sits at and what your neighbours might hear.
- 2Check your local noise limits. Many places set a night-time limit at the property boundary, often around 40–45 dB. That number, not your tolerance, is the one that matters legally.
- 3Decide what the heat is for. If you can use the heat, hydro becomes far more attractive, because warm water is easy to move and easy to use.
- 4Price the whole build, not the kit. Pump, radiator, fluid, fittings, a leak plan and your own time. The tank is rarely the expensive part.
- 5Then ask the honest question. If the quiet build costs more than the machine and still annoys someone, the machine is in the wrong building.
That last point is not a sales line, it is arithmetic. A serious immersion build can cost more than the miner it silences. At 20 sites averaging $0.0480 per kWh, the OneMiners hosting network exists precisely because industrial cooling at scale is cheaper per machine than domestic cooling for one. Both hydro and immersion models are hosted as standard.
Quiet-by-design models in the live catalogue
Hydro and immersion models run as standard across the hosting network, in buildings designed for the heat — so the loudest thing in your house stays the kettle.
Compare hosting sitesBrowse Bitcoin minersQuestions people keep asking
Where these numbers come from
- Miner power, hash rate and price figures: live OneMiners catalogue, 27 August 2026.
- Break-even calculations use a Bitcoin price of $79,264 and a network hash rate of 887.8 EH/s (mempool.space, 27 August 2026), subsidy only, before pool fees.
- Noise ranges are typical published figures for stock air-cooled, hydro and immersion setups; measured results vary with room, mounting and distance.
- Everyday sound levels are standard reference values used in occupational noise guidance.
This article is for information and education only. It is not financial advice. Mining results change with the Bitcoin price, network difficulty, your electricity rate and your hardware. Numbers shown are examples from public data on the date of writing, not a promise of any result. Always do your own research.
