MicroBT WhatsMiner Bitcoin Miners

Read the model number before you read anything else

MicroBT encodes the cooling type in the second digit, and it is the single most useful thing to know about this range:

M60 is air-cooled. M63 is hydro-cooled. M66 is immersion-cooled.

That means an M63S+ and an M66S+ are not two performance tiers — they are the same class of machine built for completely different infrastructure. Buying an M66 without a tank, or an M63 without a loop and heat rejection, leaves you with hardware you cannot switch on. Only the M50S++ and the M60 series in this category will run in a conventional air-cooled room.

After the cooling digit, S denotes the higher-performance variant, and trailing + or ++ marks mid-generation efficiency uplifts.

How to choose

Compare on W/TH, which is the same number as J/TH. MicroBT quotes efficiency in W/T and our listing titles carry it, so the comparison is direct: the M60S+ range sits around 16.5–17 W/TH, the M63S and M66S hydro and immersion machines run 15.5–18 W/TH, and the older M50S++ is materially thirstier. Lower is better, and the difference compounds every hour the machine runs.

Do not compare hashrate across cooling types. A 442 TH/s hydro unit looks enormous next to a 206 TH/s air unit, but it is drawing proportionally more power and it requires a plumbed facility. The right comparison is efficiency and total site capacity, not the headline number.

Check voltage and service. These are industrial machines on 200–240V or three-phase depending on model, and the larger hydro units draw serious current. Confirm your electrical service before ordering, not after.

Count the cooling before you shortlist anything

Of the machines in this category right now, four will run in a conventional air-cooled room, six need a water loop with heat rejection behind it, and three need an immersion tank and dielectric fluid. That is the single most expensive mistake available on this page, so it is worth stating as a count rather than a principle: the air-cooled options are the M50S++ and the M60S+ units, and everything else on this shelf is infrastructure-first hardware.

On this shelf, hydro and immersion do not buy you efficiency

This surprises people, and it is specific to what is actually in stock rather than a general claim about cooling. The most efficient air-cooled machine here, the M60S+ at 16.5 W/TH, is more efficient than five of the six hydro units and two of the three immersion units, all of which sit at 17 or 18 W/TH. What the plumbed machines buy you here is density, noise and thermal headroom — far more terahash in the same floor space, and a machine that is not screaming — not a lower power bill per unit of work. If your reason for looking at hydro is efficiency, the numbers on this page do not support it. If your reason is site density, they support it completely.

Break-even electricity rate for the machines here

MicroBT quotes W/T, which is the same number as J/TH, so the store’s usual division applies unchanged: break-even rate = hashprice ÷ (24 × W/TH). Against a hashprice of $32.18/PH/day, derived on 9 August 2026 from the last 144 blocks at a Bitcoin price of $64,959 and difficulty 127.48 T — a number that moves every day:

15.5 W/TH breaks even at 8.7¢/kWh — the M66S++ 320 TH/s, the most efficient machine in the category and an immersion unit, so the tank is not optional.

16.5 W/TH breaks even at 8.1¢ — the M60S+ 206 TH/s air unit and the M63S+ 442 TH/s hydro unit. 17 W/TH breaks even at 7.9¢, which covers most of the M63S+ range and the M66S+. 18 W/TH breaks even at 7.4¢ — the M63S 396 TH/s and the M66S 300 TH/s.

The M50S++ 146T at roughly 23 W/TH breaks even at 5.8¢, which puts it in the same economic bracket as a refurbished S19k Pro from our used shelf. It is the cheap way into MicroBT hardware and it needs genuinely cheap power to make sense. Judge it against used machines, not against the M6x line.

Read the W/T, not just the model name

MicroBT ships the same model in more than one efficiency bin, and this category shows it clearly: there are two M60S+ 206 TH/s listings, one at 16.5 and one at 17 W/TH, and the M63S+ appears at both 16.5 and 17. Half a watt per terahash sounds like nothing and is worth about a quarter of a cent on your break-even rate, which is the kind of margin that decides whether a machine survives two retargets from now. Likewise the four M63S+ listings between 414 and 442 TH/s at 17 W/TH are one machine sorted into hashrate bins — efficiency is off the table between them, so compare those on cost per terahash and on whether the number fits the circuit you have.

Who this suits

Operators building or expanding a facility, and anyone who wants a second manufacturer in the fleet so a single firmware or supply problem cannot take out the whole site. For home-scale machines look at Avalon; for the widest SHA-256 selection see Antminer or all miners.

Read the middle digit before anything else — on MicroBT machines it is the cooling, not the performance. The M60S+ is air-cooled, the M63S 396 TH/s is hydro and needs a water loop and a dry cooler behind it, and the M66S++ 320 TH/s is immersion and needs a tank and dielectric fluid. Pick the cooling your site can actually support first; the terahash figure is the easy part of this decision.

Before you commit a circuit: on this shelf the figure that decides the install is power at the wall, not terahash. The WhatsMiner M50S++ 146T is rated ≈ 3358 W ± 10% and the WhatsMiner M63S+ 442 TH/s at 7,293–7,359 W — both are one machine on one listing, and they are not one electrical job. At 240 V that is roughly 14 A against 30 A, which is a different breaker, a different feed and usually a different room, before the 80% continuous-load rule is applied to either. What a miner needs from your electrical supply works that through, and the ASIC miner buying guide is where the W/TH figures on these listings turn into a break-even electricity rate.