Is Bitcoin Mining Profitable at Home?

An honest look at home mining economics as of 28 October 2026: what a miner is expected to earn per terahash, what household electricity costs in the US, Britain and Germany, and why most home setups lose money on power alone. Worked examples show the arithmetic, the narrow exceptions (very cheap power, heat you would pay for anyway), and how pooled payouts differ from a solo lottery ticket.

Category: Mining and the Network · Published 2026-10-28

Editorial review complete · technical review pending · sources under review · time-sensitive: revisit periodically

The short answer

For most people paying an ordinary household electricity price, no. At the network difficulty and Bitcoin price of 27 October 2026, an Antminer S21 (200 TH/s for 3,500 W) was expected to earn about $7.21 a day before electricity, and it uses 84 kWh a day to do that.¹ At the average US residential price for August 2026, 18.16 cents per kWh, those 84 kWh cost $15.25.² So the machine is expected to lose about $8 a day, before it has paid back anything of its own purchase price.

That is not a quirk of one machine. At the average household prices published this autumn by the US Energy Information Administration, by Ofgem in Britain and by Germany's federal statistics office, none of the machines in this article covers its electricity on average, and older machines lose far more. The exceptions are real but narrow: electricity that costs you very little, heat you would otherwise pay for with plug-in electric heaters, and the hobbyist who runs a small board as a lottery ticket with a known running cost.

Everything below carries a date, because the two things that set a miner's revenue keep moving. Difficulty is recalculated every 2,016 blocks, roughly every two weeks,³ and the price moves by the minute. For today's figures use the calculator or the hardware pages; the worked examples here are there to show the method.

What a home miner earns per terahash

Bitcoin pays whoever adds the next block to the chain. The first transaction in each block creates new coins for the miner, and the fees paid by the transactions inside the block go to the miner as well. The whitepaper was blunt about what that costs: "it is CPU time and electricity that is expended."⁴ The new-coin part, the subsidy, has been 3.125 BTC since the April 2024 halving; with fees, a block in late October 2026 paid about 3.13 BTC in total.¹

Your expected share of those rewards comes down to one ratio. On average a block takes difficulty × 2³² hashes to find, so a machine's expected time to a block is difficulty × 2³² ÷ its hashrate in hashes per second. At the difficulty of 140.58 T set on 16 October 2026,⁵ one terahash per second would expect a block about once every 19,000 years, or about 0.000000143 of a block a day. Multiply that daily fraction by a reward of about 3.13 BTC and a price of about $80,400 (27 October 2026), and each TH/s is expected to earn about 3.6 US cents a day. Those three figures are worked out by us. Miners call this revenue per terahash, or hashprice.

Applied to real machines, SoloLuck's model pages gave these expected earnings a day on 27 October 2026, before electricity:

• Antminer S23 (305 TH/s, 3,355 W): about $11.00
• Antminer S21 Pro (234 TH/s, 3,510 W): about $8.44
• Antminer S21 (200 TH/s, 3,500 W): about $7.21
• Antminer S19 (95 TH/s, 3,250 W): about $3.42
• Avalon Q (90 TH/s, 1,674 W): about $3.24
• Avalon Mini 3 (37.5 TH/s, 800 W): about $1.35
• Antminer S9 (13.5 TH/s, 1,350 W): about $0.49
• FutureBit Apollo II (about 9 TH/s in Turbo mode, 375 W): about $0.32
• NerdQAxe++ (about 4.8 TH/s, about 76 W): about $0.17

These are long-run averages, the same whether you mine in a pool or alone; only the way the money arrives differs. They also move with the market: a 10% fall in the Bitcoin price takes 10% off every figure, and a 10% rise in difficulty takes off about 9%.

Electricity is the cost that decides it

The running cost is plain arithmetic. Multiply the machine's draw in watts by 24 and divide by 1,000 to get kilowatt-hours a day, then multiply by what you pay per kWh. A 3,500 W S21 uses 84 kWh a day, an 800 W Avalon Mini 3 uses 19.2 and a 76 W NerdQAxe++ about 1.8. Household prices, from the bodies that publish them:

• United States: 18.16 cents per kWh, the average residential price for August 2026 (a preliminary figure), released by the EIA on 23 October 2026. State averages ranged from 12.72 cents in Nevada to 47.18 cents in Hawaii.²
• Great Britain: 26.32p per kWh, the direct-debit unit rate under Ofgem's price cap for 1 October to 31 December 2026, averaged across England, Scotland and Wales, plus a standing charge of 54.83p a day. The government has removed VAT from electricity bills from 1 October 2026 to 31 March 2027, so the rate carries none.⁶ At the European Central Bank's reference rates for 27 October 2026 (1 euro = 1.1150 US dollars = 0.83658 pounds), 26.32p is about 35.1 US cents (worked out by us).⁷
• Germany: 37.97 cents per kWh, the average that households paid in the first half of 2026, including standing charges, network fees, taxes and levies; households using more than 15,000 kWh a year paid 30.55 cents.⁸ At the same ECB rates those are about 42.3 and 34.1 US cents (worked out by us).
• Canada and Australia: prices are set province by province and state by state, and they differ widely within each country, so use the per-kWh rate printed on your own bill.

Two cautions. An average spreads fixed charges over every kWh, so the price of one extra kWh, which is what a miner adds, can be a little lower than the headline; Ofgem's figure is already that extra-kWh rate. And a miner makes anyone a heavy user: a single S21 running all year adds about 30,700 kWh (worked out by us), about twice the threshold of Destatis's top household band.

Efficiency sets your break-even price

Efficiency is quoted in joules per terahash (J/TH), which is the same thing as watts per TH/s. Since every terahash earns the same, efficiency alone decides the electricity price at which a machine stops covering its running cost. A machine rated at E J/TH uses E × 24 ÷ 1,000 kWh a day for each TH/s, against about 3.6 cents of expected revenue, which gives a rule of thumb for late October 2026, worked out by us: the break-even price, in US dollars per kWh, is about 1.5 divided by the machine's J/TH. On the figures above:

• Antminer S23, 11 J/TH: about 13.7 cents
• Antminer S21 Pro, 15 J/TH: about 10.0 cents
• NerdQAxe++, about 15.8 J/TH: about 9.5 cents
• Antminer S21, 17.5 J/TH: about 8.6 cents
• Avalon Q, 18.6 J/TH: about 8.1 cents
• Avalon Mini 3, 21.3 J/TH: about 7.0 cents
• Antminer S19, 34 J/TH: about 4.4 cents
• FutureBit Apollo II, about 42 J/TH: about 3.6 cents
• Antminer S9, about 100 J/TH: about 1.5 cents

Put that list beside the tariffs and the title question answers itself. In August 2026 not one US state had an average residential price below the S21's 8.6 cents, or even the S21 Pro's 10 cents.² Seven states sat below the S23's 13.7 cents: Nevada, Louisiana, North Dakota, Utah, Idaho, Tennessee and Nebraska. Even in the cheapest, Nevada at 12.72 cents, an S23 would be expected to clear only about 76 cents a day ($11.00 − 80.52 kWh × $0.1272, worked out by us), before a cent goes towards the machine itself. In Britain and Germany every machine on the list costs more to run than it is expected to earn; an S23 would lose about $17 a day at the British cap rate and about $23 a day at the German average (worked out by us).

Older machines show why efficiency matters more than raw hashrate. An S19 does a little under half the work of an S21 for almost the same power, so it needs electricity under about 4.4 cents; the much older S9 needs under about 1.5 cents. A machine that came free still costs money every hour it runs.

Four worked examples

Each example uses the expected revenue of 27 October 2026 and the published electricity price named beside it. Revenue a day minus electricity a day gives the expected net; the arithmetic and the currency conversions are ours.

1. An Antminer S21 in an average US home.

• Expected revenue: about $7.21 a day (difficulty 140.58 T, Bitcoin about $80,400, 27 October 2026).
• Electricity: 3,500 W × 24 h = 84 kWh; 84 kWh × $0.1816 (EIA average, August 2026) = $15.25 a day.
• Net: $7.21 − $15.25 = −$8.04 a day, about −$241 over 30 days (worked out by us).

At 10 cents per kWh the same machine nets −$1.19 a day, and at 5 cents it nets +$3.01.¹ Same machine, same bitcoin, a different electricity bill, and the answer swings from a modest gain to a steady loss.

2. An Avalon Q on Britain's price cap.

• Expected revenue: about $3.24 a day, or about £2.43 at the ECB rates of 27 October 2026.
• Electricity: 1,674 W × 24 h = 40.18 kWh; 40.18 kWh × 26.32p (cap rate, October to December 2026) = £10.57 a day.
• Net: £2.43 − £10.57 = −£8.14 a day, about −£244 over 30 days (worked out by us).

Canaan built the Avalon Q for homes: it runs on 110 to 240 volts and is quoted at a noise range of 45 to 65 dB.⁹ None of that touches the arithmetic. At the cap rate it would have to earn more than four times what it is expected to earn just to cover its power.

3. A NerdQAxe++ in a German home.

• Expected revenue: about $0.17 a day, roughly €0.15.
• Electricity: 76 W × 24 h = 1.82 kWh; 1.82 kWh × €0.3797 (Destatis average, first half of 2026) = €0.69 a day.
• Net: €0.15 − €0.69 = −€0.54 a day, about −€197 a year (worked out by us).

That is what a small solo ticket costs to hold. At the 27 October 2026 difficulty, this board's chance of finding a block on any given day is about 1 in 1,500,000, and its expected time to a block is about 4,000 years.

4. The bigger machines on cheap power.

• Antminer S21 at 5 cents: $7.21 − 84 kWh × $0.05 = $7.21 − $4.20 = +$3.01 a day.
• Antminer S23 at 10 cents: $11.00 − 80.52 kWh × $0.10 = $11.00 − $8.05, about +$2.94 a day on the model page's unrounded figures.

Positive at last, but look at what is left out. These figures come before the machine's purchase price, before repairs, and before the slow fall in revenue per terahash covered below. Three dollars a day is about $1,100 a year (worked out by us), and the machine has to pay for itself out of that while the margin shrinks.

When home mining can pay its way

Electricity that costs you very little. The break-even list sets the bar: under about 8.6 cents for an S21 and about 13.7 cents for an S23, as of late October 2026. A cheap overnight rate works on the same per-hour logic, though a miner that runs only in those hours earns only for those hours. Solar surplus is another case: its real cost is what the grid would pay you for it, not your retail price, but panels produce little in winter and nothing at night, so a miner fed only on surplus runs part-time. Whatever the source, use the actual price of the extra kWh.

Heat you would pay for anyway. A miner turns practically all the electricity it draws into heat: all 3.5 kW an S21 Pro draws comes back out as heat, every hour it runs.¹⁰ In a room you would otherwise warm with a plug-in electric heater, that electricity gets spent either way, and the mining revenue becomes a rebate on the heating. Canaan sells the Avalon Mini 3 on exactly that basis, as a space heater that mines.¹¹ At the US average price an 800 W Mini 3 uses 19.2 kWh a day, which costs $3.49. An 800 W electric heater would cost the same, and the miner is expected to hand back about $1.35, so its heat costs about $2.14 a day (worked out by us).

That rebate only beats the most expensive way to heat. The US Department of Energy says today's heat pumps can cut the electricity used for heating by about 65% compared with electric resistance heating.¹² The same heat from a heat pump would take about 6.7 kWh, about $1.22 a day at the US average, which is less than the miner's $2.14 even after its rebate (worked out by us). Gas usually wins too. Under the British cap, gas costs 7.97p per kWh against 26.32p for electricity,⁶ and German households paid 11.58 cents for gas against 37.97 cents for electricity in the first half of 2026.⁸ The heat is also worth something only in the months you would heat. In summer it is a cost, and with air conditioning you pay a second time to move it back out.

Noise, wiring and space. None of this shows up in a per-day figure, but it decides whether a machine can live in a home at all. Bitmain rates the S21 Pro at 76 dBA with its fans at full speed.¹⁰ For scale, normal conversation runs at 60 to 70 dBA; the US National Institute on Deafness and Other Communication Disorders calls sounds at or below 70 dBA generally safe, and warns that long or repeated exposure at or above 85 dBA can cause hearing loss.¹³ An S21 also needs a 220 to 277 volt supply on a single 20 amp input, which in most homes means a dedicated circuit.¹ Home units such as the Avalon Q run from a normal socket and are quieter, but quiet is not the same as cheap to run.

A hobby with a known running cost. Plenty of people run a small board for fun, to learn how mining works and for a long-shot chance at a block. That is a fair reason, as long as the daily loss is counted as spending: a NerdQAxe++ costs about 16 cents a day more than it is expected to earn at the US average (1.82 kWh × $0.1816 − $0.17, worked out by us).

Pooled payouts and the solo lottery ticket

Every figure so far is an expected value. How it reaches you depends on whether you mine in a pool or alone. In pooled mining, many miners' work is combined and the block reward is split according to the work each one contributed; the pool measures that work in shares, proofs of work at a much lower difficulty than a block, so they turn up often.¹⁴ The result is a trickle of small payouts close to the expected figure, minus the pool's fee. A pool smooths the money. It does not change the electricity sum: an S21 in a pool in an average US home still loses about $8 a day, just evenly.

Solo mining pays the same expected value the other way round, as the whole block or nothing. On 27 October 2026 a block was worth about 3.13 BTC, about $251,650 at $80,400 (worked out by us). At that day's difficulty one Antminer S21 had about a 1 in 35,000 chance of finding a block on any given day, an expected time of about 96 years, and about a 1% chance across a whole year (1 − e^(−365/35,000), worked out by us). A Bitaxe Gamma at about 1.2 TH/s was looking at about 16,000 years, a 1 in 5,800,000 chance a day. An expected time is an average, not a countdown. A machine that has run for ten years is no closer to a block than it was on its first day, and nothing promises it will ever find one. The odds calculator shows the figures for your own hashrate.

People do win, which is why the lottery comparison fits. As of 27 October 2026, SoloLuck's Solo Block Record listed 44 solo-mined Bitcoin blocks since 2022, found mostly by small or home miners on public solo pools and each checked on the chain. They include a single NerdQAxe++ (block 945,601, 18 April 2026), a Bitaxe Gamma (block 957,382, 10 July 2026), an Avalon Mini 3 (block 913,593, September 2025) and an Avalon Nano 3S (block 951,771, May 2026).¹⁵ Set that against the size of the field. In the census edition for 19 to 25 October 2026, the 17,086 Bitaxes connected to a single public solo pool, Public Pool, averaged 1.05 TH/s each.¹⁶ All of them together would expect roughly one block a year at this difficulty (worked out by us). For any single owner that is a lottery ticket whose price is the electricity bill. It is not income, and it is not something to build a budget around. If a block does come, its reward cannot be spent until it has matured for 100 blocks.³

Difficulty and the halving wear returns down

Revenue per terahash does not stay put. Difficulty is recalculated every 2,016 blocks, about two weeks: up if blocks came faster than one every ten minutes, down if they came slower.³ Over the long run it rises as more efficient machines are switched on, though not in a straight line. On mempool.space's record it stood at 95.67 T on 22 October 2024, 146.72 T on 16 October 2025 and 140.58 T on 16 October 2026, and the highest in that record was 155.97 T, set on 29 October 2025.⁵ That is up about 47% over two years and down about 4% over the last one (worked out by us). If difficulty rose another 47% with the price unchanged, an S21's $7.21 a day would shrink to about $4.91 (worked out by us).

The halving is the bigger step. The block subsidy halves every 210,000 blocks, roughly every four years. The last halving, at block 840,000 in April 2024, took it to 3.125 BTC, and the next, at block 1,050,000, takes it to 1.5625 BTC.¹⁷ That block is about 78,000 blocks after the 16 October 2026 adjustment, which at ten minutes a block lands around April 2028 (worked out by us; the real date depends on how fast blocks arrive). Unless the price or fees rise to make up for it, revenue per terahash roughly halves overnight, and so does every break-even price in the list above. There is more on how the reward changes as the subsidy shrinks in what happens when all bitcoin is mined.

For a machine you mean to keep for years, judge it on its break-even price today and assume that bar will fall. A machine that only just covers its power this month, like an S21 Pro at 10 cents netting about a cent a day, has no room left for the next difficulty rise.

Running the numbers for your own home

The same steps work for any machine and any tariff:

• Find your marginal price: the unit rate per kWh on your bill, or the off-peak rate if the miner would only run then.
• Find the machine's real draw at the wall and its hashrate, from the maker or the hardware pages.
• Get the expected revenue at today's difficulty and price from the calculator, which shows the odds alongside.
• Subtract: revenue a day − (watts × 24 ÷ 1,000 × price per kWh).
• If the result is positive, set the machine's purchase price against it, remembering that the margin shrinks as difficulty rises and halves at the halving.
• If the heat would replace plug-in electric heaters, count the heating months separately, and compare against a heat pump or gas as well.
• Check again after each difficulty adjustment and after any big move in price.

The answer, as of 28 October 2026: at household prices in the US, Britain and Germany, a home miner is expected to cost more in electricity than it earns. It can pay its way on unusually cheap power, or as a heater that replaces plug-in heaters. Otherwise it is a hobby with a running cost, and solo mining adds a lottery ticket on top. For comparing machines, see how to evaluate mining hardware; for the odds, how solo-mining probability works.

Key takeaways

• At the difficulty and price of 27 October 2026, an Antminer S21 was expected to earn about $7.21 a day before electricity, using 84 kWh to do it.
• At average household prices in the US (18.16 cents, August 2026), Britain (26.32p under the October to December 2026 cap) and Germany (37.97 cents, first half of 2026), none of the machines compared here covers its electricity on average.
• Efficiency sets the break-even price: roughly 1.5 divided by the machine's J/TH, in US dollars per kWh, at late-October 2026 figures.
• Home mining can pay its way on very cheap or surplus power, or as heat in a room you would otherwise warm with plug-in electric heaters, but rarely against a heat pump or gas.
• Pooled and solo mining have the same expected value; solo is a lottery ticket priced in electricity, and an expected time to a block is an average, not a countdown.
• Rising difficulty over time and the next halving, at block 1,050,000 around 2028, keep pushing every machine's break-even price down.

Sources

1. Antminer S21 model page: expected earnings a day at the 27 Oct 2026 difficulty and price (block reward about 3.13 BTC with fees), net at $0.05 and $0.10 per kWh, and power input (220-277 V, single 20 A input) — SoloLuck
2. Electric Power Monthly, Table 5.6.A: average price of electricity by state, August 2026 (preliminary), released 23 Oct 2026; US residential average 18.16 cents per kWh — U.S. Energy Information Administration
3. Bitcoin Developer Guide, Block Chain: difficulty recalculated every 2,016 blocks (ideal 1,209,600 seconds, two weeks); coinbase outputs cannot be spent for at least 100 blocks — Bitcoin Developer Documentation
4. Bitcoin: A Peer-to-Peer Electronic Cash System (the whitepaper), section 6 'Incentive': new coins and fees go to the block's creator; CPU time and electricity are what is spent — bitcoin.org
5. Difficulty adjustment history, three years (JSON): 95.67 T on 22 Oct 2024, 146.72 T on 16 Oct 2025, 155.97 T on 29 Oct 2025, 140.58 T on 16 Oct 2026 at block 971,712 — mempool.space
6. Changes to energy price cap between 1 October and 31 December 2026 (published 26 Aug 2026): electricity 26.32p per kWh and 54.83p a day by direct debit, no VAT on electricity 1 Oct 2026 to 31 Mar 2027; gas 7.97p per kWh — Ofgem
7. Euro foreign exchange reference rates for 27 Oct 2026: 1 euro = 1.1150 US dollars = 0.83658 pounds sterling — European Central Bank
8. Press release No. 345 of 30 Sep 2026 (in German): household electricity 37.97 cents per kWh in the first half of 2026 (30.55 cents above 15,000 kWh a year), gas 11.58 cents; prices include standing charges, network fees, taxes and levies — Statistisches Bundesamt (Destatis)
9. Canaan's Avalon Q launch release (27 Mar 2025): 90 TH/s, 1,674 W, 18.6 J/TH, 110-240 V AC input, noise range 45-65 dB — Canaan Inc. via PR Newswire
10. Antminer S21 Pro model page: Bitmain's noise rating of 76 dBA with fans at full speed; all 3.5 kW it draws comes back out as heat — SoloLuck
11. Avalon Mini 3 model page: Canaan's 800 W home miner, sold as a space heater that mines — SoloLuck
12. Home Upgrades: today's heat pump can reduce electricity use for heating by approximately 65% compared with electric resistance heating — U.S. Department of Energy
13. Do you know how loud is too loud? (24 Sep 2020): normal conversation 60-70 dBA; sounds at or below 70 dBA generally safe; long or repeated exposure at or above 85 dBA can cause hearing loss — National Institute on Deafness and Other Communication Disorders (NIH)
14. Pooled mining: the block reward is split according to contributed work, measured in shares (proofs of work at a lower difficulty) — Bitcoin Wiki
15. Solo Block Record (last updated 27 Oct 2026): 44 solo-mined Bitcoin blocks since 2022, found mostly by small or home miners on public solo pools, each verified on-chain — SoloLuck
16. Solo Mining Census, edition 2026-W43 (19-25 Oct 2026, frozen 26 Oct 2026): 17,086 Bitaxes connected to Public Pool averaged 1.05 TH/s each — SoloLuck
17. Controlled supply: the block subsidy halves every 210,000 blocks (about four years); 3.125 BTC from block 840,000 (20 Apr 2024), 1.5625 BTC from block 1,050,000 — Bitcoin Wiki

Some specific claims in this article are scheduled for further source review; citations may be expanded (see our editorial policy).

Related educational reading

• How Solo-Mining Probability Works
• How to Evaluate Bitcoin Mining Hardware
• What Happens as Bitcoin's Block Subsidy Approaches Zero?
• What Happens When a Bitcoin Miner Finds a Block?