Bitcoin Mining at Home in Germany, Austria and Switzerland: Mined Against Bought Bitcoin on Household Power, Solar Surplus, as a Heater and as a Lottery Ticket

Abstract. We calculate what eleven miners on sale in November 2026, from an 18-watt Bitaxe to a 3.5-kilowatt farm machine, yield in a home in Germany, Austria or Switzerland. Each miner runs for six years against a network whose hashrate follows our hashrate model. The mined bitcoin are kept and valued at the floor of our Power Law model, and every case is compared with spending the same money on bitcoin instead. A miner bought today mines about 36’500 sats per TH/s in its life, two thirds of them in the first two years. On household power at 33 to 41 US cents per kWh, no miner covers its cost by selling what it mines, and the same money buys four to seven times as many sats. The picture changes only where the electricity is spent anyway. In place of an electric heater a miner gives the same warmth and pays for itself within one to seven years, and a heater miner that is cheap for its hashrate mines more sats than its price would buy. On solar surplus a miner covers its cost, but feeding the power in and buying bitcoin with the tariff yields two to six and a half times as many sats. A solo miner has a far better chance of its main prize than a lottery line for the same money, but returns less of the stake on average unless its power is paid for anyway. For the network one home miner's hashrate is negligible; what it adds is one more participant who chooses a pool or builds blocks of its own.

Keywords: Bitcoin, home mining, Bitaxe, electricity price, heating with a miner, solar surplus, solo mining, lottery, decentralisation, Power Law.

Version 2, 14 November 2026. Prices, hashrate and device prices of 14 November 2026. Version 1 took the network from the first version of our hashrate paper. This version takes it from version 4 of that paper, which counts the network's machines and lets the big farms count a mined coin at what it is worth a year ahead [3]. The network grows a little more slowly in the first years, so a miner mines 4 % more sats than in version 1, 36’561 instead of 35’113 per TH/s, and the test in section 11 expects 15’800 instead of 15’500 sats. No finding changed its direction. Every result can be recalculated in the mining calculator, which runs the model of this paper with its scenarios Home Mining, Solar, Heater and Whales. The experience behind the questions is in our field report Bitcoin mining in Switzerland.

How to cite: Bitcoin Power Law Research (2026). Bitcoin Mining at Home in Germany, Austria and Switzerland: Mined Against Bought Bitcoin on Household Power, Solar Surplus, as a Heater and as a Lottery Ticket. Version 2, 14 November 2026. https://bitcoin-powerlaw.com/en/magazin/home-mining-paper

1 Introduction

In Germany, Austria and Switzerland a household pays seven to nine times as much for electricity as a mining farm. Mining at home is therefore usually written off, and as a business it should be. Small miners are sold all the same, and their owners give four reasons: they want to stack bitcoin and pay the costs in francs or euros, they heat with the miner, they have solar power nobody pays well for, or they treat a solo miner as a lottery ticket. A fifth reason is the network itself, which should not be run by a few farms and pools.

One of us runs miners for these reasons: a Bitaxe on the wall and an Avalon Nano 3S in place of an 800-watt towel heater, with a family solar roof whose surplus earns little as the next use [1]. This paper puts numbers on each reason. It asks four questions.

1. Household power: what does a miner yield on a normal tariff, and how does that compare with buying bitcoin for the same money?
2. Power that is spent anyway: what changes when the miner replaces an electric heater, or runs on solar surplus?
3. The lottery ticket: is a solo miner a better ticket than the state lottery?
4. Decentralisation: what does one home miner add to the network?

We count cautiously in three ways. The mined bitcoin are valued at the floor of our Power Law model, the lowest of its lines [2]. The network grows as our hashrate model says, which is faster than a miner would wish [3]. And the comparison is always the simplest alternative: taking the same money and buying bitcoin.

2 Data

All figures are of 14 November 2026.

• Network: hashrate 905 EH/s; a block pays 3.125 new bitcoin plus 0.029 bitcoin in fees, the average of the last 144 blocks (mempool.space) [4]. The next halving falls in April 2028, the one after it in April 2032.
• Price: $65’432 per bitcoin. The Power Law floor is $67’214, the trend (fair value) $108’930. Ten years on, in November 2036, the floor is $842’624 [2].
• Exchange rates: 1 euro = 1.1052 US dollars, 1 franc = 1.1927 US dollars.
• Miners: eleven devices with their hashrate, power draw and shop price in US dollars including a power supply, as listed in the calculator (Table 1) [5].
• Household electricity: Germany 37.0 cents per kWh (BDEW, 2026) [6]; Austria 32.6 cents (Eurostat, first half of 2026, 2’500 to 4’999 kWh a year, all taxes included) [7]; Switzerland 27.7 Rappen (ElCom, median 2026) [8]. In US cents that is 40.9, 36.0 and 33.0.
• Solar feed-in: Germany 7.70 cents per kWh for new roof systems up to 10 kW that use part of their power themselves [9]; Switzerland at least 6 Rappen for systems under 30 kW [10]. In US cents: 8.5 and 7.2.
• Heating: natural gas for a German single-family home cost 11.93 cents per kWh in August 2026 [11]. A heat pump makes about 3 kWh of heat from 1 kWh of electricity.
• Lotteries: Swiss Lotto pays out 54.5 % of the stakes; a line costs CHF 2.50 and wins the jackpot with a chance of 1 in 31’474’716 [12][13]. The German LOTTO 6aus49 pays out 50 %; a line costs €1.20 and wins the top class with a chance of 1 in 139’838’160 [14][15].
• Pools: the blocks each pool found in the twelve months to 14 November 2026, 52’201 blocks in all (mempool.space) [4].

Table 1: the miners. J/TH is the power a miner needs for its work, the price per TH/s what its hashrate costs. Sats a day: in a pool with a 2 % fee, at the hashrate of 14 November 2026. The Antminer S21 is a farm machine, too loud for a home, and is listed for comparison.

3 Method

In plain words: a miner gets the share of every block that its hashrate has in the network's hashrate. Its electricity costs the same every month, while its share shrinks as the network grows. We follow one miner month by month for its lifetime, add up the bitcoin and the costs, and at the end compare three things: what was paid, what the mined bitcoin are worth, and how many bitcoin the same payments would have bought.

bitcoin mined in a month = hours in use ÷ 24 × miner's hashrate ÷ network's hashrate
                           × blocks in the month × (new bitcoin + fees per block) × (1 − pool fee)
power cost in a month    = hours in use ÷ 24 × watts ÷ 1000 × hours in the month × price per kWh
result                   = bitcoin mined × floor at the end − power cost − price of the miner
bought instead           = price of the miner ÷ buy price on day one
                           + each month's power cost ÷ buy price in that month
buy price                = Power Law trend (fair value) × 1.1

The assumptions are those the calculator starts with:

• Lifetime 6 years, counted over 10 years. The miner runs from November 2026 to November 2032. Its bitcoin are kept and valued at the floor of November 2036.
• The network follows the model of our hashrate paper, version 4, on the floor path [3]. The model counts the network's machines: a machine runs for five years and never gets better, and new machines need 20 % less power per hash each year. The big farms pay USD 0.045 per kWh for power and USD 0.0105 per kWh for their other running costs, and they count a mined coin at what it is worth on that path a year later. The closer their costs come to that value, the more slowly the network grows. It grows from 905 EH/s to 1’222 EH/s in November 2027 and 5’488 EH/s in November 2032, on average 35 % a year.
• The reward halves in April 2028 and April 2032; the fees stay at 0.029 bitcoin per block. The pool keeps 2 %.
• Buying instead means buying at the Power Law trend plus 10 %: the miner's price on day one, the power money month by month. The trend is our measure of what a bitcoin is worth in the long run, so it stands for the price a regular buyer pays over the years, and the 10 % for fees and less lucky timing. Today's price is far below that, which section 9 takes up.
• Not counted: shipping, import VAT and customs on the miner, repairs, resale value and taxes on mined bitcoin.

Three kinds of electricity are compared. Household power costs the tariff. Solar surplus costs what the grid would have paid for it. Heat: a miner turns practically all of its electricity into heat, like an electric heater of the same wattage. Where a room is heated with electricity anyway, the miner's power costs nothing extra.

Amounts are in US dollars, because that is what the miners are sold in. Local prices are given in the local currency as well.

4 What a miner earns over its life

On 14 November 2026 one TH/s earned 49 sats a day, which is 3.3 US cents at the floor. That does not last. Year by year, one TH/s that runs around the clock mines:

Table 2: the bitcoin one TH/s mines in each year of its life, with a 2 % pool fee.

A miner bought today mines about 36’500 sats per TH/s in six years. 43 % of that comes in the first year and 67 % in the first two. The sixth year adds 3 %. How long a miner lasts beyond four or five years therefore matters little, and a miner that stands in its box for the first half year has lost almost a quarter of its yield.

This number says at once whether a miner can ever return its price in bitcoin. 36’500 sats cost $44 at our buy price and $24 at the price of 14 November. A miner that costs more than $44 per TH/s mines fewer sats than its price would buy, even if its power is free and it never stops. Seven of the eleven miners in Table 1 are above that line, among them every Bitaxe. Below it are the Avalon Nano 3S and the larger machines. At today's price only the Avalon Mini 3, the Avalon Q and the Antminer S21 remain.

The electricity adds a second number. A miner mines 2’050 sats per kWh today, divided by its J/TH: 215 sats for the Bitaxe Naja Duo, 88 for the Avalon Nano 3S. Over six years the average is a third of that, 73 and 30 sats per kWh.

5 Mining on household power

Table 3 gives four power prices for every miner, each for running around the clock. The first is the usual profitability figure: the price up to which today's sats pay for today's power. The other three count the whole life of the miner, its price included.

Table 3: the highest power price, in US cents per kWh, at which each statement holds. "Never" means not even with free power. Households pay 33.0 (Switzerland), 36.0 (Austria) and 40.9 (Germany).

As a business it does not work. Whoever sells the bitcoin as they are mined needs power for about 4 US cents per kWh or less, and that with a farm machine. That is a tenth of a household tariff and below what the big farms pay. Even the first column, which leaves out the miner's price and the growing network, ends at 14 cents.

Keeping the bitcoin changes the result, but not because of the mining. Table 4 shows six miners on household power, with their bitcoin valued at the floor of November 2036.

Table 4: six years around the clock on household power in Switzerland (CH), Austria (AT) and Germany (DE). Cost: the miner and its power. Worth in 2036: the mined bitcoin at the floor of November 2036.

Only the Bitaxe Naja Duo, the most efficient miner on the list, ends above its cost in all three countries, in Germany by $67. The NerdQAxe++ does so in Switzerland alone, by $24. They do so because the floor is 12.5 times as high in 2036 as today. The mining adds nothing to that. The same money spent on bitcoin would have taken part in the same rise with far more sats:

Table 5: how many times as many sats the same money buys, spent as in Table 4 but on bitcoin at the trend plus 10 %.

On household power, buying yields four to seven times as many sats as mining. Put as a price: a bitcoin from the Naja Duo in Switzerland costs $454’000 in electricity alone and $693’000 with the device. The buyer who spends the same money over the same six years pays $173’000 on average.

For comparison, the same Antminer S21 on power at USD 0.045 per kWh, the big farms' price: it mines the same 7.3 million sats for $9’484 instead of $62’020, and that is 57 % more than the money would buy. The machine is the same. The difference is the price of the power. What a farm pays beyond its power, and what it has left when it sells coins to pay its costs, is in our hashrate paper [3].

6 Solar surplus

Surplus solar power is not free. Fed into the grid it earns the feed-in tariff, and that is what it costs to give it to a miner instead. The calculator's Solar scenario lets the miner run 8 hours a day on average over the year.

Table 6: six years on solar surplus, 8 hours a day. The power costs the feed-in tariff given up: 6 Rappen in Switzerland, 7.70 cents in Germany. Last two columns: the sats the same money buys, as a multiple of the sats mined.

Every miner in Table 6 ends above its cost, the Naja Duo only just, so a miner is a use for surplus power that pays for itself. It is not the best use. Whoever feeds the power in, takes the tariff and buys bitcoin with it and with the price of the miner ends with two to six and a half times as many sats. The reason is the miner itself: it stands still two thirds of the day and earns most in its first two years.

Two things shift this. Fewer hours make it worse: a large machine finds enough surplus for perhaps 3 hours a day over the year, and at 3 hours buying yields 3.7 times as many sats as the Avalon Q mines. Surplus that earns nothing makes it better: new German systems get no tariff in hours with a negative exchange price, and some systems may not feed in at all. But even with power that costs nothing, 8 hours a day are not enough: the price of the Avalon Q alone buys 1.1 times the sats it mines, that of the Nano 3S 2.5 times. How to tie a miner to the surplus is described in Bitcoin mining with solar power.

What remains in the miner's favour is not in the table. A small system's owner who would have to register, bill and declare a few francs of feed-in may prefer a miner that simply takes the power [1].

7 The miner as a heater

Energy is not lost. A miner that draws 140 watts warms the room with 140 watts, exactly as an electric heater does. Whoever heats a room with electricity anyway can let a miner do it and gets the bitcoin on top. Two conditions belong to this sentence.

It holds only in place of an electric heater. In Germany a kWh of heat from an electric heater costs 37.0 cents. From a miner it costs the same, less the bitcoin it mines: today 5.3 to 7.5 cents per kWh, over the miner's life 3.2 to 4.5 cents, valued at the floor when mined. A heat pump makes the kWh of heat for about 12 cents, and a kWh of gas costs 11.93 cents. Heat from a miner therefore costs nearly three times as much as heat from a heat pump or from gas. In Switzerland the sums are 27.7 Rappen, with 5.0 to 6.9 Rappen back today and 3.0 to 4.2 over the miner's life.

The miner has to suit the room. Farm machines are as loud as a vacuum cleaner. The heater miners in Table 7 are built for living space. The Avalon Q, at 1’674 watts, draws more than a household socket in Switzerland or Germany should carry for hours.

So the bitcoin give back 14 to 25 % of the electricity today and about a tenth over six years. If the bitcoin are kept, each kWh's sats are worth 21 to 30 Rappen at the floor of November 2036, about what the kWh cost.

Since the electricity is paid for the heat, only the miner's price counts against the bitcoin. The calculator's Heater scenario lets it run 12 hours a day on average over the year, the heating season.

Table 7: a miner in place of an electric heater, its power paid for the heat. Paid for: the month in which the mined bitcoin, at that month's floor, are worth the miner's price. The price buys instead: the sats the miner's price buys on day one at the trend plus 10 %.

Every heater miner pays for itself, after one to seven years. Whether it also beats the plain alternative, a simple heater plus bitcoin bought for the difference, depends on its price per TH/s and its hours. The Avalon Q mines 41 % more sats in 12 hours a day than its price buys, the Mini 3 10 % more. The Nano 3S mines 40 % fewer at 12 hours and 20 % more around the clock. The small Bitaxe-class heaters stay behind in every case.

The case this paper started from. In a Swiss bathroom an Avalon Nano 3S replaced the heating element of a towel radiator [1]. Two effects have to be kept apart.

• The element drew 800 watts, the miner draws 140. At 27.7 Rappen that is CHF 340 a year instead of CHF 1’940. This saving of CHF 1’600 a year comes from the smaller wattage. A 140-watt heater without a mining chip would save the same.
• The mining is what comes on top: by the model about 95’000 sats in the first year, worth CHF 61 at the floor when mined, 18 % of the miner's electricity. Over six years it is 219’368 sats, worth CHF 1’550 at the floor of November 2036, for a device that costs CHF 184 today.

The towels dry either way, and the miner is paid for by its bitcoin in April 2029. That is the heater case at its best: a device that is cheap for its hashrate, in use around the clock, in place of a heater that was running anyway.

8 Solo mining against the lottery

A solo miner does not share. It mines alone, and if it finds a block, the whole reward is its own: 3.15 bitcoin today, about $206’000. On average it earns exactly what a pool would pay, all at once or never (solo mining). The claim to be tested is that this is a better lottery ticket than the state lottery. It has two parts: the chance of the main prize, and what comes back on average.

The chance. As the network grows, the chance of a block shrinks like the yield in Table 2.

Table 8: the chance of at least one block, under the model's network and, for comparison, with the hashrate of 14 November 2026 held fixed.

Compare the Bitaxe Gamma 602 with lottery lines for the same money. In Switzerland the device costs CHF 56. That buys 22 lines of Swiss Lotto, a chance of 1 in 1.4 million at the jackpot, against 1 in 4’900 for the miner: 290 times the chance. With six years of household power the stake is CHF 319, which buys 127 lines and a chance of 1 in 247’000: the miner's chance is still 51 times as high. In Germany the stake with household power is €411, or 342 lines of 6aus49 and a chance of 1 in 409’000 at the top class: 84 times in the miner's favour. The prizes differ. A block is worth $165’000 to $480’000 at the floor in these six years, a jackpot is counted in millions.

What comes back on average. A lottery pays out half of what it takes. For the miner we divide the bitcoin it mines on average, valued at the floor in the month it finds them, by its stake. Solo services keep a fee as pools do; we count 2 %.

Table 9: the share of the stake that comes back on average over six years. Power paid for anyway: the miner heats in place of an electric heater, so the stake is the device.

The verdict has two halves. For the chance of the big win the claim holds everywhere: per franc or euro a solo miner is 50 to 290 times as likely to win its main prize as a lottery line. For the average return it holds only where the miner's power is paid for anyway: then 35 to 120 % of the stake come back, against 50 to 55 % from the lottery. On household power bought for the miner alone, 8 to 17 % come back, and the lottery is the better ticket on average.

Two remarks. The values are at the floor; at any price above it, more comes back. And unlike a lottery player, the miner can have the average without the luck: in a pool the same device pays the same bitcoin in small amounts.

9 How firm are the results?

The results rest on a forecast of the network and on a buy price. Table 10 changes one assumption at a time.

Table 10: one assumption changed at a time. Last two columns: the sats bought as a multiple of the sats mined.

• The network. The better mining pays, the faster the network grows and the fewer sats a miner gets. A higher bitcoin price does not help the miner in sats. It helps the buyer and the miner alike in dollars.
• Household power. Buying stays ahead in every case, by a factor of 2.3 even if the network never grew again.
• The heater is the close case, as section 7 showed. For the Nano 3S at 12 hours a day buying stays ahead unless the network stops growing. For the Mini 3 and the Avalon Q the same column reads 0.91 and 0.71 in the base case and 1.1 and 0.86 on the trend path.
• Today's price. Our buy price is the trend plus 10 %, $119’822 on 14 November 2026. On that day a bitcoin cost $65’432, just under the floor. Whoever buys near the floor gets 50 to 60 % more sats than our comparison assumes, and every case tilts further towards buying: at the floor plus 10 % the Avalon Q's price buys 1.1 times the sats it mines in 12 hours a day. Mining cannot be timed in this way. Its sats arrive as the network allows, whatever the price.

The calculator runs the cases of this paper with the day's figures. Choose the scenario, the miner and your power price; hours in use, lifetime and pool fee are under "More settings", the price path and the comparison with buying under the chart. The rows of Table 10 that change the network itself are calculated with the hashrate model [3]. The paper's figures are in US dollars:

10 Decentralisation

In the twelve months to 14 November 2026 the largest pool, Foundry USA, found 27.9 % of all blocks, AntPool 18.3 % and F2Pool 13.1 %. The three together found 59.3 %, the largest five 77.4 % [4]. In most pools the operator alone decides which transactions go into a block (mining pools).

What can a home miner do about that? In hashrate, close to nothing. A Bitaxe Naja Duo is one part in 206 million of the network. A million of them would be 4.4 EH/s, half a percent. One percent of the network in Avalon Nano 3S heaters would be 1.5 million devices drawing 211 megawatts.

What a home miner adds is of a different kind.

• A choice. Every miner decides which pool gets its hashrate and can change it in minutes. A small or mid-sized pool grows through such decisions.
• Its own blocks. A miner that mines solo against its own node, or in a pool that lets its miners build the block (Stratum V2, or the DATUM protocol of the pool OCEAN), decides for itself which transactions it includes. The solo services together found 27 of the 52’201 blocks of the last twelve months.
• Knowledge. People who have run a miner know what hashrate, difficulty and a block template are, and are harder to mislead about how the network works.

These are reasons, and they cannot be counted in sats. They cost what sections 5 to 7 say: on household power the difference to buying, in place of an electric heater the price of the device.

11 A test

The yield per TH/s is the number everything here depends on, and it can be checked within a year.

A miner that runs around the clock from 14 November 2026 to 14 November 2027 in a pool with a 2 % fee should earn about 15’800 sats per TH/s. That is the model on the floor path, with the network at about 1’220 EH/s at the end. With the network unchanged at 905 EH/s it would be 18’000 sats, on the trend path 14’400.

If the year's yield is below 14’400 sats per TH/s, the network has grown faster than our model allows even on the trend path, and every mining result in this paper is too favourable. If it is above 18’000, the network has shrunk and the results are too cautious. For the Avalon Nano 3S of section 7 the three values are 95’000, 108’000 and 86’000 sats. We will compare them with the device.

12 Limitations

• The hashrate forecast. The model behind the network's growth missed 2026: it called for growth and the hashrate fell [3]. If the network grows more slowly than assumed, mining does better than stated here (Table 10).
• The floor. The mined bitcoin are valued with our Power Law model, which has its own limits [2]. On 14 November 2026 the price was 3 % below the floor.
• Device prices are shop prices in US dollars on one day. Shipping, import VAT and customs come on top in Germany, Austria and Switzerland, which makes mining look somewhat better here than it is.
• Device data are the makers' and shops' figures for hashrate and power draw. We measured none of the devices.
• Lifetime. Six years is an assumption. In small devices the power supply often fails first. Because most of the yield comes early, the result depends little on the last years.
• Heat. We assume every kWh of the miner replaces a kWh of an electric heater. A miner that also runs when no heat is needed buys that electricity for the bitcoin alone.
• Electricity prices are national averages or medians of 2026. Tariffs differ widely between suppliers and municipalities, and they change over six years.
• Solar hours. Eight hours a day on average over the year fit a small miner on a large roof. Large machines find enough surplus far less often.
• Lotteries. Only the payout share and the top prize class are compared. The Austrian lottery is left out because we found no reliable payout figure.
• Taxes on mined bitcoin, on sales and on lottery winnings are not counted.
• No financial advice. The paper describes what follows from stated assumptions. It does not say that a mining device will pay, or that bitcoin's price will follow the floor.

13 Data and reproducibility

Every input is public: the hashrate, the reward per block and the pools' blocks from mempool.space (/api/v1/mining/hashrate/1m, /api/v1/mining/reward-stats/144, /api/v1/mining/pools/1y), floor and trend from our daily file /power-law/btc-daily.csv, the electricity prices, tariffs and lottery figures from the sources below. The formulas are in section 3, the hashrate model in [3] and on the methodology page. The mining calculator runs the same model with the day's hashrate and prices, so its results move away from the figures of 14 November 2026 as time passes. You are welcome to use, reproduce and extend this work. Please cite it as shown at the top.