Is Bitcoin mining still profitable in 2026?

I took the 9 machines from my reviews, with the figures from their pages: measured at the wall when I did it, read from their interface or rated by the maker otherwise. For each one: what it would earn on average in a pool A server that machines connect to. A regular pool shares the earnings, a solo pool does not. Solo or pool?, its electricity over a year, and the price per kWh below which it would pay for its own power.

None of them makes it. The Thor X1 comes closest because it is the most efficient, and only loses about €14 a year. The heaviest loss is the NerdQX 8T’s: about €196 a year, for €93 of earnings.

Average earnings are what a regular pool would pay, sharing its blocks according to everyone’s work, before its fee, often a few percent. Electricity counts the machine running day and night, as it really does. The purchase price is not counted: as long as the balance stays negative, it never pays itself back. Even with free electricity, the machine that would earn back its price fastest, the NerdQX 8T (about €284), would need about 3.1 years to do it, at the same difficulty, and the 2028 halving The reward in new bitcoin is cut in half every 210,000 blocks, roughly every four years. The last one, in April 2024, brought it to 3.125 BTC. would come first.

For ranges, the math uses the value from my comparison: 1.2 TH/s and 18 W for the Bitaxe Gamma, 9.65 TH/s and 145 W for the NerdOctaxe, 0.16 TH/s and 10 W for the NerdNOS. “From the interface”: figures shown by the machine, without power supply losses. The NerdQAxe++ Hydro’s rounded rated figures put it high in this ranking, but my review gives it slightly worse efficiency than the NerdOctaxe, above 15 J/TH The energy used for a trillion attempts: the lower, the more efficient the machine. A Bitaxe Gamma runs at about 15 J/TH. The comparison.

Three figures are enough, those of the network as of 17 October 2026:

1. Earnings: multiply the machine’s hashrate in TH/s by €11.9. That is what it would earn on average over a year in a pool, before fees.
2. Electricity: multiply its power draw in watts by 8.76, then by your price per kWh.
3. Break-even price per kWh: divide the earnings by the year’s kWh, the watts times 8.76. If your kWh costs less, the machine pays for its power.

It all comes down to efficiency, power draw divided by hashrate, in J/TH. At €0.20 per kWh, a machine only covers its electricity below 6.8 J/TH. The Thor X1, measured at the wall, runs at 10.6 J/TH; most others sit around 15 J/TH or above. A computer or a phone is out of the game: a CPU hashes thousands of times slower than a small Bitaxe, and even a high-end graphics card stays nearly a hundred times behind, for far more watts (mining on a PC or a phone).

The €0.20 used across this site is not random: it is the price per kWh of EDF’s regulated tariff in France, Base option, €0.2001 including tax for 3 and 6 kVA since 1 August 2026. Even the cheapest kWh of that tariff, off-peak hours on a blue Tempo day at €0.1356, stays above the Thor X1’s break-even price (€0.129), and a miner also runs during peak hours. On Tempo, one day of my 550 W shelf costs me about €2.05 on a blue day, €2.37 on a white day and €7.13 on a red day. Outside France, compare the last column of the table with your own price per kWh. Plans, subscription and off-peak hours are covered in the guide to a miner’s power consumption.

The earnings per TH/s are also a guardrail against cloud mining offers: a contract promising more than €0.033 a day for each rented TH/s promises more than the network itself pays on average (spotting a cloud mining scam).

Five things move the table, and you only choose two of them: the machine and the pool.

• The bitcoin price. Earnings follow it: up 10%, earnings up 10%, and the reverse on the way down. The table below gives the price each machine would need to cover its electricity.
• Network hashrate and difficulty. The more machines in the world, the smaller your share. Difficulty adjusts every 2,016 blocks, about two weeks, to keep one block every ten minutes whatever the network hashrate (mining difficulty explained). When it rises 10%, your earnings drop about 9%. It has been multiplied by 2.3 since late October 2023 (62 T), even though it is about 10% lower than in late October 2025 (156 T), according to mempool The waiting room for transactions not yet written into a block. Whoever builds the block chooses which ones go in..space’s history.
• The 2028 halving. Every 210,000 blocks, the subsidy of each block is cut in half: it is written in Bitcoin Core’s code. The next one comes at block 1,050,000. On 28 October 2026, the chain was at block 973,490: about 76,500 blocks to go, a year and a half at ten minutes per block, so around April 2028. The subsidy will then drop from 3.125 to 1.5625 BTC, and at the same price and difficulty, every machine’s earnings would fall to about 50% of today’s.
• Fees. Transaction fees, paid to the block’s miner, currently add about 0.012 BTC per block to the 3.125 BTC subsidy, 0.4% of the reward over recent blocks. Pool fees, on the other hand, come out of your earnings.
• Your machine: its efficiency sets its break-even price per kWh, and it is the only lever that does not depend on the network.

All else being equal, the Thor X1’s break-even price per kWh would fall to €0.065 after the halving. What price and difficulty will do by then, nobody knows: this table only shows where the thresholds are.

Not on average. Solo, you do not receive those small amounts: you receive the whole block, about 3.137 BTC, roughly €227,000 at the price of 17 October 2026, on the day your machine finds it, and nothing on other days. With the Thor X1, the odds are about 1 in 9,120 per year. Multiply the block by those odds: about €25 a year, the average earnings in the table. Pool or solo, the average is the same; only the way you receive it changes (solo or pool, which one).

The lottery is sometimes won: in July 2026, block 957,382 was found on Public Pool by a miner who reports a single Bitaxe Gamma, about 1 TH/s (solo blocks found). But most small machines will never find one. And to build up one whole bitcoin in a pool, the Thor P2, the most powerful machine in my reviews, would need about 610 years at the same difficulty (how long to mine one bitcoin). The solo mining calculator redoes these odds for your machine.

I mine solo myself, on FindMyBlock, my own pool: 0% fee, and the reward goes straight to the address of the miner who finds the block (the settings).

It can tip the balance. A miner turns almost all the electricity it uses into heat, just like an electric heater. In a room heated by thermostat-controlled electric heaters, each kWh from the miner therefore saves roughly one kWh of heater use: during those months, its electricity is no longer a mining cost, it is heating.

The math then changes: the Thor X1 pays for its whole year of electricity if its heat replaces an electric heater for about 4 months a year; the Thor P2, measured at the wall at 15.3 J/TH, needs about 7 months. At home, in winter, the heat from my shelf warms the room, and the Avalon Nano 3S, with its adjustable air outlet, even doubles as a small space heater.

Three limits. The reasoning only holds against electric heating at the same price per kWh: if you heat another way, count the price of your heat instead. In summer, that heat adds to the house’s own. And a desk miner is only a small space heater. The detailed math, with the machines that heat best: heating your home by mining Bitcoin.

I mine at a loss, knowingly: my shelf of 13 machines draws 500 to 600 W around the clock, about €70 of electricity a month on my French Tempo plan, and the block still has not come. I talk about it in the video at the top of this page. My reasons, for what they are worth, are the same as in the Bitaxe math:

• Understanding Bitcoin by running it at home, far better than reading about it.
• Taking part: over the twelve months to late October 2026, two pools, Foundry and AntPool, found almost one block in two (mempool.space). Every machine mining solo in someone’s home is a bit of hashrate none of these giants control.
• The thrill of the ticket: a tiny but real chance of winning a whole block.
• The heat, put to use in winter.
• Depending on no one: my node A computer that keeps a full copy of Bitcoin and checks every block itself. Run your own node checks every block itself.

If the only goal is to own bitcoin, let us be honest: at household prices, buying is cheaper. A year of the Thor X1’s electricity, about €39, would buy about 53,200 sats at the price of 17 October 2026, before exchange fees; the machine mines about 34,400 in the same time. This is not investment advice, only today’s math: it moves with the price and the difficulty.

Not on average. Bitcoin Cash uses the same calculation, SHA-256: the same machines mine it. Its network is about 270 times smaller (Bitcoin Cash figures from 28 October 2026): the same machine has about 270 times the chance of finding a block there, but a block is only worth about €840 (3.125 BCH). In the end, 1 TH/s earns about €12.0 a year there before pool fees, against €11.9 on Bitcoin: practically the same, because miners move from one network to the other until the two balance out. What changes is the lottery: about 1 chance in 59 per year of finding a block with a Bitaxe Gamma, about 13% with a Thor P2. My FindMyBlock pool also has a Bitcoin Cash version, with a 1.5% fee taken only on a block found: solo mining Bitcoin Cash.