Bitcoin Mining for Beginners in 2026: How It Works, What It Costs, and What the Numbers Say Today

Mining is how new bitcoin are created and how transactions are added to the chain, and it is also a business with a very small margin for error. Whether it pays for you depends on three inputs you can look up: the network's hashrate, the bitcoin price, and your electricity cost per kilowatt-hour. This guide explains the rules, shows the network's current numbers, and works through a break-even calculation with live data from 1 November 2026, so that you can replace our assumptions with your own. It makes no promise of profit, and at the figures below many home setups would lose money.

The short version

• What miners do: they repeatedly hash a candidate block header until the result is below a target set by the network's difficulty, and the first to succeed adds the block and collects the reward (Bitcoin.org developer guide).
• The reward today is about 3.14 BTC per block, made of the 3.125 BTC subsidy plus transaction fees. In the last 144 blocks, fees averaged about 0.0148 BTC per block, around 0.47% of the reward (mempool.space).
• The network is very large. Average hashrate is about 1.07 zettahashes per second (1.07 x 10^21 hashes per second) and difficulty about 148.4 trillion (mempool.space, 3-day data).
• Revenue per unit of hashrate is small: at the data above and a bitcoin price of $77,417, one terahash per second earns roughly $0.033 per day before costs.
• Electricity decides the outcome. With the same data, a machine using 20 joules per terahash breaks even at roughly $0.068 per kWh, before hardware cost, pool fees and cooling.
• Solo mining at home is a lottery. A 100 TH/s machine would be expected to find a block about once in 200 years.
• The next halving is at block 1,050,000, about 75,957 blocks away, which at ten minutes per block is roughly 527 days, around April 2028.

How mining works, in the protocol's own terms

The Bitcoin.org developer guide describes the process. Mining software builds a block using a template and creates a block header; it sends the 80-byte header to the mining hardware, an ASIC, "along with a target threshold (difficulty setting)". The hardware "iterates through every possible value for the block header nonce and generates the corresponding hash". If no hash is below the threshold, it gets an updated header and tries again. When a hash is below the target, the block is valid, and the miner broadcasts it.

The guide also separates two ways to mine. In solo mining, "the miner attempts to generate new blocks on his own, with the proceeds from the block reward and transaction fees going entirely to himself, allowing him to receive large payments with a higher variance (longer time between payments)". In pooled mining, the miner "pools resources with other miners to find blocks more often, with the proceeds being shared among the pool miners". A pool trades the chance of an occasional large payout for a steadier, smaller stream.

Rules fixed in the software

Three parameters are in Bitcoin Core's source code, the reference implementation of the consensus rules. We read them directly:

• Block spacing: the target time between blocks is ten minutes (nPowTargetSpacing = 10 * 60).
• Difficulty adjustment: the target timespan is two weeks (nPowTargetTimespan = 14 * 24 * 60 * 60), so, at ten minutes per block, difficulty is recalculated every 2,016 blocks. If blocks come faster than ten minutes, difficulty rises; if slower, it falls.
• Subsidy: the subsidy starts at 50 BTC and is cut in half every 210,000 blocks (nSubsidyHalvingInterval = 210000). A block's total reward is the subsidy plus the fees of its transactions: the validation code computes blockReward = nFees + GetBlockSubsidy(...) and rejects a block whose first transaction pays out more.

Because of the second rule, adding more machines to the network does not make blocks come faster for long: difficulty rises to bring the average back to ten minutes. The rule is what makes the economics a competition over a fixed number of coins per day.

The network today (1 November 2026)

Two readings of this table matter. First, daily hashrate estimates swing a lot: mempool.space's daily values over the last month range from roughly 0.87 to 1.23 ZH/s, which reflects how hashrate is estimated from the pace of block production, not a real swing of that size in hardware. Second, the Cambridge power estimate is a model. Its own page gives a theoretical lower and upper bound around the best-guess estimate, and the best guess depends on assumptions about hardware and electricity cost.

Who finds the blocks

Over the last week, mempool.space attributed 1,049 blocks to pools as follows: Foundry USA 278 (26.5%), AntPool 214 (20.4%), F2Pool 167 (15.9%), ViaBTC 102 (9.7%), Luxor 56 (5.3%), SpiderPool 49 (4.7%), MARA Pool 45 (4.3%) and SECPOOL 45 (4.3%). Two cautions. Attribution is by the pool tag in the block, which pools can change. And a pool's share is the combined hashrate of its customers, not a single owner's. Even so, the top three pools together found about 62.8% of recent blocks, which is why pool choice is a decentralization question as well as a payout one.

The economics: from network data to your break-even

Step 1: revenue per terahash

Your expected revenue is your share of the network's hashrate times the coins issued per day. At ten minutes per block there are 144 blocks per day on average. Using the average of 3.14 BTC per block over the last 144 blocks:

Daily BTC for the whole network = 144 x 3.1398 = 452.1 BTC.
Network hashrate in TH/s = 1.069 x 10^21 / 10^12 = about 1.069 x 10^9 TH/s.
BTC per TH/s per day = 452.1 / 1.069 x 10^9 = about 4.23 x 10^-7 BTC.
At $77,417 per bitcoin, that is about $0.0327 per TH/s per day (EUR 0.0294 at EUR 69,403).

The industry calls this "hashprice". It rises with the bitcoin price and falls when hashrate grows. It does not account for pool fees, which pools charge on top.

Step 2: electricity per terahash

A machine's efficiency in joules per terahash (J/TH) equals the watts it draws per TH/s. A machine rated at 20 J/TH draws 20 W for each TH/s, so 1 TH/s costs 20 x 24 / 1000 = 0.48 kWh per day. At EUR or dollar prices per kWh this gives the daily electricity cost.

Step 3: the break-even table

The table shows net daily result per TH/s in dollars, before hardware cost, pool fees, cooling, taxes and any change in the bitcoin price or difficulty. Efficiencies of 17, 20 and 30 J/TH are our illustrations: 17 is the Cambridge network-wide estimate, and 30 is its upper bound.

To make it concrete, consider a machine of 100 TH/s (an illustrative size). At 20 J/TH it draws 2,000 W, or 48 kWh a day. Daily revenue at today's hashprice is about $3.27. At $0.05 per kWh, electricity costs $2.40 and the machine makes about $0.87 a day before everything else; at $0.08, it costs $3.84 and loses about $0.57 a day; at $0.12, it costs $5.76 and loses about $2.49 a day. At 17 J/TH and $0.05, the same hashrate would clear about $1.23 a day. These are small margins on a machine that costs real money to buy and to run.

What the table leaves out

• Hardware cost. Even in the profitable cells, the daily margin must pay back the purchase price. We do not quote prices because they vary by model and seller and we did not verify any.
• Pool fees. They reduce revenue by a percentage that depends on the pool and payout scheme. Read the pool's documentation.
• Difficulty and price. Both change. A fall in the bitcoin price cuts revenue at once; a rise in hashrate cuts it gradually. The mempool.space estimate points to a small difficulty decrease at the next adjustment, which would raise hashprice slightly, but it is only an estimate.
• The halving. At block 1,050,000 the subsidy falls from 3.125 to 1.5625 BTC. Unless the bitcoin price rises or fees grow enough, revenue per TH/s roughly halves at that point. Our halving guide and the halving countdown cover it. The subsidy-and-fees structure matters here: fees are currently only about 0.47% of the reward.
• Cooling, noise and space. A machine drawing 2,000 W becomes about 2,000 W of heat and is loud. In a cold climate the heat can offset heating costs; in a hot one it adds cooling costs. We did not model either.
• Taxes. Mining income is taxable in most places, as we explain below.

How sensitive the result is: price, difficulty and the halving

The break-even electricity price is not a fixed number. It moves one-for-one with revenue per terahash, and revenue per terahash depends on two things you do not control. A short sensitivity check, holding everything else at the 1 November values and using a 20 J/TH machine (0.48 kWh per TH/s per day):

The same logic applies to hashrate in the opposite direction: if network hashrate rises by 10% and the price stays put, revenue per terahash falls by about 9%. Miners see their margin shrink as more machines join, and difficulty is the mechanism that passes that pressure on.

The halving is the largest scheduled change. At block 1,050,000 the subsidy drops from 3.125 to 1.5625 BTC. If fees stayed at today's 0.0148 BTC per block, the total reward would fall from about 3.14 to about 1.58 BTC, 50.2% of today's level. At an unchanged bitcoin price and an unchanged hashrate, revenue per TH/s would fall from about $0.0327 to about $0.0164 a day, and the break-even electricity price for a 20 J/TH machine from about $0.068 to about $0.034 per kWh. In reality price, hashrate and fees will all be different by 2028, and hashrate has historically adjusted to such shocks as less efficient machines switch off. We show the arithmetic only to make the point that a profit calculated today is not a profit guaranteed for the life of the hardware.

A short glossary

• Hashrate: how many hashes per second a miner or the network computes. Units step by a thousand: TH/s (10^12), PH/s (10^15), EH/s (10^18), ZH/s (10^21).
• Difficulty: a number that sets how hard it is to find a valid block. It adjusts every 2,016 blocks.
• ASIC: a chip built only for Bitcoin's hashing. General computers and graphics cards cannot compete at this task.
• J/TH: joules per terahash, a measure of efficiency. Lower is better. It equals watts per TH/s.
• Hashprice: expected revenue per unit of hashrate per day, as computed above.
• Subsidy: the newly issued bitcoin in a block, now 3.125 BTC, as distinct from the fees paid by transactions.
• Pool: a service that combines miners' work and splits the rewards, usually for a fee.
• Nonce: the number a miner changes in the block header to produce a different hash on each attempt.

Mining versus simply buying bitcoin

Mining and buying are different bets. Buying exposes you to the bitcoin price. Mining exposes you to the bitcoin price too, because revenue is paid in bitcoin, and in addition to electricity costs, hardware depreciation, difficulty and the halving. A miner who sells each day's output to pay the electricity bill is, in effect, selling bitcoin at the going price to fund a machine. The arithmetic above shows that at $0.05 per kWh a 100 TH/s machine at 20 J/TH nets about $0.87 a day, which is about $26 a month, before anything else. Whether that is worth the hardware outlay, the noise and the effort is a personal calculation we cannot make for you. If your aim is simply to hold bitcoin, buying it is simpler and does not depend on your electricity contract. If your aim is to learn how the network works, a small setup can be a fair price for the education, as long as you treat the spend as a cost.

Solo mining odds, with arithmetic

A miner with 100 TH/s holds a share of the network of 100 x 10^12 / 1.069 x 10^21, about 9.4 x 10^-8. With 144 blocks a day, the chance of finding at least one block in a day is about 1.3 in 100,000 (1.35 x 10^-5), so the expected wait is around 74,000 days, about 200 years. That is the expected value of the wait; luck can shorten or lengthen it, as in any lottery, and each block would pay about 3.14 BTC. Small "lottery miners", which are hobby devices with a fraction of a terahash, have a proportionally lower chance. The sensible use of solo mining at this scale is education or entertainment, not income.

That is the logic of pools: a pool combines many such shares so that blocks are found regularly, and distributes the proceeds in proportion to contributed work, less a fee. The pool concentration discussed above is the price of that smoothing.

Reading a pool dashboard without being misled

Once you point a machine at a pool, you will see figures such as a current hashrate, an average over a day, and an estimated daily or monthly earning. Three habits help you read them. Look at the longer averages and not the instantaneous figure, because a miner's reported hashrate swings with luck and reporting delays. Treat any "estimated earnings" line as an extrapolation from the pool's own assumptions about price and difficulty, and compare it with your own hashprice calculation. And check how the pool pays: some schemes pay for work submitted regardless of whether a block is found, others pay only when a block is found, so the timing and variance of your income differ. We did not review individual pools' payout rules, so read the pool's own documentation before you commit hardware to it.

Energy: what the best public estimate says

The Cambridge Centre for Alternative Finance (CBECI) estimates Bitcoin's network power demand at about 19.74 GW, equivalent to about 173 TWh a year, with a theoretical lower bound of 10.12 GW (88.7 TWh) and an upper bound of 37.97 GW (332.9 TWh). Its estimate of average hardware efficiency is about 17 J/TH, between 9.5 and 30. These are modelled figures that rest on the assumption that miners are rational and on an average electricity cost the model lets you adjust. They are the best publicly available estimate we found, and they should be read as a range, not a measurement.

Practical steps for a beginner

1. Start with your electricity price, including all taxes and network charges, per kWh. Compare it with the break-even column above. If it is above about $0.07 per kWh (roughly EUR 0.06), the tables suggest home mining of current hardware does not cover its own power cost at today's price. Check your own numbers with the formula.
2. Compute your own hashprice with current data: 144 x block reward / network hashrate in TH/s x bitcoin price. Public explorers show the inputs.
3. Treat hardware quality and sourcing as a risk. The 2026 hardware wallet incidents showed how products can be tampered with in distribution (see our reconstruction of the Ledger case). Buy from the manufacturer or an authorized seller, and keep the machine's management interface off the open internet, with changed default passwords.
4. Use a wallet you control for payouts, ideally a hardware wallet bought directly from its maker. See our hardware wallet guide.
5. Be wary of "cloud mining" and guaranteed-return offers. A promise of fixed returns from mining is a warning sign in our crypto scams guide. Mining returns are not guaranteed by anyone, as the table shows.
6. Plan for tax from day one. Keep records of every payout: date, amount, fiat value at receipt. In Germany, the finance ministry's 2025 letter treats income from block creation that is not a commercial activity as service income under section 22 no. 3, with an exemption limit of EUR 256 per calendar year, and says expenses such as hardware (including tax depreciation) and electricity are deductible (paragraphs 45 and 47). Our guide to staking, airdrop and DeFi taxes reads that letter and HMRC's manual. Other countries differ.
7. Start small, if at all. Treat the first purchase as a test of your electricity and cooling assumptions, not as an investment you expect to be paid back on a schedule.

What we could not verify

• Hardware prices and the efficiency of specific current models. We use illustrative efficiencies only.
• Pool fee levels and payout schemes. They differ by pool and change.
• Mining regulation and electricity-contract rules in individual countries.
• Future bitcoin price, difficulty and fees, which determine real outcomes. The table is a snapshot of 1 November 2026.
• The CBECI figures are modelled estimates, and we did not review its methodology in detail.

Frequently asked questions