How Bitcoin mining works: hardware, pools, and the economics of securing the network

Bitcoin mining creates new BTC and secures every transaction on the network.  Miners compete to solve a computational puzzle, and the first to find a valid answer earns the right to append the next block of transactions to the blockchain. That miner collects the block reward, the newly issued BTC the protocol mints, along with the network fees attached to the transactions in the block. The mechanism behind this competition is called Proof of Work, and it is what makes Bitcoin's ledger tamper-resistant without any central authority keeping the books.

Disclaimer: This guide is for educational purposes only. It is not financial advice, not a solicitation, and not for UK audiences. Bitcoin mining and cryptocurrency mining are risky and not suitable for all users.

What Bitcoin mining actually does

Bitcoin mining does two jobs at once. It issues new bitcoin, and it secures every transaction on the network. 

Every Bitcoin transaction waits in a holding area called the mempool until a miner picks it up. Miners pull transactions from the mempool, usually favoring those offering higher network fees, and assemble them into a candidate block. From there, the hardware goes to work.

A miner's machine hashes the block header over and over, running through billions of values called nonces every second, hunting for an output that lands below a target the network sets. Bitcoin's hash function, SHA-256 (Secure Hash Algorithm 256-bit), turns any input into a fixed-length, effectively random-looking output. There is no clever shortcut. Finding a valid hash is brute-force trial and error.

Once a miner lands a valid hash, it broadcasts the finished block. Other nodes check it within milliseconds, add it to their copy of the chain, and the winning miner takes the block reward of 3.125 BTC set at the April 2024 halving, plus every network fee in that block.

The security comes from accumulated work. To rewrite a past transaction, an attacker would have to redo the proof of work for that block and every block after it, then outpace the combined hash rate of the honest network. The deeper a transaction sits in the chain, the more computation it would take to undo, which is why exchanges and custodians wait for several confirmations before treating a deposit as final.

For the wider picture of how Bitcoin is structured, including the unspent transaction output (UTXO) model, the 21 million supply cap, and how blocks link together, see How Bitcoin works.

Key terms about Bitcoin mining

A few terms recur throughout this guide to Bitcoin mining. Each definition below is meant to stand on its own.

Mining hardware: from CPUs to ASICs

Bitcoin mining hardware has run through four clear generations, and each one pushed the previous into obsolescence.

Four hardware generations

Current-generation ASICs

As of June 2026, the dominant ASIC makers are Bitmain (Antminer line), MicroBT (Whatsminer line), and Canaan (Avalon line). A representative current-generation unit, the Bitmain Antminer S21 Pro, is rated at 234 TH/s and about 15 joules per terahash (J/TH) at 3,510 watts, according to Bitmain's published specification. A single current-generation machine runs roughly $5,000 to $15,000 as of June 2026, according to the Hashrate Index ASIC Price Index, depending on model and market conditions, and industrial operators deploy thousands of them in purpose-built facilities.

Three numbers decide whether a machine is worth running: hash rate, or how many hashes it produces; efficiency in J/TH, or how much electricity each hash costs; and the unit's purchase price.

Cooling has become its own arms race. Air-cooled units remain the default, but immersion and hydro cooling let operators run machines harder and quieter, which matters at industrial density where thousands of units share a building.

The hardware arms race

Each new ASIC generation lifts the network's total hash rate, which forces difficulty upward (covered in the next section), which strands older machines below break-even, which feeds demand for the next generation. That loop has run without pause since 2013. A machine bought today generally stays competitive for two to four years before newer hardware prices it out at prevailing electricity rates.

Mining pools: why solo mining is impractical

Bitcoin network  hash rate is hovering near 1,000 EH/s as of June 2026, according to hash rate data from mempool.space (It briefly crossed 1 zettahash per second in January 2026 before winter-storm curtailment in Texas pulled it back.) A single ASIC running at 234 TH/s is a vanishingly small slice of the whole, on the order of 0.00002% of total network power. Left to mine alone, that machine could go years between blocks.

Mining pools exist to smooth out that randomness. A pool combines the hash rate of many miners and splits block rewards among them in proportion to the work each contributed.

How a mining pool works

The pool operator keeps a live connection to the Bitcoin network and builds candidate blocks. Individual miners connect to the pool and receive work assignments, slices of the search space to grind through. Miners send back shares, partial solutions that prove they are working even when they are not full blocks. Shares are the accounting unit the pool uses to measure each miner's contribution. When any member finds a valid block, the pool pays out across all contributors according to their share of the work.

Pool payout methods

Pool concentration and centralization

As of June 2026, a handful of pools, including Foundry USA, AntPool, F2Pool, and SpiderPool, direct the majority of Bitcoin's hash rate, according to onchain pool data from mempool.space. Foundry USA alone has accounted for roughly 25%–34% of blocks across recent reporting windows. Foundry USA is backed by Digital Currency Group, and AntPool is owned by the ASIC maker Bitmain, so a few corporate parents sit behind a large share of block production. That concentration raises a familiar worry: if one pool, or a small group acting together, controlled more than 50% of hash rate, it could, in principle, reorder recent transactions in what is known as a 51% attack. In practice, pools are coalitions of independent miners who can repoint their machines elsewhere in minutes, and the cost of mounting such an attack generally outweighs the payoff.

The industry has also moved to address the narrower question of who chooses the transactions in a block. In May 2026, seven of the largest pools, together representing close to 75% of global hash rate, joined the Stratum V2 working group. Stratum V2 lets individual miners, rather than pool operators, select which transactions go into the blocks they help find.

Difficulty adjustment: Bitcoin's self-regulating clock

Bitcoin aims for one new block every 10 minutes on average. But hash rate never sits still. Machines switch on and off, electricity prices move, and whole regions drop offline for weather or regulation. Without a correction, blocks would arrive faster as miners pile in and slower as they leave.

The protocol corrects for this on a fixed schedule. Every 2,016 blocks, roughly two weeks, the network recalculates the difficulty target from how long the previous 2,016 blocks actually took. If they come in faster than the 20,160-minute target, difficulty rises, and the target hash becomes harder to hit. If they come in slower, difficulty falls. The change is capped at a factor of four in either direction per period, though real adjustments are usually within single-digit percentages.

Why difficulty adjustment matters

This self-regulating feedback loop is what keeps Bitcoin's issuance schedule on track no matter how much mining power shows up. Whether hash rate doubles or halves, blocks still land about every 10 minutes, and the supply schedule, including the roughly four-year halving cycle, holds. For how the halving reshapes miner economics over time, see What is Bitcoin halving.

The economics of Bitcoin mining

Mining profitability reduces to one comparison: revenue per BTC against cost per BTC. The industry tracks this through hashprice, the expected daily revenue per unit of hash rate, which folds the BTC price, network fees, and difficulty into a single figure. When hashprice falls, the least efficient machines are the first to switch off.

Revenue

Two streams make up a miner's revenue. The block reward is the predictable part, 3.125 BTC per block since the April 2024 halving, dropping to 1.5625 BTC at the next halving (block 1,050,000, expected around April 2028). Network fees are the variable part, rising and falling with transaction demand. As of mid-2026, fees have run in the rough range of 0.015–0.03 BTC per block, averaging about 0.02 BTC, based on block fee data from mempool.space, with sharp spikes during periods of heavy demand. As the block reward keeps shrinking with each halving, fees become a steadily larger share of total revenue. 

Costs

Electricity is the dominant operating cost. At $0.05/kWh, a 3,510-watt machine costs about $4.21 a day to run, and industrial operations typically secure rates of $0.03 to $0.06 per kilowatt-hour (kWh), well below retail. Coinshare’s Q1 2026 mining report estimated that electricity makes up the majority of cash costs for publicly listed miners.

Hardware depreciation matters nearly as much. A $10,000 machine with a competitive life of three years costs roughly $9 a day in depreciation alone, assuming little residual value. Facility costs, including cooling, racking, networking, site lease, insurance, security, and staff, vary widely by location and scale. Pool fees typically take another 1%–3% of gross mining revenue, depending on the pool and payout method.

Breakeven

The all-in cost to produce one BTC swings widely by operation. CoinShares' Q1 2026 mining report put the weighted-average cash cost for publicly listed miners at roughly $80,000 per BTC in Q4 2025. Large operators with cheap power and current hardware produce well below that, while private miners paying retail electricity rates and running older machines often sit well above it.

The variables an operation can influence are its electricity rate, hardware efficiency, and scale. The ones it cannot touch are the BTC price, network difficulty, and prevailing fee levels.

Energy consumption and sourcing

Bitcoin mining's electricity use draws steady scrutiny from regulators, journalists, and environmental groups, and the headline figures depend heavily on how they are measured.

Estimates of Bitcoin's annual electricity consumption cluster around 175 TWh for 2025. The Digiconomist Bitcoin Energy Consumption Index put 2025 usage near 175 TWh, with higher-end readings approaching 200 TWh, while the Cambridge Bitcoin Electricity Consumption Index (CBECI) reported figures in a similar range depending on its model assumptions. For scale, that is roughly the annual electricity consumption of a mid-sized country such as Poland.

The energy mix debate

How clean that power is remains contested. The Cambridge Centre for Alternative Finance's April 2025 Digital Mining Industry Report found that 52.4% of the hash rate it sampled ran on non-fossil sources: 42.6% renewables such as hydro, wind, and solar, plus 9.8% nuclear. Industry self-reported figures tend to run higher. The Bitcoin Mining Council has reported sustainable-energy shares above 60% among its members, though that data is voluntary and covers only participants. Because the Cambridge sample captured roughly half of the global hash rate, the unsampled remainder, concentrated in regions like Russia and Central Asia, likely leans more heavily on fossil fuels.

Less contested is the direction of travel. Mining is increasingly sited next to stranded or surplus energy: flared natural gas, curtailed wind and solar, and excess hydro that would otherwise go to waste. MARA Holdings (formerly Marathon Digital), for instance, has built part of its operations around renewable and otherwise curtailed power.

Miners are also unusually flexible electricity consumers. Because a rig can power down in seconds, large operations increasingly sell that flexibility back to grid operators, curtailing during demand spikes and absorbing power that would otherwise be wasted. The Cambridge Digital Mining Industry Report recorded members curtailing hundreds of gigawatt-hours of load to support grids, which reframes mining as a potential balancing resource rather than only a drain.

The efficiency trend

Even as total hash rate climbs, the energy each hash costs keeps falling with every hardware generation. Current machines run at 15–20 J/TH, down from 50+ J/TH about five years ago, so the network buys more security per unit of energy than it used to.

Industrial mining vs home mining

Mining has tilted decisively toward industrial-scale operations for several reasons.

Mining at home is technically possible, but rarely profitable at retail electricity rates in most countries. A single ASIC is loud, around 75–76 decibels, comparable to a vacuum cleaner running nonstop, according to Bitmain's specification. It throws off real heat and needs a dedicated circuit. At a $0.12/kWh retail rate, it would cost roughly $10 a day to run while earning a statistically lumpy fraction of a block reward through a pool. Some hobbyists still mine for reasons beyond profit: acquiring BTC without an exchange or identity check, supporting decentralization, capturing the waste heat, or soaking up spare solar capacity.

How mined BTC reaches a wallet

When a miner or pool earns BTC, the coins are sent to the Bitcoin address specified in the mining software's configuration. From there, miners commonly move BTC to an exchange to sell or to a self-custodial wallet to hold.

For example, MetaMask supports native BTC through SegWit addresses, so anyone receiving BTC, miners included, can hold it directly alongside assets on Ethereum, Solana, and other supported networks, all under a single set of accounts with full Private Key control. For a step-by-step on receiving or buying BTC in self-custody, see how to buy Bitcoin, and for more on wallet types and security, see What is a Bitcoin wallet.