How Bitcoin Mining Works: A Step-by-Step Technical Breakdown

Bitcoin mining is the process by which new blocks are added to the blockchain, new bitcoins are issued, and the network reaches consensus on the state of all transactions — without any central authority. It is the engine that makes Bitcoin work. But beneath the familiar metaphor of “digital gold mining” lies a precise, deterministic computational process that is worth understanding in detail.

This article walks through the mining process step by step, from the moment a miner begins assembling a candidate block to the moment that block is accepted by the network.

Step 1: Constructing the Block Template

Mining does not begin with hashing. It begins with block template construction — the process of deciding which transactions to include in the next block and assembling the data structures that the miner will hash.

Fetching Transactions from the Mempool

Every Bitcoin full node maintains a mempool (memory pool): a collection of valid, unconfirmed transactions that have been broadcast to the network but not yet included in a block. When a miner begins working on a new block, they select transactions from their mempool to include.

Transaction selection is not random. Miners are economically rational actors, and they select transactions to maximize fee revenue within the block’s size constraints. Since the SegWit upgrade, blocks are limited to 4 million weight units (approximately 1-1.5 MB of raw data, depending on the transaction types). The miner constructs a set of transactions that fits within this limit while maximizing the total fees.

In practice, this means transactions are sorted by fee rate (satoshis per virtual byte, sat/vB). Higher fee-rate transactions are included first. A transaction paying 50 sat/vB will be included before one paying 10 sat/vB. This market-based fee mechanism is how Bitcoin allocates scarce block space without a central authority.

The Coinbase Transaction

Every block must contain exactly one special transaction as its first entry: the coinbase transaction. This transaction has no inputs (it does not spend any existing UTXOs) and creates new bitcoins according to the current block subsidy schedule.

As of 2026, the block subsidy is 3.125 BTC per block (following the April 2024 halving). The coinbase transaction output is the block subsidy plus the sum of all transaction fees in the block. If the block contains 0.5 BTC in total fees, the coinbase output is 3.625 BTC.

The coinbase transaction also contains an arbitrary data field (the coinbase scriptSig) where miners can embed up to 100 bytes of data. Satoshi Nakamoto famously included the headline “The Times 03/Jan/2009 Chancellor on brink of second bailout for banks” in the genesis block’s coinbase. Today, miners use this field for pool identification, the extra nonce (discussed below), and occasionally messages.

Building the Merkle Tree

Once the miner has selected transactions, they construct a Merkle tree — a binary hash tree where every leaf node is the hash of a transaction, and every non-leaf node is the hash of its two children. The root of this tree, the Merkle root, is a single 32-byte hash that cryptographically commits to every transaction in the block.

The Merkle tree construction proceeds as follows:

1. Hash each transaction (using SHA-256d, i.e., double SHA-256) to get the transaction IDs (txids)
2. Pair the txids. If there is an odd number, duplicate the last one
3. Hash each pair together to produce the next level
4. Repeat until only one hash remains — the Merkle root

The Merkle root is critical because it allows anyone to verify that a specific transaction is included in the block without downloading the entire block (this is called a Merkle proof or SPV proof). It also means that changing even a single byte in any transaction would produce a completely different Merkle root.

Assembling the Block Header

The miner now assembles the 80-byte block header from six fields:

Field	Size	Value
Version	4 bytes	Block version (currently 0x20000000 plus any BIP9 signaling bits)
Previous Block Hash	32 bytes	The SHA-256d hash of the most recent valid block header
Merkle Root	32 bytes	The Merkle root computed from the selected transactions
Timestamp	4 bytes	Current Unix time (with some flexibility — must be within a 2-hour window)
Bits (Target)	4 bytes	Compact encoding of the current difficulty target
Nonce	4 bytes	Initially set to 0

This 80-byte header is the data that the miner will hash repeatedly.

Step 2: The Hashing Process

SHA-256 Double Hashing

Bitcoin’s proof-of-work uses SHA-256d: the SHA-256 hash function applied twice in succession. Given the 80-byte block header as input:

hash = SHA-256(SHA-256(block_header))

SHA-256 produces a 256-bit (32-byte) output. This output is a number between 0 and 2^256 - 1. The key property of SHA-256 that makes proof-of-work possible is that it is a one-way function: given an output, there is no way to compute the input except by trying inputs and checking. Small changes to the input produce completely unpredictable changes to the output (the avalanche effect).

The Difficulty Target

The miner’s goal is to find a block header whose SHA-256d hash, when interpreted as a 256-bit number, is less than the current difficulty target. The target is encoded in the header’s “Bits” field using a compact representation.

Conceptually, the target defines how many leading zeros the hash must have. A lower target means more leading zeros are required, which means a smaller number, which means the hash is harder to find. When the network says “difficulty has increased,” it means the target has been lowered, requiring more computational work to find a valid hash.

At the time of writing, the difficulty target requires roughly 79-80 leading zero bits, meaning a valid hash must be a number smaller than approximately 2^176. The odds of any single hash attempt satisfying this condition are approximately 1 in 2^80 — about 1 in 10^24.

Iterating the Nonce

The miner’s search process is straightforward:

1. Set the nonce field to 0
2. Compute SHA-256d of the 80-byte header
3. Compare the result to the target
4. If the hash is below the target: valid block found
5. If not: increment the nonce by 1 and repeat

The nonce is a 32-bit field, providing 2^32 (approximately 4.3 billion) possible values. At modern ASIC hash rates, a single miner exhausts the entire 32-bit nonce space in under one second. This is why additional sources of variation are needed.

The Extra Nonce and Other Search Dimensions

When the 4.3 billion nonce values are exhausted without finding a valid hash, the miner must change something else in the block header to create a new set of hashes to try. The primary mechanism is the extra nonce:

Extra nonce: A field within the coinbase transaction’s scriptSig. Incrementing the extra nonce changes the coinbase transaction, which changes its txid, which changes the Merkle root, which changes the block header — giving a completely new set of 2^32 nonce values to try.

Timestamp: The miner can adjust the timestamp within the allowed range (the current time plus or minus a window), producing different headers.

Transaction selection: The miner can add, remove, or reorder transactions, changing the Merkle root.

In practice, modern mining operates in a three-dimensional search space: nonce (inner loop, exhausted in milliseconds), extra nonce (middle loop), and timestamp/transaction set (outer loop).

Step 3: Finding a Valid Block

When a miner finally computes a hash that falls below the target, they have found a valid block. This hash serves as the proof of work — evidence that a specific, quantifiable amount of computational effort was expended. Anyone can verify this proof instantly by hashing the block header once and checking that the result is below the target. Verification takes microseconds; finding the proof took billions of attempts.

This asymmetry — hard to produce, trivial to verify — is the fundamental property that makes proof-of-work useful for consensus.

Step 4: Block Propagation

Once a valid block is found, the miner broadcasts it to the network as fast as possible. Speed matters: if another miner finds a valid block at the same height first and that block propagates more widely, the slower miner’s block becomes an orphan (or more precisely, a stale block), and they receive no reward.

Compact Block Relay (BIP 152)

Modern Bitcoin nodes use compact block relay to minimize propagation time. Instead of transmitting the full block (which can be over 1 MB), the node sends:

1. The block header
2. A list of short transaction identifiers (6 bytes each instead of 32-byte txids)

The receiving node matches these short identifiers against transactions already in its mempool. Since most transactions in the new block were already known to the receiving node, only a few missing transactions need to be explicitly requested. This reduces block propagation time from seconds to milliseconds in most cases.

Block Validation

Every node that receives the block independently validates it:

1. Header validation: Is the hash below the target? Is the previous block hash correct? Is the timestamp within the acceptable range?
2. Transaction validation: Is every transaction valid? Are the inputs unspent? Are the signatures correct? Do the outputs not exceed the inputs (plus the block subsidy)?
3. Merkle root verification: Does the computed Merkle root match the one in the header?
4. Block size/weight: Is the block within the 4-million-weight-unit limit?

Only after passing all checks does a node add the block to its copy of the blockchain and relay it to its peers. This independent validation by every node is what makes Bitcoin trustless — no node needs to trust the miner or any other node.

Step 5: Difficulty Adjustment

Bitcoin’s difficulty adjustment algorithm ensures that blocks are produced, on average, every 10 minutes, regardless of how much hash power is on the network. This adjustment occurs every 2,016 blocks (approximately every two weeks).

The algorithm is straightforward:

new_target = old_target × (time_to_mine_2016_blocks / 20160_minutes)

If the last 2,016 blocks were mined in 10 days instead of 14, blocks were coming too fast, and the target is lowered (difficulty increases) by a factor of 14/10 = 1.4. If they took 20 days, the target is raised (difficulty decreases) by a factor of 14/20 = 0.7. The adjustment is capped at a factor of 4 in either direction to prevent extreme changes.

This self-regulating mechanism means that Bitcoin’s issuance schedule is predictable regardless of the price of bitcoin, the number of miners, or the sophistication of mining hardware. Hash rate can double, and after one difficulty adjustment, blocks will again arrive every 10 minutes.

Mining Pools: Collective Hash Power

The probability of a solo miner finding a block is proportional to their share of the total network hash rate. With the network hash rate exceeding 700 EH/s (exahashes per second) in 2026, even a miner operating thousands of ASIC machines represents a tiny fraction of the network. A miner with 1 PH/s (petahash per second) has approximately a 1 in 700,000 chance of mining each block, or roughly one block per 13.5 years on average.

This variance problem led to the creation of mining pools, where many miners combine their hash power and share rewards proportionally.

How Pools Work

1. The pool operator provides a block template to all pool members
2. Each miner works on a modified version of this template (with their own extra nonce range)
3. Miners submit shares — proof-of-work solutions that meet a lower difficulty target than the actual Bitcoin difficulty. These shares prove that the miner is performing work
4. If any miner’s hash also meets the actual Bitcoin difficulty target, the pool broadcasts the valid block
5. The block reward is distributed among all pool members proportionally to the shares they contributed

Payout Schemes

PPS (Pay Per Share): The pool pays a fixed amount for each valid share, regardless of whether the pool actually finds a block. This eliminates variance for miners but transfers risk to the pool operator.

PPLNS (Pay Per Last N Shares): The pool pays proportionally based on shares contributed during a window around when a block is found. This distributes risk more evenly but introduces some variance.

FPPS (Full Pay Per Share): Like PPS but also includes an estimated share of transaction fees, not just the block subsidy.

Centralization Concerns

As of 2026, the top 4 mining pools control over 60% of the network hash rate. This concentration raises concerns about censorship and potential 51% attacks. However, it is important to note that pool miners can and do switch pools — the hash power is distributed among individual miners who voluntarily point their machines at a particular pool’s server. If a pool acts maliciously, miners can migrate to other pools within minutes.

The development of Stratum V2, a new mining protocol, addresses some centralization concerns by allowing individual miners to construct their own block templates rather than accepting the pool operator’s template. This prevents pool operators from censoring specific transactions.

ASIC Evolution: The Hardware Arms Race

Bitcoin mining hardware has evolved through four generations:

CPU Mining (2009-2010)

Satoshi Nakamoto mined the first Bitcoin blocks on a standard CPU. Bitcoin’s early blocks were mined at hash rates measured in megahashes per second (MH/s). An Intel Core i7 of that era could achieve roughly 10-30 MH/s.

GPU Mining (2010-2013)

Graphics processing units (GPUs) offered 10-100x improvement over CPUs for SHA-256 hashing due to their parallel processing architecture. A high-end GPU could achieve 200-800 MH/s. GPU mining made CPU mining economically obsolete virtually overnight.

FPGA Mining (2011-2013)

Field-Programmable Gate Arrays (FPGAs) offered better energy efficiency than GPUs, achieving similar hash rates with significantly less power consumption. However, FPGAs were expensive and difficult to program, and their dominance was brief.

ASIC Mining (2013-Present)

Application-Specific Integrated Circuits (ASICs) are chips designed exclusively for SHA-256 hashing. The first Bitcoin ASICs appeared in 2013, and they immediately rendered all previous mining hardware obsolete.

Modern ASICs (2026) achieve hash rates of 200-300 TH/s (terahashes per second) per unit while consuming 15-25 watts per terahash. Key milestones:

• Bitmain Antminer S9 (2016): 14 TH/s, 98 W/TH — defined an era
• Bitmain Antminer S19 XP (2022): 140 TH/s, 21.5 W/TH
• Bitmain Antminer S21 (2024): 200 TH/s, 17.5 W/TH
• Latest generation (2026): 250-300+ TH/s with sub-15 W/TH efficiency

The ASIC arms race has driven two fundamental trends: increasing hash rate (and thus network security) and decreasing energy consumption per hash. The result is that Bitcoin mining has become a highly specialized industrial operation — the era of profitable home mining with commodity hardware ended years ago.

Energy and Economics

Mining is fundamentally an energy conversion process: electricity is converted into cryptographic security for the Bitcoin network. A miner is profitable when the value of bitcoin earned (block subsidy plus fees) exceeds the cost of electricity, hardware depreciation, cooling, and operational overhead.

The break-even electricity price varies by ASIC generation and Bitcoin price, but as a rough guideline in 2026: with a modern ASIC achieving 17 W/TH and Bitcoin at $80,000, mining is profitable at electricity costs below approximately $0.08-0.10 per kWh. This economic pressure drives miners to seek the cheapest electricity on Earth, which increasingly means stranded energy (energy that would otherwise be wasted) and renewable energy (which is often the cheapest per kWh at scale).

The Difficulty-Price Feedback Loop

When Bitcoin’s price rises, mining becomes more profitable, attracting more hash power, which increases difficulty, which reduces profitability per unit of hash power. When the price falls, unprofitable miners shut down, difficulty decreases, and remaining miners become more profitable. This feedback loop maintains a dynamic equilibrium where mining is always marginally profitable for the most efficient operators.

Why Proof of Work Matters

Proof of work is often criticized as “wasteful.” This criticism misunderstands the function of the energy expenditure. The energy consumed by mining is not wasted — it is converted into an unforgeable proof that a specific amount of physical work was performed. This proof has three critical properties:

1. Objective cost: Creating a valid block requires provably expending real-world resources (electricity). This objective cost prevents Sybil attacks — an attacker cannot cheaply create fake blocks.

2. Thermodynamic finality: As blocks are buried under subsequent blocks, reversing them requires re-expending all the energy used to create the subsequent blocks. After six confirmations, reversing a transaction would require more energy than a small country consumes in a year.

3. Fair issuance: New bitcoins are distributed to those who expend real resources to secure the network, creating a direct link between physical work and monetary issuance. This is analogous to how gold mining distributes new gold — through expenditure of energy, not through political decree.

Proof of work is the mechanism that anchors digital money to physical reality. Without it, Bitcoin would be just another database that requires trust in its operators.

For a deeper understanding of block structure, see Bitcoin Block Structure. For the economic implications of mining, see Bitcoin Halving and Bitcoin and Energy.