How Bitcoin Actually Works: A Technical Foundation

Why the Mechanics Matter

Most people who own bitcoin could not explain how a transaction works. This is not an insult — most people who use email could not explain SMTP either. But Bitcoin is different from email in one critical respect: it asks you to be your own bank. And if you are going to be your own bank, you should understand the vault, the ledger, and the locks.

What follows is a technical explanation written for intelligent readers who are not engineers. We will not simplify to the point of inaccuracy, but we will also not drown you in jargon. The goal is comprehension — the kind that lets you make real decisions about your own financial infrastructure, not the kind that impresses people at parties.

We draw primarily from Satoshi Nakamoto’s 2008 whitepaper, which remains the most concise description of the system, supplemented by Saifedean Ammous’s treatment in The Bitcoin Standard for broader context.

Transactions: The Fundamental Unit

Everything in Bitcoin begins with a transaction. Forget the blockchain for a moment; forget mining. A Bitcoin transaction is simply a signed message that says: “I am transferring control of this amount of bitcoin from this address to that address.”

To understand how this works, you need to understand a concept called a UTXO — an Unspent Transaction Output. Think of UTXOs like individual bills in a wallet. If someone sends you 0.5 BTC, you now have a UTXO worth 0.5 BTC sitting at your address. If someone else sends you 0.3 BTC, you have another UTXO worth 0.3 BTC. Your “balance” is not a single number in an account somewhere; it is the sum of all the UTXOs controlled by your private keys.

When you send bitcoin, you are spending one or more UTXOs as inputs and creating new UTXOs as outputs. Suppose you want to send 0.6 BTC to a friend. You take your 0.5 BTC UTXO and your 0.3 BTC UTXO as inputs — totaling 0.8 BTC — and create two outputs: 0.6 BTC to your friend’s address, and roughly 0.2 BTC back to yourself as change. The small difference between inputs and outputs becomes the transaction fee, paid to the miner who includes your transaction in a block.

Every input in a transaction must reference a specific previous output, and it must be signed with the private key that controls that output. This is how ownership is proven — not by an account balance in a database, but by cryptographic proof that you control the keys to a specific UTXO. The system does not know your name, your account number, or your credit score. It knows only that a valid signature was provided.

The Blockchain: A Linked Ledger

Now we need a way to record these transactions so that everyone agrees on who owns what. This is the blockchain — and despite the breathless hype that has attached itself to the word, the concept is straightforward.

A blockchain is a linked list of blocks, where each block contains a set of transactions and a reference to the previous block. That reference is a cryptographic hash — a fixed-length string of characters produced by running the previous block’s data through a hash function (Bitcoin uses SHA-256). Change even a single bit of the previous block’s data and the hash changes completely. This means that each block is cryptographically chained to the one before it, all the way back to the genesis block mined on January 3, 2009.

The practical consequence is immutability. If someone wanted to alter a transaction buried 100 blocks deep, they would need to recompute the hash for that block and every subsequent block — all while the rest of the network continues adding new blocks to the chain. As we will see, the energy required to compute these hashes makes this effectively impossible.

Each block has a maximum size of roughly 1 megabyte (with some nuance introduced by the SegWit upgrade in 2017, which effectively allows up to about 4 megabytes of data per block when accounting for witness data). This limits the number of transactions that can fit in each block, which is one reason Bitcoin’s base layer throughput is constrained.

A new block is added approximately every ten minutes. This cadence is not accidental; it is enforced by the system’s most elegant mechanism.

Proof-of-Work: The Engine of Consensus

Here is the central problem Bitcoin solves: how do you get thousands of computers around the world to agree on the state of a shared ledger without trusting any single one of them? This is the Byzantine Generals Problem, and Nakamoto’s solution was proof-of-work.

When a new block of transactions needs to be added to the chain, miners compete to find a specific number. More precisely, they take the block’s data — the transactions, the reference to the previous block, a timestamp, and a variable number called a nonce — and run it through the SHA-256 hash function. The goal is to find a nonce that produces a hash below a certain target value. Because hash functions are one-way (you cannot work backward from a desired output to find the input), the only way to find a valid nonce is brute force: try billions of possibilities until one works.

This process is deliberately expensive. It requires real electricity, real hardware, real time. When a miner finds a valid hash, they broadcast the block to the network. Other nodes verify that the hash is valid, that all transactions in the block are legitimate, and that the block follows all consensus rules. If everything checks out, the block is accepted and the miner receives the block reward — currently 3.125 BTC — plus all the transaction fees in the block.

The beauty of this system is that cheating is more expensive than playing by the rules. To alter the blockchain, an attacker would need to control more than 50% of the network’s total computing power and sustain that control long enough to rewrite the chain faster than honest miners extend it. The cost of mounting such an attack on the current network runs into the billions of dollars with no guarantee of success.

The Difficulty Adjustment

Proof-of-work alone is not enough. If more miners join the network, blocks would be found faster. If miners leave, blocks would slow down. Bitcoin needs a stable block time for its monetary policy to function, and it achieves this through the difficulty adjustment.

Every 2,016 blocks — roughly every two weeks — the network recalculates the difficulty target. If blocks have been coming in faster than one every ten minutes, the target is lowered, making valid hashes harder to find. If blocks have been coming in slower, the target is raised. This self-correcting mechanism ensures that regardless of how much computing power is directed at the network, a new block will be found approximately every ten minutes.

This is worth sitting with for a moment. The difficulty adjustment means that adding more mining power to the network does not produce more bitcoin. It only makes the puzzle harder. This is fundamentally different from, say, gold mining, where deploying more resources generally yields more gold. In Bitcoin, the issuance rate is fixed by the protocol. More miners means more security, not more supply.

Ammous describes this as one of Bitcoin’s most underappreciated properties. It severs the link between production effort and supply increase — a link that has undermined every other commodity money in history. When the price of gold rises, miners extract more of it, eventually increasing supply and dampening the price. Bitcoin’s supply is indifferent to demand.

The 21 Million Cap

This brings us to the supply schedule. The Bitcoin protocol specifies that only 21 million bitcoin will ever exist. This is enforced not by a promise, a contract, or a government decree, but by the consensus rules that every node on the network independently validates.

When Bitcoin launched in 2009, the block reward was 50 BTC. Every 210,000 blocks — approximately every four years — this reward is cut in half. After the first halving, it became 25 BTC. Then 12.5. Then 6.25. After the April 2024 halving, it stands at 3.125 BTC. This halving process continues until the reward becomes so small it rounds to zero, which will occur around the year 2140.

The math is elegant. The sum of a geometric series starting at 50 and halving indefinitely: 50 + 25 + 12.5 + … converges to 100. Multiply by 210,000 blocks per era and you get 21 million. Every node on the network verifies that each new block follows this schedule. Any block that claims a larger reward than allowed is rejected by every honest node. No vote is taken. No amendment is proposed. The rule simply is.

By approximately 2035, over 99% of all bitcoin that will ever exist will have been mined. The remaining fraction trickles out over the following century in increasingly tiny increments.

Nodes vs. Miners: A Critical Distinction

There is a common misconception that miners control Bitcoin. They do not. Miners and nodes perform different functions, and understanding the distinction is essential.

Miners propose new blocks. They gather unconfirmed transactions from the mempool (a kind of waiting room for transactions that have been broadcast but not yet included in a block), assemble them into a candidate block, and expend energy searching for a valid proof-of-work. When they find one, they broadcast the block.

Nodes validate everything. A full node is a computer running the Bitcoin software that independently verifies every block and every transaction against the consensus rules. Does this transaction spend a valid UTXO? Does the block reward match the current halving schedule? Is the proof-of-work valid? Does the block reference the correct previous block?

If a miner produces a block that violates any consensus rule, nodes reject it. The miner has wasted energy and earns nothing. This is why nodes are the true enforcers of Bitcoin’s rules. Anyone can run a node — the software is free, and the hardware requirements are modest (a basic computer with a few hundred gigabytes of storage). When people say “don’t trust, verify,” they are talking about running a node.

The relationship between miners and nodes is sometimes described as a division of powers. Miners propose; nodes dispose. Neither controls the network alone. The result is a system of checks that has no CEO, no board of directors, and no single point of failure.

Honest Limitations

We committed in the first article of this series to honest assessment, and that applies to the technical layer as well.

Throughput. Bitcoin’s base layer can process approximately seven transactions per second. This is not a bug being fixed in the next update; it is a fundamental design choice. The small block size and ten-minute block time are what make decentralization possible — a larger block would require more bandwidth, more storage, and more processing power to run a node, which would centralize the network over time. The trade-off is real: base-layer Bitcoin cannot serve as a global payment rail for everyday transactions. Layer 2 solutions like the Lightning Network attempt to address this by handling high-frequency, small-value transactions off-chain and settling them on the base layer periodically.

Energy consumption.Proof-of-work consumes significant electricity — comparable to the annual consumption of some mid-sized countries . This is not waste in the way critics suggest; the energy expenditure is what secures the network. But it is a genuine cost, and dismissing it as irrelevant does not serve an honest analysis. The question is not whether Bitcoin uses energy, but whether the security it purchases with that energy is worth the cost. Reasonable people disagree.

Finality time. A single confirmation (one block) takes an average of ten minutes, but convention holds that six confirmations — roughly an hour — provides sufficient security against double-spend attacks for large transactions. For someone accustomed to instant payment confirmations, this is a meaningful friction. Again, Layer 2 solutions partially address this for smaller transactions.

Complexity of self-custody. The UTXO model, the importance of fee estimation, the risks of address reuse — these are not intuitive for most users. The protocol is robust, but the user experience remains a frontier. Mistakes can be irreversible and costly.

What You Now Understand

If you have followed to this point, you understand more about Bitcoin than the vast majority of people who hold it. You know that Bitcoin is not a database with account balances but a chain of signed transactions referencing unspent outputs. You know that the blockchain is a linked sequence of blocks secured by computational work. You know that the difficulty adjustment maintains the ten-minute heartbeat regardless of how much mining power enters or leaves the network. You know that 21 million is not a marketing number but a mathematical certainty enforced by code and consensus. And you know that nodes — not miners — are the ultimate arbiters of the protocol’s rules.

This is the foundation. In the next article, we will examine Bitcoin’s monetary policy in detail: the halving schedule, its implications for security, and how it compares to the monetary policy of the institutions it was designed to exist alongside. Understanding the protocol is the first step. Understanding its economics is the next.