Proof of work: what the miner pays for and what the network buys with that money

Proof of work is the rule by which a network with no central authority decides which version of the transaction history counts as the real one: the right to append the next block goes to whoever presents the result of a deliberately expensive computation. What is being proven is not the participant's identity, nor their honesty, nor the validity of the transactions. What is proven is that a resource has been spent. The contents of a block are checked separately, against the rules of the protocol; the work answers only the question "who gets to speak", and it makes that answer costly.

What problem the mechanism actually solves

A distributed network faces a problem that a bank does not have: nodes receive messages in different orders, some nodes lie, and there is no referee. If the right to write were granted on the basis of one vote per address, an attacker would simply create as many addresses as they liked. This is known as a Sybil attack. Proof of work ties the vote to something that cannot be multiplied for free: computation that has actually been performed. The idea predates bitcoin. The Hashcash scheme was proposed in 1997 as a fee for sending an email, so that spam would become more expensive than ordinary correspondence. In 2008 the idea was applied to keeping a ledger of transactions, and in 2009 the network went live.

The mechanics: brute force, the hash and the target

A block header contains a reference to the previous block, a digest of the transactions included in the block, a timestamp and a utility field called the nonce. All of this is fed through a cryptographic hash function, which produces an output of fixed length that cannot be predicted from the input and cannot be reversed. The miner changes the nonce and recomputes the hash until the result falls below the target value set by the network.

The whole point lies in the asymmetry. A suitable hash can only be found by brute force, which means by burning electricity. Verifying a hash once it has been found takes a single recomputation, instantaneous for any node. The work is expensive to produce and almost free to check, and that is precisely why the network can be open to anyone without drowning in junk.

Difficulty and the pace of block production

The target value is not fixed. After a number of blocks defined by the protocol, the network looks at how long it took to find them and recalculates the difficulty so that the average interval between blocks stays the same. When more hardware comes online, the puzzle becomes harder; when computing power leaves, it becomes easier. This leads to an important conclusion for understanding the economics: an increase in the network's total computing power does not speed up the issuance of coins. It merely reshuffles market share among miners and raises everyone's costs.

Why rewriting history is expensive

To reverse a record that has already been confirmed, an attacker has to build an alternative chain and redo all the work for the section being reversed, and do so faster than the rest of the network keeps extending the honest chain. This requires control of the majority of the network's total computing power, held for the entire duration of the attack. The protection here is economic rather than mathematical: the protocol does not forbid rewriting history, it makes rewriting unprofitable compared with mining honestly. Hence the practical rule: a record becomes more valuable as more blocks are built on top of it, not at the moment it is included in a block.

The miner pays for this security, with hardware, with electricity and with what that electricity costs, and in return receives the block reward and transaction fees. The block reward is cut on a schedule hard-coded into the protocol (this is what is called the halving), so over time fees make up a growing share of a miner's revenue, while the industry's margin of safety depends on the coin's price and on energy tariffs.

How it differs from proof of stake

Under proof of stake, the right to sign a block is distributed among those who have locked up their own coins as collateral, and part of that collateral is confiscated for breaking the rules. An external cost is replaced by an internal one: not burnt electricity, but the risk of losing capital inside the system itself. Ethereum switched to this model in 2022. The argument over which guarantee is stronger cannot be settled by arithmetic. With proof of work the cost is real and sunk, but it demands constant expenditure; with proof of stake the expenditure is minimal, but security rests on the value of the very asset the system secures.

What the site's data covers on this topic

A plain caveat is needed here: the TradeAlmanac database contains no crypto-asset quotes, no profiles of mining companies and no network difficulty figures, so this article gives no measurable values for proof of work, and inventing them in place of a directive is not allowed. The site covers the Russian market: stocks, government bonds, funds, issuers' financial reports and the news feed. Definitions of the terms used in this text are collected in the glossary.