Proof of Work: What It Is and How It Works

Proof of Work is an algorithm that requires participants in a blockchain network to prove their right to add a new block by incurring real computational costs: they try different hash values until they find a suitable one, consuming electricity and hardware capacity in the process. It is precisely this “costliness” of the solution that protects the network from falsifying its history and allows thousands of computers that do not know one another to reach agreement without an intermediary.

What Is Proof of Work in Simple Terms

Proof of Work is a way to confirm that a network participant has actually spent computational resources before gaining the right to record a new block on the blockchain. A computer must solve a problem that takes time and electricity, while the other participants can verify the completed solution almost instantly. This “expensive to solve, cheap to verify” asymmetry is the foundation of the entire mechanism.

This asymmetry is easiest to understand through the everyday analogy of sudoku: solving the puzzle yourself may take an hour of trying different options and crossing things out, but checking someone else’s completed solution takes just a second by verifying the cells against the rules. Proof of Work works in exactly the same way, except that instead of sudoku cells there is a 256-bit hash value, and instead of an hour there are seconds or minutes of intensive trial and error on specialized hardware. The difference in scale is enormous: the space of possible hash values is so vast that it is physically impossible to search all of it, leaving miners to rely on luck while making billions and trillions of attempts per second.

A blockchain needs such a mechanism because in an open network without a central operator, someone must decide which version of the transaction history is correct when several competing versions appear. Proof of Work solves this simply: the valid chain is the one that has consumed the greatest total amount of computational work. Falsifying the history means recalculating all the blocks faster than the entire rest of the network combined—and in practice, this costs more than the falsification could yield.

Before cryptocurrencies, the idea of Proof of Work served a different purpose: combating email spam. In 1997, Adam Back proposed the Hashcash system, in which the sender of an email had to spend a couple of seconds of computing power solving a small hash puzzle. This was imperceptible for a single email but became prohibitively expensive when sending millions of emails at once. In 2008, Satoshi Nakamoto took this very idea and applied it not to emails but to transaction blocks, adding an economic reward for solving the puzzle. In this way, a computational challenge became the foundation of both monetary issuance and network security.

Besides protecting against falsifying the history, Proof of Work has a second function: the fair distribution of new coins. The block reward goes to whoever finds the solution first, rather than to whoever has more money in their account or a higher reputation in the community. This keeps entry to the network open: anyone with hardware and access to electricity can become a miner, without permission or registration with a network operator.

How Mining Works: From a Transaction to a New Block

Several steps take place between the moment a user sends coins and the moment a transaction becomes permanently part of the blockchain: the transaction first waits in the shared pool, then a miner builds a candidate block from the pool, tries different hash values in search of a solution, and, after finding one, broadcasts the completed block to the network, which verifies and accepts it.

A key detail of this process is that the search proceeds almost blindly. Changing the nonce by even one completely changes the result of the hash function, so it is impossible to gradually “approach” the solution or predict it from previous attempts. The only strategy is to try billions and trillions of options per second, which is what the specialized hardware discussed below is designed to do. When the options in the 32-bit nonce field run out without a solution, the miner changes the timestamp or the transaction set and starts searching again.

Thousands of other miners around the world are solving the same problem at the same time, and the winner is whoever finds a suitable hash first. The other participants, whose work was not used in that round, simply switch to assembling the next block. Their computations do not disappear without purpose; they support the network’s overall security: the greater the total computing power trying different options, the harder it is to falsify a block retroactively.

Why Mining Difficulty Is Needed and How It Changes

Difficulty is a parameter that determines how small a block’s hash must be for the solution to be considered valid. The network periodically adjusts it so that blocks continue to appear at roughly the same interval, regardless of how much computing power is currently connected to mining.

Without this mechanism, the time between blocks would fluctuate with the network’s total computing power: if twice as much hardware joined mining over a month, blocks would be found twice as often at the same difficulty, reducing the 10-minute interval to 5 minutes. This would disrupt everything tied to a predictable issuance rate and to the time required for a transaction to be considered sufficiently confirmed. That is why the Bitcoin protocol compares the actual time taken to find every 2016 blocks—approximately two weeks—with the target and recalculates the difficulty in proportion to the discrepancy.

The difference in approaches is clear from the table: networks with long recalculation intervals, such as Bitcoin, respond slowly to changes in hashrate and maintain a stable pace over the long term, but may “slump” for a couple of weeks after a sharp miner exodus. Networks with a sliding window, such as Monero, adjust almost instantly after each block, but their difficulty fluctuates more frequently and noticeably in the short term.

An increase in difficulty is directly linked to an increase in the network’s total computing power — its hashrate. When a major mining pool brings new equipment online, blocks are found more quickly for a short time, and at the next recalculation point the protocol raises the difficulty, returning the pace to its target value. The reverse process also occurs: if mining becomes unprofitable — for example, because of a drop in the coin’s price or higher electricity costs — some equipment is switched off, blocks slow down temporarily, and difficulty decreases at the next recalculation. This is exactly how the Bitcoin network behaved in 2021, when a significant share of the hashrate was suddenly shut down following China’s mining ban: the interval between blocks temporarily almost doubled, and then the largest difficulty drop in the network’s history brought the pace back to normal.

How Proof of Work Differs from Proof of Stake

Proof of Work uses real computing power and electricity to secure the network, while Proof of Stake uses coins locked in the network: instead of searching through hashes, the right to create a block in PoS is won by lottery among those who have provided a stake, and the larger the stake, the higher the chance. PoS consumes thousands of times less energy, while PoW has been tested for longer against attacks under real-world conditions.

The difference appears everywhere — from the entry threshold for a network participant to how each mechanism penalizes dishonest players. In Proof of Work, the penalty is natural: if you spend electricity and equipment resources on an invalid block, you lose those resources permanently; they have already been burned. In Proof of Stake, the penalty is imposed artificially through slashing: if a validator signs conflicting blocks or remains offline longer than allowed, the protocol deducts part of their staked coins.

In terms of security, both mechanisms are resilient in practice for large networks, but they are vulnerable to smaller ones in different ways. Proof of Work can be attacked by renting or purchasing enough computing power — for small PoW coins with low hashrate, this costs tens of thousands of dollars per hour to rent on specialized hashpower marketplaces, and such attacks have occurred in practice (Bitcoin Gold, Ethereum Classic). Proof of Stake must be attacked by buying up or borrowing a large share of the staked coins, which raises their price in liquid networks and makes the attack conspicuous and suicidal for the attacker — their own stake would lose value along with the attacked network.

The entry threshold is also fundamentally different. Becoming a Proof of Work miner means buying or assembling specialized equipment, finding cheap electricity, and often joining a pool to smooth out income — this requires capital investment of at least several thousand dollars for profitable operation. Becoming a Proof of Stake validator means locking up the required amount of coins — a substantial amount for an independent node, but accessible through staking pools and exchanges, where even a small amount can be used to participate while sharing the reward with the pool’s other contributors.

What Equipment Is Needed for Proof-of-Work Mining

Mining Bitcoin and most major PoW coins today requires an ASIC—a specialized chip designed for one specific hash function and incapable of doing anything else; less popular coins using other algorithms still rely on graphics cards (GPUs), while a CPU has been unsuitable for serious mining for more than a decade.

It was not always this way—the equipment used for mining has gone through three evolutionary stages, and each successive one made the previous stage unprofitable.

CPUs and GPUs are general-purpose processors: they can perform any calculations, and part of their transistors sits idle during mining. An ASIC (application-specific integrated circuit) is designed for exactly one operation—searching through hashes for a specific algorithm—and does not waste silicon on anything unnecessary, so it calculates hashes orders of magnitude faster at the same power consumption. The downside of this specialization is that an ASIC for SHA-256 is useless for any other algorithm: if a coin changes its hashing algorithm or mining becomes unprofitable, the device turns into an expensive piece of metal.

GPU mining has not disappeared completely—it remains a viable option for coins whose algorithm was intentionally designed to be “ASIC-resistant” (difficult to optimize through narrow silicon specialization), or for networks where ASICs for the specific algorithm are not yet mass-produced. Such coins use algorithms like RandomX (Monero) or Ethash, on which the GPU mining once used for Ethereum still made sense until the network switched to Proof of Stake in 2022. CPU mining has also survived in a narrow niche—RandomX was specifically designed so that an ordinary processor would remain competitive, while the advantage of an ASIC or GPU would be minimal, in the interests of decentralizing Monero mining.

In practice, when choosing equipment for profitable mining of a major PoW coin today, the question is not whether to use a CPU or GPU, but which ASIC model to buy: their hashrate per watt of energy differs fundamentally, and this ratio—not the rated power—determines actual profitability given the current electricity price and network difficulty.

How Much Energy Does Proof of Work Consume, and Why Is It Criticized?

The Bitcoin network consumes around 120–150 terawatt-hours of electricity per year—comparable to the annual consumption of a country such as Argentina or Norway—and this is why Proof of Work is most often criticized for its environmental footprint, although some of this energy already comes from renewable sources or sources that would otherwise go unused.

The argument against Proof of Work is straightforward: the energy spent searching through hashes produces nothing useful outside the network itself—in contrast to, for example, electricity used to smelt steel, there is no physical product here that can be touched. Critics add that some mining has historically been concentrated in regions with cheap coal, while a rise in the coin’s price directly encourages higher network energy consumption because the larger reward justifies bringing increasingly inefficient equipment online.

Supporters of Proof of Work respond that energy consumption alone is not a measure of harm—it matters where the energy comes from and what would happen to it without mining. Mining equipment can be physically moved to places where electricity would otherwise go to waste: for example, to associated gas from oil wells that would otherwise be flared without any benefit, or to surplus hydropower during flood seasons that cannot be delivered to consumers through the grid in time. From this perspective, mining acts as a “buyer of last resort” for energy that would otherwise have no demand, rather than competing for electricity needed by households.

Nevertheless, a comparison with Proof of Stake honestly favors PoS: after Ethereum switched to Proof of Stake in September 2022, its energy consumption fell by more than 99.9% because validators no longer need to search through hashes—only to keep an ordinary server online. This is the main reason why major projects launched after Bitcoin almost universally choose Proof of Stake or its variations, leaving Proof of Work as a niche for networks that value its time-tested resilience more than energy efficiency.

What Is a 51% Attack, and How Realistic Is It for Proof of Work?

A 51% attack is a situation in which one participant or a colluding group controls more than half of the network’s total computing power and can thereby rewrite the recent history of blocks, for example by canceling its own transactions after receiving goods (double-spending). For large networks such as Bitcoin, such an attack costs more than any potential benefit, while small PoW coins have experienced it repeatedly in practice.

The attack mechanism is based on the rule that “the longest chain with the greatest total work wins.” Without publishing its blocks, the attacker secretly starts calculating its own parallel chain while simultaneously paying for a purchase on the open public chain and receiving goods or coins on an exchange. When the hidden chain becomes longer than the public one, the attacker releases it all at once—the network accepts the longer chain as genuine under the rules, the payment transaction in the old chain is invalidated, and the spent coins return to the attacker as if they had never been spent.

The cost of an attack depends directly on the size of the network. Taking control of more than half of Bitcoin’s hashrate would mean either building ASIC farms worth tens of billions of dollars or renting comparable capacity—but there is nowhere to rent it physically, because such a volume of spare capacity simply does not exist on the market. The situation is the opposite for small coins: their total hashrate is low, and specialized computing-power rental marketplaces such as NiceHash make it possible to “borrow” enough computing power for several hours at a reasonable cost.

Practical protection against an attack is based on waiting for confirmations: the more blocks have been added on top of a transaction, the more expensive it would be for an attacker to recalculate the chain and reverse it, because they would need to outpace not just one block but all the network’s accumulated work since then. That is why major exchanges and sellers of expensive goods wait for not one confirmation but six or more for Bitcoin — after that many blocks, the cost of rewriting history is clearly higher than any reasonable transaction amount. For small coins with low hashrate, even a large number of confirmations does not provide the same guarantee, because the absolute dollar cost of an attack remains low.

Which cryptocurrencies use Proof of Work today

The largest and most resilient PoW network is Bitcoin, with several other notable coins around it: Litecoin, Dogecoin, Monero, Bitcoin Cash, and Ethereum Classic — each with its own hashing algorithm and its own balance between speed, privacy, and ASIC resistance.

The hashing algorithm determines what equipment can be used to mine a coin at all, and this is the key distinction between the networks in the table. SHA-256, used by Bitcoin and Bitcoin Cash, has the most developed ASIC market, so both networks compete for the same equipment and electricity. Scrypt, used by Litecoin and Dogecoin, also moved to ASIC mining long ago, but these chips work only with Scrypt coins, while merged mining allows miners to produce blocks on both networks simultaneously without additional power consumption. Monero’s RandomX is the opposite extreme: the algorithm was specifically designed so that a standard server CPU would be almost as efficient as a specialized chip, keeping mining accessible without major investments in equipment.

Outside the table are dozens of lesser-known PoW coins — from Zcash with the Equihash algorithm to Ravencoin and Kaspa, focused on GPU mining and serving their own niches. The general trend in recent years is that major projects launched from scratch almost never choose Proof of Work, while existing PoW networks remain viable thanks to accumulated network efficiency and reputation — changing the consensus mechanism of a large active coin is technically difficult and risky for community trust, so the overwhelming majority of PoW networks today are older than 2018.

What are the advantages and disadvantages of Proof of Work

The main advantage of Proof of Work is more than 15 years of real-world testing without a single successful breach of a major network, while the main disadvantage is the enormous amount of energy and physical equipment consumed by a process that produces nothing beyond blockchain protection.

Centralization is a disadvantage that was not obvious when the protocol launched. The idea behind Proof of Work assumed that computing power would be distributed among many independent participants, but in practice mining consolidated around several large pools and chip manufacturers: a handful of mining pools regularly control more than half of Bitcoin’s total hashrate, although the participants in each pool are technically independent and can switch to another at any time. This is not the same as a 51% attack by a single actor, but it reduces the network’s safety margin against collusion among several large pools.

In practice, the choice between the advantages and disadvantages of Proof of Work depends on what matters most for a particular application. For a network securing trillions of dollars in value, such as Bitcoin, conservative, time-tested resilience outweighs the energy costs—the community deliberately chooses not to change a mechanism that works. For new projects that do not yet have the same reserve of trust and have the opportunity to design a system from scratch, the energy efficiency and low entry barrier of Proof of Stake usually outweigh the historical track record that PoW has yet to build.

Where to start if you want to try mining

Start not by buying equipment, but by doing the calculations: use an online profitability calculator to determine whether a particular ASIC or graphics card will pay for itself at your electricity rate and the network’s current difficulty, and only if the figures add up should you move on to choosing a pool and buying hardware.

A mining pool is needed by almost everyone except the owners of truly large farms, because the chance of an individual average miner finding a block is extremely small—you could mine solo for years without receiving a single reward. A pool combines the computing power of many participants, finds blocks regularly, and distributes the reward among everyone in proportion to each participant’s contribution, smoothing income into predictable daily or weekly payouts instead of a rare large windfall.

The main financial risk is that equipment does not pay for itself immediately, while network difficulty and the coin’s price may change in either direction during that time. A payback period of six months to a year at the time of purchase can double if difficulty rises faster than expected or the coin’s price falls, while a specialized ASIC cannot be repurposed for anything else—in the worst case, the investment may never pay back at all. For this reason, industry professionals usually calculate not “how much will I earn?” but “how many months will it take to recover my investment under current conditions?” and build in a margin for those conditions to worsen before investing a significant amount.

What the future holds for Proof of Work

Proof of Work will not disappear in the foreseeable future, but it will not become the mainstream choice for new projects either: major networks where the mechanism is already working and has built up trust—above all Bitcoin—will continue using it as is, while the overwhelming majority of new blockchains will continue choosing Proof of Stake or its variants for their energy efficiency and low entry barrier.

Ethereum’s transition to Proof of Stake in 2022 was a turning point: before it, PoW was seen as the default standard for serious projects; afterward, it became a deliberate, more conservative choice. The Bitcoin community has repeatedly and openly discussed a hypothetical transition to another consensus mechanism and rejected the idea each time—a change to a network with a trillion-dollar market capitalization would risk splitting the community and losing the trust built precisely on the immutability of the basic rules.

Regulatory pressure on PoW is increasing primarily through the energy angle: the European Union has discussed requirements to disclose the carbon footprint of crypto-assets, while individual regions—from New York State to Chinese provinces—have introduced direct or indirect restrictions on mining, citing the strain on power grids. At the same time, the opposite trend is also growing: legalizing and attracting mining to places where surplus energy would otherwise find no demand, as is happening in some US states and Gulf countries.

Over a horizon of several years, the most likely scenario is the coexistence of two models: Bitcoin and a handful of networks around it remain on Proof of Work as benchmarks of resilience and predictability, while everything new in the industry—from DeFi protocols to gaming blockchains—is built on Proof of Stake and its modifications, where energy efficiency and speed matter more than a decade and a half of real-world testing.

Frequently asked questions about Proof of Work