Hashing algorithms sit at the core of cryptocurrency networks. They define how block data is checked, how transactions are referenced, and what kind of hardware miners need to compete. Bitcoin relies on SHA-256, Litecoin uses scrypt, and Monero moved to RandomX, with each choice producing a different mining model.
A hash function takes input data of almost any size and turns it into a fixed-length output, usually called a hash or digest. That is different from encryption. Encryption is reversible with a key; hashing is generally one-way and is used for data integrity, transaction IDs, block hashes, and proof-of-work.
Why hashing matters across a blockchain
Blockchains need a compact fingerprint for data that is easy to verify and difficult to tamper with. Hashing provides that. A tiny change in a transaction, timestamp, nonce, or Merkle root produces a very different result, a property commonly described as the avalanche effect. Honest nodes can recompute hashes quickly. Attackers, by contrast, cannot realistically work backward from a digest or find two different inputs that produce the same output within practical limits.
This is also what links blocks together. Each block header stores the previous block’s hash. Change an older block and its hash changes too, breaking the chain of references. To make altered history look valid, an attacker would need to redo the proof-of-work for that block and every block after it.
Block hashes, transaction IDs, and Merkle roots
Hashing is used far beyond mining. A block hash is usually derived from the block header. In Bitcoin, the block header is serialized in an 80-byte format and includes fields such as version, previous block hash, Merkle root, timestamp, nBits, and nonce. The resulting hash acts as the block’s compact identifier and, in proof-of-work systems, proves that enough computation was performed.
Transaction IDs are typically hashes of transaction data. Wallets, explorers, and nodes use them to reference a specific transfer without repeating the full payload each time. Merkle roots serve a different job: they commit a block to its complete transaction set while allowing efficient verification that a transaction belongs to that block.
How proof-of-work turns hashing into a competition
In a PoW network, miners repeatedly change the nonce and hash a candidate block header until the result falls below the network’s target threshold. In Bitcoin, miners apply Double SHA-256 to a candidate 80-byte block header. If the hash does not satisfy the target, they modify the nonce or other variable data and try again.
Several terms matter here. The target defines how low a valid hash must be, difficulty expresses how hard it is to find one, and hash rate measures how many hashing attempts hardware can perform per second. Producing a valid block is expensive. Verifying it is not; full nodes simply recompute the hash and check it against the consensus rules.
SHA-256, scrypt, and RandomX lead to different hardware outcomes
SHA-256, used by Bitcoin, produces a 256-bit, or 32-byte, output. Over time, this workload became highly suitable for specialized chips, which is why ASIC hardware came to dominate Bitcoin mining. The source frames this as a practical reason to study hashing in the first place: once the algorithm is understood, the hardware trend becomes easier to explain.
scrypt, Litecoin’s proof-of-work algorithm, also produces a 256-bit output but places more emphasis on memory usage. That gave consumer hardware and GPUs more room in earlier stages and slowed the rise of ASIC dominance compared with Bitcoin, though it did not stop specialized miners from arriving later.
RandomX is the current algorithm used by Monero. According to the source, Monero switched to RandomX in 2019 after earlier CryptoNight-based mining became more vulnerable to ASIC development. RandomX does not make ASICs impossible, but it aligns with Monero’s goal of keeping mining more accessible to consumer hardware, especially CPUs.
Equihash, X11, and Ethereum’s past use of Ethash
Zcash uses Equihash, a memory-oriented proof-of-work algorithm based on the generalized birthday problem. Its parameters are written as Equihash(n, k), and Zcash originally adopted Equihash(200, 9). That made GPU mining practical in the network’s early years, but ASICs for Equihash eventually appeared, and Zcash did not keep changing the algorithm to block them.
Dash uses X11, which chains together 11 hash functions in sequence: BLAKE, BMW, Grøstl, JH, Keccak, Skein, Luffa, CubeHash, SHAvite, SIMD, and ECHO. The output of one function becomes the input of the next. The design aimed to make mining hardware harder to specialize, yet X11 ASICs still arrived and mining on Dash is now generally ASIC-dominated.
Ethereum also appears in the comparison because it historically used Ethash during its proof-of-work period, though it no longer relies on mining. Taken together, these examples show that a hashing algorithm does more than process data. It affects verification, hardware concentration, and the practical structure of a network’s security model.