Proof of History is Solana’s cryptographic clock: a continuously running chain of SHA-256 hashes that stamps every event with a verifiable position in time before validators ever vote on it. Each hash takes the previous hash as its input, so the chain can only be built one step at a time — and the number of steps between two points proves that real time passed between them. This is not a consensus mechanism and it does not replace proof of stake. It is a shared sense of ordering that lets validators skip the slow part of every other blockchain: arguing about what happened first.
Key Facts
- Anatoly Yakovenko published the Proof of History whitepaper in November 2017, three years before Solana mainnet beta went live.
- The chain runs on plain SHA-256. Under default parameters a slot contains 64 ticks of 12,500 hashes each — roughly 800,000 hashes per 400 ms slot.
- Solana targets a 400 ms slot; 432,000 slots make one epoch, which lands close to two days.
- Generating the chain is strictly sequential on one core. Verifying it is parallel — a GPU can check thousands of segments at once.
- The leader schedule for an entire epoch is known in advance, so nodes forward transactions to the next leader instead of broadcasting blindly.
- Each leader holds four consecutive slots — about 1.6 seconds of block production — before the slot rotates.
- PoH orders events. Tower BFT and staked validators decide which fork is canonical. Two separate jobs.
Why Time Is the Hard Part of a Blockchain
Nodes in a distributed network cannot trust each other’s clocks. A machine in Frankfurt and a machine in Singapore will disagree about the exact moment a transaction arrived, and any node can lie about its timestamp to get its own transaction placed first. Classic blockchains solve this by making ordering itself the thing that gets voted on: validators exchange messages, converge on a shared view, and only then append a block. Every round of that conversation costs latency, which is why block times are usually measured in seconds rather than milliseconds. Anatoly Yakovenko’s original insight was that ordering does not need a vote if it can be proven mathematically instead.
How the Hash Chain Actually Works
A node starts with an arbitrary value and hashes it. Then it hashes the result. Then it hashes that result, forever. Because SHA-256 output cannot be predicted without running the function, there is no shortcut: producing the ten-millionth hash requires performing all 9,999,999 hashes before it. That property turns the counter into a clock. When a transaction arrives, its data is mixed into the chain at the current position, permanently pinning it between the hash before and the hash after.
The asymmetry is what makes the design practical. One CPU core grinds the sequence out step by step, but a verifier can split the finished chain into segments and check them all simultaneously across many cores. Producing the proof is slow and serial; checking it is fast and parallel. Cryptographers call this shape a verifiable delay function, though PoH is a simplified cousin rather than a textbook VDF — it proves elapsed work without a compact proof of correctness.
| Property | Proof of History | Proof of Stake (Tower BFT) |
|---|---|---|
| Question it answers | What happened first? | Which fork is real? |
| Input | SHA-256 hash chain | Validator votes weighted by stake |
| Requires communication | No | Yes |
| Can it be faked | Not without redoing the hashes | Not without controlling stake |
| Cost to run | One CPU core, continuously | Staked SOL plus vote fees |
What PoH Buys the Network
Because the clock exists before consensus, the leader schedule can be computed for a whole epoch ahead of time. Every node knows which validator produces blocks in which slot, so transactions get routed straight to that leader instead of gossiped to the entire network and back. Blocks also stream out as they are built rather than arriving as one finished lump — a leader broadcasts entries continuously across its four slots while the validators that run the network verify them in flight. That pipeline is the reason Solana measures block time in hundreds of milliseconds. In practice slots often run a little past the 400 ms target, and you can watch the drift yourself on Solscan, where slot height climbs visibly faster than block height whenever a leader misses.
What Proof of History Is Not
Three misreadings show up constantly. PoH is not a consensus algorithm — it cannot tell you which of two competing forks the network accepts, and Solana still needs staked validators voting for that. PoH is not a security mechanism against reordering by the leader either: the current leader can still choose which transactions to include and in what sequence within its own slot, which is exactly the gap MEV bots trade inside. And PoH is not mining. The hash chain runs on a single core at negligible power draw, so comparing it to Bitcoin’s energy budget misses the point entirely.
The role of the clock is also narrowing. Alpenglow, the consensus overhaul approved by validator governance in 2025, moves voting off-chain and replaces Tower BFT, which reduces how much of the protocol depends on PoH as a timing source. The hash chain stays as an ordering primitive; the parts of it wired into consensus are being rebuilt. Second client work such as Firedancer is being written against that same future rather than the old design.
Why the Clock Matters When You Buy $DOLAN
DOLAN Duck ($DOLAN) is an SPL token on Solana with a fixed 98.3M supply, roughly 10,700 holders and no presale allocation — a fair launch where every entry, including the first, had to queue through the same slot machinery as everybody else. That is the practical meaning of Proof of History for a memecoin buyer: your swap is not accepted because you clicked first, it is accepted because it landed in a leader’s slot and got pinned into the hash chain at a specific position. When a token trends and thousands of buys hit the same 400 ms window, the ones that fail are the ones that never made it into a slot before the blockhash expired. That is also why Solana became the default chain for memecoins rather than a chain with twelve-second blocks — and why routing through an aggregator like Jupiter, which builds transactions with slot timing in mind, beats firing swaps by hand.
Proof of History is a chain of SHA-256 hashes where each hash feeds the next. Because the sequence cannot be skipped ahead, the number of hashes between two events proves that time passed between them. Solana uses it to order transactions before validators vote.
No. Proof of History only establishes ordering. Solana still runs proof of stake, where validators weighted by staked SOL decide which fork is canonical. PoH removes the need to negotiate timestamps, not the need for consensus.
Solana is the only major network using it as designed. Proof of History was introduced in Anatoly Yakovenko’s 2017 whitepaper and remains specific to Solana and chains forked from its codebase.
Generating the chain is sequential and cannot be parallelised, but verifying it can be split across many cores at once. That asymmetry is the whole point: proofs are expensive to fake and cheap to check.
No. The hash chain runs on a single CPU core with negligible power draw. There is no competitive hashing race and no block reward for producing hashes, so the comparison to Bitcoin mining does not hold.
Solana targets 400 milliseconds per slot, with 432,000 slots per epoch — close to two days. Real slots frequently run slightly longer when a leader is slow or skips its turn.
Not fully. The clock fixes the order of slots, but the leader still chooses the order of transactions inside its own slot. That discretion is what sandwich bots exploit, and PoH was never designed to prevent it.