Why do ETH supply numbers differ across sites?
Look up "ETH supply" on three sites and you will get three different numbers. None of them is necessarily wrong — they measure different things, at different times, with different definitions. This page puts the numbers side by side and explains each gap.
Live comparison — all values measured in the same collection run, 2026-07-27T05:33:55Z (UTC)
Burn rate: direct measurement vs ultrasound.money
| Source | Burn (ETH / 24h) | How it is measured |
|---|---|---|
| This site (direct) | 29.5 | Sum of base-fee and blob burn over 7,200 raw block headers |
| ultrasound.money d1 | 29.5 | Their indexer's rolling 24h burn sum |
These two should agree within a few percent — the residual gap is window boundaries (our window ends at our head block, theirs at their indexer head). A persistent large gap would indicate a methodology bug on one side; that is exactly what this cross-check exists to catch.
Supply: consensus-derived vs CoinGecko circulating
| Source | ETH supply | Definition |
|---|---|---|
| Consensus-derived total (ultrasound.money supply-parts) | 121,884,168 | execution balances + beacon balances − processed deposits |
| CoinGecko circulating | 120,682,636 | Aggregator's circulating-supply estimate |
Current gap: 1,201,532 ETH (0.99% of supply).
The three real reasons numbers differ
1. "Circulating" is not "total"
Total supply counts every ETH in existence. Circulating supply subtracts what an aggregator considers non-circulating — and for ETH there is no official foundation-locked allocation schedule like many tokens have, so each aggregator applies its own judgement. That judgement is the gap.
2. Update lag
ETH supply changes every block (issuance up, burn down). A number cached hours ago is already off by the net issuance accrued since. Sites refresh at different cadences; this site stamps every figure with its measurement time so lag is visible instead of hidden.
3. Methodology
Some sites read the execution layer only, some include beacon-chain balances, some track the deposit contract (which only ever grows and therefore overstates stake after withdrawals). Two honest measurements with different system boundaries produce different numbers.