EIP-1559 (London) — fee burning

Before London, gas was a blind first-price auction: you guessed a gas price, overpaid or got stuck. EIP-1559 restructured the fee into a protocol-set base fee (that adjusts each block with demand and is burned) plus an optional priority tip to the validator.

fee = base fee  +  priority tip
        │             │
     BURNED      → validator
   (removed from
    supply)      block > 50% full → base fee ↑ next block
                 block < 50% full → base fee ↓ next block

It is not a fee cut

The most persistent myth about 1559: that it made gas *cheaper*. It didn't. When demand is high the base fee still rises — it just rises *predictably*, block by block, instead of via blind bidding. 1559 fixed the price discovery, not the price. Cheap fees were always going to come from scaling (rollups + blobs), not from a fee-market redesign.

block N    > 50% full ──→ base fee ↑ (~12.5%) next block
block N+1  < 50% full ──→ base fee ↓ next block
           → fees converge toward real demand, predictably

Two consequences

UX: wallets can estimate the base fee reliably, so fees became far more predictable. Monetary: burning the base fee permanently removes ETH from supply. When the network is busy, ETH can be net deflationary — the "ultrasound money" meme. After the Merge cut issuance, that effect sharpened.