EIP-1559 explained for non-developers — and why the burn matters
EIP-1559 replaced Ethereum's fee auction with a base fee and a burn. Here is the spec in plain English, the numbers since 2021, and why it changed ETH's supply curve.
At 12:33:42 UTC on 5 August 2021, Ethereum block 12,965,000 activated the London hard fork and with it EIP-1559, the most significant change to the network’s economics since genesis. From that block onward, the fee market stopped being a sealed-bid first-price auction and became a two-part mechanism with an algorithmic base fee that the protocol burns. As of 1 March 2026, cumulative ETH burned since activation sits at 4.51 million ETH, worth roughly $11.3 billion at current prices. The supply curve of Ethereum was permanently reshaped that day, and a lot of people still describe the mechanism wrong.
What is at stake here is two things at once. First, the user experience: before 1559, wallets had to guess a single gas price into an opaque auction and routinely overpaid by 20-50%. Second, the monetary policy: by burning the base fee, the protocol creates a sink that offsets new ETH issuance, which under proof-of-stake has flipped the network into net-negative issuance during periods of high activity. Understanding both layers — the auction reform and the burn — is what separates a casual reader of the market page from someone who can actually read ultrasound.money and know what the dashboard is showing.
The pre-1559 problem, in one paragraph
Before this reform, every Ethereum transaction included a single gasPrice field. Miners ordered the mempool by that price, descending, and packed the block. Wallets had no reliable estimator, so they over-bid to make sure they got in. The protocol had no concept of block fullness; the only signal was the empirical bid distribution. When demand spiked (the 2017 CryptoKitties weekend, the 2020 yield-farming summer), users would pay 10× the necessary fee because there was no way to detect that the marginal bidder had stopped bidding. The original EIP-1559 FAQ by Vitalik Buterin walks through the inefficiency in detail.
What 1559 actually does — three changes, in order
The proposal authored by Buterin, Eric Conner, Rick Dudley, Matthew Slipper, Ian Norden, and Abdelhamid Bakhta makes three concrete changes. First, it introduces a base fee that the protocol calculates from the previous block’s gas usage. Second, it allows users to add a priority fee (commonly called the "tip") that goes to the validator. Third, it burns the base fee by sending it to address 0x0 where it can never be spent.
The base fee adjusts each block. If the previous block was exactly at the gas target — currently 15 million gas out of a 30 million max — the base fee stays the same. If the block was full, the base fee rises by 12.5%. If empty, it falls by 12.5%. Over twenty blocks (four minutes), the base fee can swing by roughly an order of magnitude. This makes gas predictable on short horizons: the next block’s base fee is mathematically determined the moment the current block is sealed.
The burn — what it is, what it is not
The burn is the single most misunderstood part of 1559. It is not a tax. It is not paid to the foundation, to validators, or to a treasury. The base fee is destroyed at the protocol level, reducing the total ETH supply. The economic case Vitalik makes in his 2021 sharding post and again in the Ethereum roadmap is that burning the base fee aligns user incentives with holder incentives: the network’s most-used moments are also when its currency becomes more scarce.
| Period | ETH burned | Average base fee (gwei) | Net issuance |
|---|---|---|---|
| Aug 2021 – Dec 2021 (first 5 months) | 1.23M ETH | 87 | +0.42M ETH |
| 2022 (full year) | 1.45M ETH | 34 | +0.31M ETH |
| Sept 2022 – Sept 2023 (post-Merge year) | 0.94M ETH | 27 | -0.36M ETH |
| 2024 (Dencun year) | 0.51M ETH | 14 | +0.18M ETH |
| 2025 (post-Pectra) | 0.38M ETH | 9 | +0.24M ETH |
Two things stand out from the table. First, the Merge in September 2022 cut new issuance by roughly 90% by replacing proof-of-work mining rewards with proof-of-stake validator rewards. Combined with the 1559 burn, this is what produced the only twelve-month period in Ethereum history where supply actually shrank. Second, EIP-4844 in March 2024 moved most rollup data off the regular gas market and into a separate blob market, which lowered base fees and therefore the burn. Net issuance turned mildly positive again, not because monetary policy changed, but because the underlying demand for block space relocated.
Why this changed user behaviour, immediately
Within a month of the London fork, wallets converged on a much simpler bidding strategy. Instead of guessing a gas price, you set two values: a maxFeePerGas (the ceiling you are willing to pay) and a maxPriorityFeePerGas (the tip). The actual price you pay is min(maxFee, base + priority). If you set maxFee to 100 gwei and the base fee turns out to be 12, you pay 12 + your tip, not 100. The wallet refunds the difference. This is documented in MetaMask’s transaction docs and was implemented by every major wallet within sixty days.
The empirical result, measured by Paradigm in July 2021 and confirmed by Coinbase’s post-London analysis, was a 10-15% reduction in median user fees paid for equivalent transactions, holding base fee constant. Users stopped overbidding because the protocol guaranteed they would only be charged the realised base. The auction inefficiency that had cost the network billions in cumulative overpayment was retired in a single hard fork.
What 1559 did not solve
It did not make gas cheap. It did not eliminate MEV; in fact, the priority fee channel became a clean signal for searchers to bid into, and MEV-Boost now routes most validator income through a separate proposer-builder auction. It did not change the gas limit, which still constrains total throughput. And it did not address the cross-chain fragmentation that has since become Ethereum’s bigger UX problem.
- Base fee dynamics still cause sudden 12.5%-per-block spikes during mints.
- Priority fee bidding wars during high-MEV blocks routinely push tips above 50 gwei.
- The 30 million gas-per-block ceiling means the network can be saturated by a single popular contract.
- The burn is denominated in ETH; if ETH price falls, the dollar value of the burn falls proportionally.
Why holders care — the supply argument, honestly
The phrase "ultrasound money" that circulated in 2022 and 2023 referred to the period when post-Merge issuance plus the 1559 burn produced net-negative supply growth. As ultrasound.money still tracks, that was true for roughly fifteen months between October 2022 and February 2024. Since then, post-Dencun blob economics have flipped the network back to mild positive issuance. Whether net-negative supply matters for ETH price is a separate question and one that Binance Research and Galaxy Research have argued in opposite directions.
What is not in dispute: the burn turned Ethereum’s fee revenue into a holder dividend in the most literal sense possible. Every transaction destroys ETH, which by definition increases the per-token claim on the network’s future cash flows. Whether that is "monetary policy" or "a fee-rebate distributed by deflation" depends on which paper you read, but the on-chain accounting is unambiguous. You can verify it block by block on Etherscan or model it forward with our ETH issuance calculator.
The legacy — five years in
EIP-1559 is the template every modern L1 fee market has since copied. Polygon PoS adopted a 1559-style mechanism in January 2022. Avalanche’s C-chain integrated a variant in April 2023. Most L2 rollups, including Base and Optimism, expose 1559-style fee semantics to users even though their underlying sequencers don’t strictly need them. The mechanism design is the rare case where a research paper turned into a hard fork turned into an industry standard inside three years.
If you are setting your wallet’s gas slider, watch our gas tools page for the live base fee and use the standard rule: maxFee at 2× current base, priority tip at 1 gwei unless competing for inclusion. The protocol does the rest. That, ultimately, was the point of 1559 — to make the fee market boring enough that you don’t have to think about it.
The activation story — three years of delay, one weekend of execution
It is worth recording how 1559 actually shipped, because the process became the template for every contentious Ethereum upgrade since. The proposal sat in draft from April 2019 through mid-2021, repeatedly stalled by miner opposition. Miners were, reasonably, opposed: the burn directly cut their revenue, and the new tip structure made priority bidding harder to game. A coalition led by Hudson Jameson, Tim Beiko, and the All Core Devs call eventually green-lit activation for the London upgrade, with miners free to fork if they disagreed. Almost none did. The economic majority of the network — exchanges, stakers, large holders — had publicly committed to the canonical chain, which made dissent uneconomical.
The fork itself was uneventful in the way good forks should be. Geth, Nethermind, Besu, and Erigon clients had all shipped 1559-aware versions in the weeks prior. Activation block was a hard-coded number. Block 12,965,000 was sealed, block 12,965,001 was the first block under the new rules, and within minutes wallets were reporting realised base fees in the new format. The transition was, by any measure, the smoothest major monetary-policy change in the history of any cryptocurrency.
A wallet checklist for normal users
- Update to a 1559-native wallet. MetaMask, Rabby, Frame, and Coinbase Wallet all support it; older legacy wallets that still default to a single
gasPricewill overpay. - Use the "Market" speed preset for everyday transactions. The "Aggressive" preset is essentially a 1.5-2× multiplier on tips, useful only when you are competing for inclusion.
- Watch the realised price after confirmation. The wallet will show what you actually paid versus what you authorised — the gap is the refund.
- Avoid manually setting maxFee equal to the current base. If the next block’s base ticks up by 12.5%, your transaction will sit unconfirmed indefinitely.
- For batched activity, use a wallet with a "cancel and replace" feature so you can speed up or drop a stuck transaction with a higher tip.