An Ethereum burn address is an address nobody can spend from, so any ETH or tokens sent to it are gone for good. The two best known are the null address 0x0000000000000000000000000000000000000000 and 0x000000000000000000000000000000000000dEaD. The fee burn that Ethereum has run since August 2021 works differently. Under EIP-1559 the base fee of every transaction is destroyed by the protocol and never sent to any address. As of 25 October 2026, Etherscan counts 4,641,004.72 ETH burnt this way. Burning has slowed to a trickle in 2026, though. CryptoSlate's ledger to 9 October 2026 shows fee burns offset only 2.07% of newly issued ETH, so total supply is growing.
The newest numbers: burn is close to zero relative to issuance
CryptoSlate published a supply ledger on 9 October 2026, based on data from ethsupply.fyi, covering 1 January 2026 00:00:11 UTC to 9 October 2026 15:44:23 UTC:
| Item | ETH |
|---|---|
| Gross issuance to validators | 796,623.377 |
| Fee burn (execution and blob fees) | 16,524.553 |
| Consensus penalties | 1,685.919 |
| Net added to supply | 778,412.846 |
Over the period, supply grew about 0.64%, to roughly 122.116 million ETH. For a second check, Etherscan's supply page on 25 October 2026 showed a total supply of 122,161,279.51 ETH. These figures come mainly from one secondary ledger. Etherscan's totals match it, but we did not reconcile the two line by line.
Two things explain the gap.
- Base fees are very low. CryptoSlate cites a sample block base fee of about 0.335 gwei on 9 October 2026. At today's 60 million gas limit, the outlet estimates the base fee would have to average about 13.85 gwei for burns to cancel out issuance.
- Issuance is higher than the often-quoted 1,700 ETH a day. ethereum.org's figure of about 1,700 ETH a day assumed roughly 14 million ETH staked. CryptoSlate reports about 43.754 million ETH of active effective stake across 853,325 validators in October 2026. Spread over the 2026 window, the issuance figure above works out to roughly 2,800 ETH a day.
What is the Ethereum burn address?
The null address: 0x000…000
Etherscan labels 0x0000000000000000000000000000000000000000 "Null: 0x000...000". It notes that the address is not owned by any user and tags it Burn and Genesis. Many token contracts use it as the formal destination of a burn. The OpenZeppelin ERC-20 implementation, for example, "destroys a value amount of tokens from account, lowering the total supply" and emits a Transfer event with to set to the zero address. In that case the tokens are removed from the contract's total supply; nothing is actually stored at address zero.
The address also holds ETH that people sent there directly. Etherscan showed a balance of 14,156.62 ETH when we checked on 25 October 2026. That ETH is locked and out of circulation, but it is a separate pool from the EIP-1559 burn.
The dead address: 0x…dEaD
Etherscan labels 0x000000000000000000000000000000000000dEaD "Null: 0x00...dEaD". It describes the address as "commonly used by projects to burn tokens (reducing total supply)". Projects like it because the suffix is easy to read, and because standard ERC-20 transfers to the zero address are blocked. OpenZeppelin's transfer requires that "to cannot be the zero address", so a plain transfer has to go to a different unspendable address. Etherscan showed 12,642.66 ETH at this address on 25 October 2026.
Why tokens sent there cannot be recovered
Spending from an Ethereum account requires a valid signature from its private key. Nobody is known to hold a key for either address, and these addresses are agreed by convention to be unspendable. The protocol does not lock them in any special way.
How the EIP-1559 fee burn works
EIP-1559 went live in the London upgrade on 5 August 2021. Its status is Final, and its authors are Vitalik Buterin, Eric Conner, Rick Dudley, Matthew Slipper, Ian Norden and Abdelhamid Bakhta.
- Two fees. Each transaction pays a base fee set by the protocol and an optional priority fee. The specification says "the base fee is not given to anyone (it is burned)". The priority fee goes to the block producer, called the miner in the original text and the validator since the Merge.
- Target and limit. The gas target is the block gas limit divided by an elasticity multiplier of 2. Blocks can be up to twice the target.
- Adjustment. With
BASE_FEE_MAX_CHANGE_DENOMINATOR = 8, the base fee moves by at most 12.5% per block. It rises when blocks are above target and falls when they are below.
So burned ETH never moves to the null address or the dead address. It is subtracted from the sender and never credited to anyone. Burn trackers count it by summing base fees across blocks; they do not read the balance of a burn address.
Blob fees are burned too
Dencun (13 March 2024) added EIP-4844 blob transactions, a cheaper data lane for rollups with its own base fee. Galaxy Research reported on 21 August 2024 that 2,408.41 ETH, or 89.45% of blob fee revenue, had been burned in the first 150 days. Pectra (7 May 2025) raised blob throughput through EIP-7691. Fusaka (3 December 2025) shipped EIP-7918, which sets a reserve price for the blob base fee at 1/16 of the execution base fee (BLOB_BASE_COST = 2**13). The aim is to keep blob users paying a share of what they cost nodes. Even with that floor, total fee burn stayed far below issuance in 2026, as the ledger above shows.
What base fee would make ETH supply shrink?
Enter an average base fee in gwei to see roughly what share of new ETH issuance the fee burn would cancel, measured against CryptoSlate's 13.85 gwei break-even.
| Average base fee | Base fee at which burn equals issuance | Share of new ETH issuance cancelled by the fee burn |
|---|---|---|
| 0.335 gwei | 13.85 gwei | 2.42 % |
| 13.85 gwei | 13.85 gwei | 100 % |
| 0.335 gwei | 13.85 gwei | 2.42 % |
- Burn is treated as rising in direct proportion to the base fee, as in CryptoSlate's break-even model; the real ledger burn to 9 October 2026 offset 2.07% of issuance.
- Below 100% the supply keeps growing on these inputs; above 100% the fee burn would outpace new issuance.
This is a rough proportional estimate: it ignores blob fees, consensus penalties, changes in staking and how full blocks actually are, and it cannot forecast base fees.
Issuance: what the Merge changed
The Merge took place on 15 September 2022. According to ethereum.org, miners had received about 13,000 ETH a day, and execution-layer issuance has been zero since then. Stakers then received about 1,700 ETH a day, based on about 14 million ETH staked. ethereum.org says this cut total new issuance by about 88%. Validator issuance rises as more ETH is staked, which is why the 2026 average is above that figure.
Is ETH deflationary in 2026?
No. In the 2026 window, net supply grew by about 778,413 ETH. Ethereum can become deflationary only when base fees stay high enough for long enough. CryptoSlate puts that level at about 13.85 gwei at a 60 million gas limit, against a sampled 0.335 gwei. The gas limit reached 60 million in late November 2025 after validators signalled for it, with no hard fork. Higher capacity tends to keep base fees down unless demand grows to fill it. The next upgrade, Glamsterdam, has no mainnet date on ethereum.org as of this writing. A higher gas limit discussed for after Glamsterdam is a proposal, not a schedule.
How to check burns yourself
- Cumulative burn and supply: Etherscan's supply page (etherscan.io/stat/supply) lists total supply, burnt fees and staking rewards.
- Daily burn: Etherscan's daily ETH burnt chart.
- Live burn and net issuance: ultrasound.money and beaconcha.in/burn.
- Manual burns of a token: open the token contract on a block explorer and look for
Transferevents to0x000…000or0x…dEaD. Check whethertotalSupplyfalls; if it does not, the tokens were locked but not burned.
Limitations: the 2026 supply figures rely mainly on one ledger (ethsupply.fyi via CryptoSlate), and live numbers change by the block.
