A token burn permanently destroys tokens, either by sending them to an address with no known private key or, in a properly specified mechanism, by never crediting them to anyone in the first place. It changes one side of the market. Supply falls. Nothing about a burn creates a buyer, which is why deflationary therefore bullish is an inference the mechanism itself does not support.
A burn is a supply operation. Whether it moves price depends on demand holding up and on the burn outrunning new issuance, and Ethereum's EIP-1559 says in its own specification text that the second condition is not guaranteed.
Scroll to see the full diagram
The canonical burn, as specified
The reference implementation is Ethereum's transaction fee mechanism. EIP-1559 puts a base fee per gas in protocol, moving up or down each block as a function of the gas used in the parent block against that block's gas target, where the gas target is the block gas limit divided by the elasticity multiplier. The base fee per gas is burned. Only the priority fee, the tip, reaches the block producer.1
Two properties are worth separating out. The burn is automatic, so no governance vote or treasury action triggers it. And it is denominated in demand for blockspace, so the quantity burned is set by how much people want to use the network, not by a policy decision to reduce supply.
That combination is rarer than the word burn suggests. Most burns in the market are discretionary, funded from a treasury, and scheduled by a team. Those are a different instrument wearing the same name.
The arithmetic, from the spec's own constants
The specification's constants make the control loop calculable. The elasticity multiplier is 2, so the gas target is half the maximum block size and a completely full block carries twice the target. The base fee maximum change denominator is 8, so a full block raises the next block's base fee by one eighth, or 12.5%, and a completely empty block cuts it by the same proportion. The initial base fee was set at 1 gwei.1
Run that forward and you can price congestion. Ten consecutive completely full blocks take a base fee of 1 gwei to roughly 3.25 gwei. Ten consecutive empty blocks take it back down to roughly 0.26 gwei. Every unit of gas paid at those rates is destroyed rather than transferred.
So the burn rate is a demand meter. It rises when people compete for blockspace and falls when they do not, and a founder copying the design is copying a mechanism whose output is entirely outside the team's control.
Why the base fee is burned rather than paid to the validator
The burn is not there to reduce supply. It is there to remove an incentive. If the base fee were paid to the block producer, the producer would have a direct reason to manipulate the number the mechanism reads off the previous block, and to arrange things with users outside the protocol.
Roughgarden's analysis formalises this by introducing two forms of incentive compatibility specific to blockchains: one protecting against deviations by profit-maximising miners, and off-chain-agreement proofness, protecting against collusion between miners and users outside the protocol. EIP-1559 satisfies both and is dominant-strategy incentive compatible except during a sudden demand spike.2
That is the difference between a burn that is mechanism design and a burn that is marketing. This one exists because destroying the fee is what makes the fee unmanipulable. Reducing supply is a side effect of the design, not its purpose, and the paper's stated exception during demand spikes is the kind of qualification an unanalysed burn never carries.
Deflationary is a conditional, and the specification says so
EIP-1559's own security considerations include a section headed ETH burn precludes fixed supply. Its text is unusually direct: by burning the base fee, a fixed Ether supply can no longer be guaranteed. If more is burned on base fee than is generated in mining rewards then ETH will be deflationary, and if more is generated in mining rewards than is burned then ETH will be inflationary. Because user demand for blockspace cannot be controlled, the document states that it cannot be asserted whether ETH ends up inflationary or deflationary.1
The authors of the most cited burn mechanism in the industry wrote down that the direction of net supply is undetermined. Anyone quoting the same mechanism as proof that burning makes a token deflationary is contradicting the specification they are citing.
The live position is tracked publicly. The ultrasound.money dashboard is the reference the industry uses for net ETH issuance against the EIP-1559 burn.3 It is a running number that moves with usage, so read it as a gauge rather than as a settled property of the asset.
Supply is one side of a two-sided market
Here is the claim to be careful with, because it is the one that gets asserted most and sourced least. A burn reduces supply. That is mechanical and provable. Whether it raises price requires demand to hold constant and requires the burn to outpace issuance, and no source in the evidence base behind this page establishes a proven causal price effect from burning on its own.
Joel Monegro of Placeholder argued in 2020 that for a token whose value comes from participation in a pool of resources rather than from being spent, burning does not create new value; it redistributes existing value among a smaller group. His sharper point is that whenever the token price does not immediately grow at the same rate as the burn, which he says is most of the time, the burn reduces the network's overall market capitalisation.4
He also disposes of the scarcity guarantee. The people with the power to burn are usually the same people with the power to issue, so a burn does not by itself protect against future dilution; what controls supply is the governance process. His example is Maker, which burns MKR as it earns income and issued new MKR in a solvency emergency in March 2020.4
How to tell a real burn from a headline
Four questions, and they take a minute. What funds the burn: user fees, protocol revenue, or a treasury balance nobody was going to sell anyway. Is it automatic or discretionary. Were the tokens burned actually circulating, or were they an unsold allocation that had never reached the market. And what is the net figure, burn minus issuance, over the same period.
That last one catches most of it. A protocol emitting eight percent a year and burning one percent has an inflation rate of seven percent and a burn announcement. Both numbers are true. Only one of them is in the press release.
The burn we are comfortable signing off looks like this: funded by revenue the business genuinely earned, executed by rule rather than by decision, and reported net of issuance. If those three hold, the burn is a value-accrual route with a known cost. If they do not, it is a supply headline, and supply headlines do not survive the first quarter in which usage falls.
Common questions
What is a token burn?
A token burn permanently destroys tokens, usually by sending them to an address with no recoverable private key or by never crediting them at all. Ethereum's EIP-1559 does the second: the base fee per gas paid on every transaction is burned rather than paid to the block producer, and only the priority fee reaches the validator.1 The result is a supply reduction tied to network usage.
Does burning tokens increase the price?
A burn reduces supply. It does not create demand, and no primary source establishes a proven causal price effect from burning alone. The price effect requires demand to hold and the burn to outpace new issuance. Joel Monegro argued that where the price does not grow at the same rate as the burn, the burn actually reduces the network's overall market capitalisation.4
Is Ethereum deflationary because of EIP-1559?
Conditionally, and the specification says so itself. Its security considerations state that burning the base fee means a fixed supply can no longer be guaranteed, that ETH is deflationary when the burn exceeds issuance and inflationary when it does not, and that because blockspace demand cannot be controlled the outcome cannot be asserted.1 The live balance is tracked on the ultrasound.money dashboard.3
What is the difference between a fee burn and a buyback and burn?
A fee burn destroys tokens users already paid, so no purchase happens and only supply changes. A buyback and burn spends protocol income buying the token on the open market first, which puts real money on the bid, then destroys what was bought. The purchase is the demand-side event. The burn afterwards is a bookkeeping choice about the tokens the protocol now holds.
See Tokenomics Design for how this applies in practice.
Sources
- EIP-1559: Fee market change for ETH 1.0 chain
Ethereum Improvement Proposals, 2019
The base fee is burned and only the priority fee reaches the block producer. Constants read from the specification 3 August 2026: ELASTICITY_MULTIPLIER 2, BASE_FEE_MAX_CHANGE_DENOMINATOR 8, INITIAL_BASE_FEE 1 gwei. The security-considerations section headed ETH burn precludes fixed supply states that the inflationary or deflationary outcome cannot be asserted. - Transaction Fee Mechanism Design
Tim Roughgarden, arXiv:2106.01340, 2021
Introduces myopic miner incentive compatibility and off-chain-agreement proofness, and proves EIP-1559 satisfies both and is dominant-strategy incentive compatible except during a sudden demand spike. - ultrasound.money, ETH issuance and burn dashboard
ultrasound.money
The commonly cited live tracker of net ETH issuance against the EIP-1559 burn. The figure changes continuously, so it is cited as a gauge rather than for any point-in-time total. - Stop Burning Tokens - Buyback and Make Instead
Joel Monegro, Placeholder, 2020
Argues burning a capital-like token redistributes existing value among fewer holders rather than creating value, that a burn outpacing price growth reduces market capitalisation, and that burning does not guarantee protection against dilution. Cites Maker's March 2020 MKR issuance during a solvency emergency.
Last reviewed 2026-08
Know the terms but not sure how they apply to your project? That is what an engagement is for. We design, document, and stress-test the whole token economy inside the Tokenomics Data Room.
100+ projects advised. Complete tokenomics in 4 to 6 weeks.