A token burn permanently removes tokens from circulation by sending them to an address nobody controls. On the blockchain the tokens still exist. Moving them again is impossible, because no private key for that address exists.
Ethereum runs the largest example. Since EIP-1559 activated on 5 August 2021, roughly 4.6 million ETH have been burned through transaction base fees. That is over $9 billion destroyed.
How burning actually works
Simpler than it sounds. No delete function exists on a blockchain.
Tokens go to a burn address, typically one consisting almost entirely of zeros, such as 0x000...000. No corresponding private key exists, and generating one for a chosen address is computationally impossible. So the tokens sit there permanently, visible and unspendable.
In some contracts a burn function is implemented directly, which reduces recorded total supply. Same effect, different accounting.
The two kinds of burn
Manual burns are discretionary. Deciding to destroy a quantity, a team announces it and executes the transaction. Frequency and amount are choices.
Protocol burns are automatic. On every Ethereum transaction the base fee burn happens with no human involvement. BNB Chain runs a similar mechanism, burning a portion of gas fees, alongside a quarterly auto-burn.
When assessing a project, the distinction matters. Tied to real usage, an automatic burn is a property of the system. A manual burn is a marketing decision that can stop at any time.
Why projects burn tokens
- Offsetting issuance. Where new tokens are created as staking rewards or mining subsidies, burning counteracts the dilution.
- Fee capture. Directing fees to a burn returns value to all holders proportionally rather than to a treasury.
- Correcting supply mistakes. Some projects burned oversized initial allocations after launch.
- Removing unsold tokens. Common after a sale that did not reach its cap.
- Signalling. Frequently the real reason. A burn announcement generates attention.
Ethereum's burn, in detail
Worth walking through, since it is the most consequential implementation.
Before August 2021, Ethereum users bid blindly against each other for block space, and all gas paid went to miners. Part of the London hard fork, EIP-1559 split every fee into two parts. Set algorithmically by network demand, the base fee is destroyed. A priority fee still goes to the validator.
Immediately it took effect. Over 100,000 ETH burned within 21 days. More than a million by November 2021.
Combined with the Merge in September 2022, which cut new issuance sharply, Ethereum turned deflationary during periods of heavy activity. Supply actually shrank.
Then the picture changed. In March 2024, the Dencun upgrade moved Layer-2 transaction data off the main chain, cutting fees for users and cutting the burn along with them. By 2026, Ethereum's supply had grown by roughly 950,000 ETH rather than shrinking.
That reversal is the most instructive thing about token burns. Exactly as designed, the mechanism worked. What changed was the volume of activity feeding it.
What a burn does not do
- Demand stays unaffected. Reducing supply matters only if demand holds. Burning tokens nobody wants changes nothing.
- No price effect is guaranteed. Ethereum's burn removed over $9 billion in ETH, a small fraction of market capitalisation.
- Scarcity does not follow. Burning 1% while minting 5% is inflating, whatever the announcement says.
- Appearances mislead. Burning tokens the team was never going to sell is theatre.
- Reversal is impossible, which is genuinely useful and also means mistakes are permanent.
How to assess a burn claim
- Automatic or discretionary? Protocol burns tied to usage are structural. Manual burns are decisions.
- What is the net supply change? Compare burned against newly issued over the same period. Only the net figure means anything.
- Verify it on-chain. Burn transactions are public. Check the address and the amount rather than the announcement.
- What share of supply? Burning 0.01% is a rounding error dressed as news, and a whitepaper rarely spells this out.
- Where did the tokens come from? Burning treasury tokens that were locked anyway costs the team nothing.
Where mb.io fits
$MBG is the utility token of the mb.io ecosystem, providing fee discounts and ecosystem access.
mb.io is a regulated crypto spot exchange backed by MultiBank Group, a financial institution founded in 2005 that serves more than 2 million clients across 100+ countries.
- Regulated by VARA in the UAE and AUSTRAC in Australia
- 10/10 security score from Hacken, an independent blockchain security auditor
- Institutional-grade MPC custody powered by Fireblocks, with segregated client funds
- A curated list of assets, so you're not sorting through thousands of tokens to find the ones worth trading
- Buy, sell, and swap in three steps, from sign-up to purchase
- 24/7 customer support, on web and on the iOS and Android apps
Open your account and start trading on mb.io.
Frequently asked questions
What is a burn address?
An address with no known private key, usually consisting almost entirely of zeros. Sent there, tokens remain visible on the blockchain and can never be moved, which is what makes the burn permanent.
Does burning tokens increase the price?
Not automatically. Supply falls, and price depends on demand as well. Ethereum has burned over $9 billion in ETH, a small proportion of its market capitalisation.
How much ETH has been burned?
Roughly 4.6 million ETH since EIP-1559 activated on 5 August 2021, according to Glassnode data, worth over $9 billion.
Is Ethereum deflationary?
During periods of heavy activity it has been, particularly after the Merge cut issuance. Following the Dencun upgrade in March 2024, burn rates fell and supply grew by around 950,000 ETH into 2026.
What is the difference between a manual and an automatic burn?
Written into the protocol, an automatic burn runs on every transaction, as with Ethereum's base fee. A manual burn is a discretionary decision by a team, which can change or stop.
Can burned tokens be recovered?
No. Without a private key for the burn address, the tokens stay permanently immobile. Intentional, and irreversible.
Do burns show up in total supply figures?
That depends on the implementation. Contract-level burns reduce reported total supply directly. Burns to a dead address leave total supply unchanged while reducing circulating supply, which is why the two figures can differ.
Are token burns a good sign?
Sometimes. Tied to real usage, an automatic burn is a structural property worth understanding. A discretionary burn announced for attention says more about marketing than economics.
