Token buybacks became the loudest idea in crypto tokenomics in 2026, and the pitch is simple. A network earns money. Instead of keeping every dollar, it spends part of that income buying its own token on the open market. Fewer tokens float around, the argument goes, so each remaining one should be worth more. It borrows the logic of a corporate share buyback, where a company uses profit to shrink its own share count.
That is the theory. The reality is messier, and the gap between the two is where most people get the idea wrong. A buyback is only ever as strong as the revenue paying for it, and a headline number tells you almost nothing about how much buying actually reaches the market. This is a guide to how token buybacks really work, what happens to the coins afterward, and the questions worth asking before you treat one as a reason to buy.
A buyback is fees turned into open-market demand
Start with where the money comes from. A protocol, say a lending market or a perpetuals exchange, charges users fees. Those fees are real income denominated in stablecoins or blue-chip crypto. A buyback program takes an agreed slice of that income and routes it to a wallet that buys the project's token on exchanges, the same way any trader would.
The fee itself can come from several places. A decentralized exchange skims a cut of every swap. A perpetuals venue collects trading fees and, on top of that, the periodic funding payments that longs and shorts pass between each other. If you have never looked at how those work, they are worth understanding on their own, because they are a major revenue line for the biggest buyback programs. Coinliva broke the mechanism down in a piece on how funding rates set an eight-hour fee.
The key point is that a buyback is not the protocol printing value out of thin air. It is recycling money that outside users already paid in. When people describe a token as having "real yield" or "cash flow," this is usually what they mean: income that exists whether or not the token price cooperates. That grounding is exactly what makes buybacks more interesting than the average tokenomics gimmick, and it is also why the revenue figure matters more than any promise about percentages.
What happens to the tokens is not one answer
Once the wallet has bought the tokens, projects do one of a few different things with them, and the choice changes what the buyback actually does to supply.
| Approach | What happens to the bought tokens | Effect on circulating supply |
|---|---|---|
| Burn | Sent to an address no one can spend from, removed permanently | Falls, and the reduction is irreversible |
| Treasury or fund | Held by the project in a wallet it controls | Falls for now, but the tokens can return later |
| Rewards recycling | Redistributed to stakers or liquidity providers | Roughly flat, ownership shifts rather than shrinks |
The difference is bigger than it looks. A burn is a one-way door. A treasury buyback parks the tokens somewhere, and those tokens are still an overhang, because a project that accumulates a large stack can always decide to sell it, use it to pay contributors, or fund an acquisition. Supply that is held is not the same as supply that is destroyed, even if a dashboard counts both as "bought back." Reading which of these a program uses is the first thing that separates a durable design from a marketing line.
Hyperliquid shows the model at full scale
The clearest example of a buyback running hard is Hyperliquid, the perpetuals exchange behind the HYPE token. Its structure routes roughly 99% of trading fees into an on-chain wallet, the Assistance Fund, that buys HYPE continuously. The remaining sliver goes to the protocol's liquidity providers.
The numbers are large. Hyperliquid has taken in more than 1.3 billion dollars in fees since launch, and the Assistance Fund had accumulated over 44 million HYPE by the middle of 2026, a position worth billions at the time. On an average day the fund has bought on the order of a million dollars of HYPE, with some single days running far higher. One estimate put Hyperliquid at about 46% of all token buyback activity across the entire crypto industry in 2025, which tells you how concentrated this trend still is. A handful of high-revenue protocols do most of the buying.
Scale like that is only possible because the fee income is genuinely large. That is the lesson to carry to every other project that announces a buyback: the design is downstream of the revenue, never the other way around.
Why a buyback can be announced and still do nothing
Here the theory and the reality come apart. A buyback needs money flowing through it, and plenty of programs are quieter than their announcements suggest.
The first failure is a revenue shortfall. Some designs only trigger buying when daily income clears a fixed threshold. When revenue drops below that line, the program simply does not fire, and the token gets none of the support holders were promised. Coinliva tracked a stretch where Lido's buyback needed far more daily revenue than it was earning to switch on at all.
The second failure is the gap between the listed rate and the real one. A project can advertise that it directs a headline share of fees to buybacks, while the amount that actually hits the market over a given window comes in well under that. Even Hyperliquid, the strongest case, has shown this: Coinliva found a period where a buyback listed at 99% of fees effectively delivered closer to 61%. Timing, gas, and how "fees" are defined all chip away at the number on the page.
The third failure is price. A buyback funded by the token's own richly valued fees can end up buying near a local high, which is a poor use of the treasury even when the mechanism runs exactly as designed. A program that buys steadily through every price, high or low, is spending some of its income badly, and a token that has already run hot is an expensive thing to keep purchasing.
And a buyback does not run in a vacuum. On the other side of the ledger, most young tokens are still unlocking supply to insiders and investors on a vesting schedule. If a project buys back a few hundred thousand tokens a month while an unlock releases far more into circulation, the net supply is still growing. Buybacks shrink the float; unlocks expand it, and the honest question is which force is bigger this quarter.
How to read a buyback claim without getting fooled
You do not need on-chain forensics to sanity-check one of these programs. A few questions get you most of the way.
- Where does the revenue come from, and is it big and steady, or is it a thin trickle dressed up in a large percentage?
- Are the bought tokens burned, or just held in a treasury that can sell them again?
- Does the buyback fire on a schedule regardless of revenue, or only when income clears a threshold?
- How much supply is unlocking over the same period, and does the buying outrun it?
- Is the published buyback rate matched by what an on-chain tracker shows actually being spent?
Answer those and you will already be ahead of most of the commentary, which tends to stop at the headline figure.
Frequently asked questions
Are token buybacks the same as a stock buyback?
The logic is similar, but the legal and structural details are not. A company buying its own shares reduces the share count and often boosts earnings per share. A token buyback uses protocol revenue to purchase tokens, yet those tokens may be burned, held, or recycled, and there is no company balance sheet or securities regulator standing behind the process the way there is in equities.
Does a buyback guarantee the price goes up?
No. Buying pressure can support a price, but it competes with sellers, with new supply from unlocks, and with the broader market. A buyback improves the odds over a long horizon only if the revenue behind it is real and durable. Over any short window it can be swamped by everything else moving the market.
What is the difference between a buyback and a burn?
A buyback is the act of purchasing tokens with revenue. A burn is one thing a project can choose to do with tokens afterward, sending them to an unspendable address so they are gone for good. A project can buy back and burn, buy back and hold, or burn tokens it never bought, so the two words are not interchangeable.
How can I check whether a buyback is actually happening?
Look for the wallet the project uses for buybacks and watch its activity on a block explorer, then compare that against the protocol's reported revenue on a data site like DefiLlama. If a project claims a large buyback but the wallet is quiet or revenue is thin, the claim is louder than the reality.
Why did buybacks become such a big theme in 2026?
A small group of protocols started earning serious fee income, and returning that money to token holders through buybacks gave their tokens something most cryptocurrencies lack, which is a link between usage and value. That made the model attractive to copy, even for projects whose revenue could not really support it, which is exactly why the scrutiny matters.