Token buybacks top $600M in 2026: why they lift some altcoins and fail others
Projects are spending real revenue on their own tokens at record pace. The results range from Hyperliquid's success to programs that could not outrun dilution.
ALTCOINSDEFIANALYSIS
Daniel Hartley
10/27/20263 min read
For years the standard complaint about altcoins was that their tokens captured little of the value their networks produced. In 2026 a growing number of projects have answered by spending revenue on buying back their own tokens. According to Allium Labs data cited by the Financial Times, buybacks totalled about $638 million between January 1 and August 31, up from $545 million in the same period of 2025 and from almost nothing in 2024. The trend has turned token economics into one of the main ways investors now compare altcoins. The harder question is whether buybacks create lasting value or simply recycle cash into assets that keep falling.
The activity is highly concentrated. Hyperliquid and Pump.fun accounted for nearly 90% of this year's repurchases. Hyperliquid directs about 99% of its trading-fee revenue to buying HYPE through its Assistance Fund and has bought back and retired roughly $1.3 billion worth of tokens since December 2024. HYPE gained about 70% in the year to late August, a period in which most large altcoins fell, making it the clearest case study for supporters of the model.
How a token buyback works
The mechanics are simple. A protocol collects fees from users, usually in stablecoins or its chain's base asset. A smart contract or treasury desk then buys the native token on the open market, either on a fixed schedule or at the team's discretion. What happens next varies. Some projects burn the tokens and reduce supply permanently. Others hold them in a treasury, which keeps flexibility but leaves the door open to resale. A third group distributes them to stakers, which works more like a dividend than a buyback.
The comparison with equity markets is useful. Share repurchases tend to support a stock when they are funded by recurring cash flow and when the company is not issuing large amounts of new stock at the same time. The same two conditions decide whether a token buyback works: real revenue on the way in, and limited new supply on the way out.
Where the model breaks
The list of programs that have not helped prices is long. Jupiter spent nearly $14 million on buybacks this year while its token lost about 55% over twelve months, according to the same Allium figures. Chainlink's token has roughly halved despite its reserve purchases, and Helium dropped its buyback program altogether in February. In each case, the money spent was small relative to market value, to new emissions or to the selling pressure from a weaker market.
Pump.fun shows the tension between buybacks and dilution most clearly. In the week ending August 9 it spent about $5 million to buy and burn 2.15 billion PUMP, and by then its program had offset around 15.7% of the original supply. Yet in July, vested PUMP worth about $86.5 million went to 121 team and investor wallets. A buyback that runs alongside large unlocks often does little more than absorb insiders' supply. Revenue durability is the other weak point: Hyperliquid's income depends on derivatives volume and Pump.fun's on memecoin launches, both of which are cyclical. That makes buybacks pro-cyclical, largest when prices are already high and smallest when support is most needed.
A checklist for reading buyback announcements
Five questions separate substance from marketing. How large is the program relative to market capitalisation and to scheduled unlocks? Is it funded by genuine fee revenue, or by selling treasury tokens or venture money? Are the purchased tokens burned, held or redistributed? Can governance switch the program off or redirect the funds? And are the buying addresses public, so that anyone can verify the purchases on-chain?
Some projects are now writing guardrails into their designs. Lido's proposed NEST framework would only activate once annualised revenue exceeds $40 million, with a limit of $50,000 a day and $10 million over any rolling twelve months. Sky, the protocol formerly known as MakerDAO, bought about $26 million of SKY this year from protocol surplus. These programs are smaller than Hyperliquid's, but their explicit rules make them easier to evaluate. Regulation adds another layer. A token that routes protocol revenue to holders starts to resemble a claim on cash flow, and the SEC and CFTC rule proposals now in progress will shape how such mechanisms are treated in the US. For now, the lesson of 2026 is that buybacks amplify a working business model rather than create one. The projects that benefit are those whose users pay meaningful fees month after month, and whose token supply is not growing faster than the program can absorb.