Crypto protocols have spent $18.8 billion on token buybacks and burns since January 2025, according to Tokenomist (formerly Unlocks.app) data through late July 2026. Of that total, $16.0 billion went to burns and $2.8 billion to buybacks. The trend accelerated in 2025 ($14.9 billion) but has slow...
"Outside of Bitcoin, the value of crypto assets will increasingly be defined by the same metric that defines stocks and bonds: revenue." — Matt Hougan, Chief Investment Officer, Bitwise Asset Management
Crypto protocols have spent $18.8 billion on token buybacks and burns since January 2025, according to Tokenomist (formerly Unlocks.app) data through late July 2026. Of that total, $16.0 billion went to burns and $2.8 billion to buybacks. The trend accelerated in 2025 ($14.9 billion) but has slowed in 2026, with $3.9 billion deployed year-to-date.
The gap between announcement and execution is wide. Of 11 major tokens running buyback or burn programs, only two — BNB and RAY — are net-deflationary over a trailing 12-month window. The rest, including high-profile programs from Hyperliquid, Pump.fun, and Jupiter, remain net-inflationary once scheduled token unlocks are factored in. On August 13, Bitwise CIO Matt Hougan published a memo arguing that crypto valuations "could double or more" as protocols increasingly link revenue to token value — a thesis the data only partially supports.
The buyback-and-burn mechanism is not new — Binance has burned BNB quarterly since 2017 — but its adoption across DeFi protocols represents a structural change in how crypto projects return value to token holders. According to Tokenomist research covering 27 tokens, the breakdown is as follows:
| Period | Burns | Buybacks | Total | |--------|-------|----------|-------| | 2025 Full Year | $12.7B | $2.2B | $14.9B | | 2026 YTD (Jul) | $3.3B | $0.6B | $3.9B | | Combined | $16.0B | $2.8B | $18.8B |
Burns account for 85% of the total. The distinction matters: a burn permanently destroys tokens via a zero-access wallet, while a buyback can end in three outcomes — burn (supply reduction), hold (temporary float removal), or redistribute (no net effect). Only burn or buyback-then-burn reduces total supply.
The mechanism's funding source is equally critical. Programs funded by recurring protocol fees (Hyperliquid, Aave, Raydium) are structurally durable. Programs funded by treasury reserves run until the treasury empties — as KAITO demonstrated when it paused its buyback after revenue collapsed.
Hyperliquid (HYPE): The perpetual futures exchange has deployed $1.3 billion in HYPE buybacks since the token launched in November 2024. Its Assistance Fund routes 97% of protocol fees into continuous, automated market purchases. Annualized revenue stands at roughly $800 million, though quarterly revenue fell 51% from approximately $290 million in Q3 2025 to $149 million in Q2 2026. The decline stems from builder-deployed markets under HIP-3, where Trade.xyz's real-world asset perps now drive roughly half of Hyperliquid's volume but pass through less fee revenue to the protocol, according to CoinDesk.
Pump.fun (PUMP): The Solana-based token launchpad executed one of crypto's largest single-day burns on April 29, 2026, destroying $370 million worth of PUMP — 36% of circulating supply. The platform subsequently scaled back from burning 100% of revenue to a 50/50 split: half to buyback-and-burn, half to operations. Cumulative revenue exceeded $1 billion since the January 2024 launch. Annual revenue, however, has fallen from $971 million in 2025 to an annualized rate of approximately $322 million in 2026, according to Tokenomist data.
Aave (AAVE): The lending protocol activated Aavenomics 3.0 on June 27, 2026, replacing discretionary buybacks with an automated, non-discretionary engine. The mechanism removes approximately 292 AAVE from circulation daily, funded by roughly $400 million in annualized protocol revenue. Aave's buyback does not burn tokens — purchased AAVE is sent to the ecosystem reserve for redistribution. The protocol has repurchased more than 1.2% of total supply since the program began in April 2025.
Uniswap (UNI): Governance Proposal 100 activated protocol fees across selected v4 liquidity pools on seven networks on July 27, 2026. Daily protocol revenue rose to approximately $325,000 from a prior run rate of $114,000. The initial UNIfication vote in December 2025 burned 100 million UNI (approximately $590 million). Since then, roughly 7 million additional UNI have been burned via the TokenJar mechanism, implying an annualized burn rate of about $34 million at current trading levels. Annual protocol revenue sits at approximately $100 million.
Raydium (RAY): The Solana-based DEX has quietly operated a fee-funded auto-burn since 2022, directing 12% of trading fees to token destruction. Cumulative burns total $216 million (90.8 million RAY). Unlike most peers, Raydium is genuinely net-deflationary — supply shrank 6.8% over the trailing 12 months.
Lighter (LIT): The perpetual futures DEX burned 15.64 million LIT on July 10, removing 6.3% of circulating supply through its Q2 2026 revenue-funded buyback. Annualized revenue sits at $26.3 million with $69 million in cumulative fees. The protocol is net-deflationary today at a burn rate of approximately 30.6 million LIT annually, but faces a structural cliff: when vesting opens on December 27, 2026, gross vesting of roughly 166 million LIT per year will run 5.4 times the current burn pace.
Tokenomist's analysis applied a straightforward test: does the trailing 12-month recurring burn rate exceed forward 12-month scheduled token unlocks? The results are sobering.
| Token | Net 12-Month Supply Change | Status | |-------|---------------------------|--------| | BNB | -4.5% (6.0M tokens) | Deflationary | | RAY | -6.8% (18.4M tokens) | Deflationary | | BGB | +0.4% (3.0M tokens) | Near breakeven | | PUMP | +14.2% (56.5B tokens) | Inflationary despite burns | | ASTER | +23.7% (636.9M tokens) | Inflationary | | HYPE | +47.1% (104.7M tokens) | Inflationary despite $1.3B buyback | | KAITO | +99.9% (241.2M tokens) | Paused buyback, massive unlocks |
HYPE's result is notable: despite $1.3 billion in buybacks, scheduled token unlocks add 104.7 million HYPE to supply over 12 months, resulting in 47.1% net inflation. Pump.fun shows a similar dynamic — the $370 million burn was offset by team and investor unlocks, including 6.875 billion PUMP tokens that unlocked on August 12, 2026.
The buyback-and-burn trend is no longer limited to DeFi protocols. Layer-1 blockchains are restructuring their tokenomics toward supply reduction.
Solana: SGP-0003, a governance proposal currently in signaling, bundles two SIMDs. SIMD-0553 introduces resource-based transaction fees, replacing Solana's relatively uniform fee model. SIMD-0550 doubles the annual disinflation rate to 30%, pulling Solana's 1.5% inflation floor forward to 2029 from 2032 and removing approximately 18.9 million SOL in emissions over six years. Daily burns would rise from roughly 650 SOL ($48,000) to between 7,500 and 9,000 SOL ($650,000), according to CoinDesk estimates. The proposal has backing from 24.94 million SOL in stake but requires approximately 40 million more to clear the 15% signaling threshold before an August 18 deadline.
Aptos: On April 14, 2026, Aptos increased base gas fees tenfold while maintaining stablecoin transfer costs at approximately $0.00014. All base gas fees are directed to on-chain burns. Staking rewards were cut from 5.19% to 2.6%, and the protocol introduced a hard supply cap at 2.1 billion APT with 210 million APT locked. Projected annual burns exceed 32 million APT, driven largely by DEX volume.
Buyback announcements have not consistently translated into outperformance. Tokenomist tracked 30-day price performance relative to Bitcoin following each program's announcement:
| Token | 30-Day Return vs. BTC | |-------|-----------------------| | OKB | +318.7% | | AAVE | +29.1% | | PUMP | +4.8% | | ASTER | -3.8% | | HYPE | -13.0% | | UNI | -21.5% | | JUP | -22.7% |
OKB's outlier performance followed a one-time burn to a fixed 21-million-token supply. Aave's fee-funded recurring buyback showed meaningful outperformance. The remaining five underperformed Bitcoin in the month following their announcements.
Valuation multiples vary widely across protocols operating buyback programs:
| Protocol | P/S Ratio | Revenue Trend | |----------|-----------|---------------| | PUMP | 2.4x | Declining | | RAY | 5.7x | Declining | | JUP | 5.6x | Declining | | PENDLE | 10.8x | Declining | | AAVE | 13.2x | Declining | | HYPE | 15.5x | Declining | | UNI | 88.8x | Mixed |
The common denominator: every protocol's revenue is currently declining, complicating the "revenue repricing" thesis.
The buyback narrative arrives during a period of broad fee compression across crypto. According to CryptoRank data from June 2026, average crypto fees have fallen 44.6% year-to-date. Sector breakdowns:
Total DeFi TVL fell $43.4 billion in H1 2026, according to Binance Research. The fee decline reflects a broad deceleration in on-chain activity, not a structural collapse, but it raises a question about whether protocols can sustain buyback programs at current rates. Hyperliquid's 51% quarterly revenue decline and Pump.fun's revenue deceleration from $971 million to a $322 million annualized rate illustrate the risk.
The shift from governance-only tokens to revenue-linked instruments represents a structural change in crypto market design. The comparison to equity buybacks is directionally accurate — U.S. public companies spent $1.37 trillion on buybacks in 2024, according to S&P data — but crypto's version faces complications that equities do not: concurrent token unlocks, declining fee revenue, and mechanisms that range from permanent supply destruction to temporary float management.
Hougan's thesis that crypto valuations "could double or more" requires two conditions: that protocol revenue stabilizes or grows, and that buyback mechanisms genuinely reduce circulating supply. At present, neither condition is broadly met. Revenue is falling across the sector, and most programs are net-inflationary after accounting for vesting schedules. The two protocols that pass both tests — BNB and Raydium — are outliers, not the norm.
The economic logic is sound: protocols that generate real revenue and return it to token holders via permanent supply reduction should, all else equal, command higher valuations. The question is whether the current crop of programs represents that model or merely borrows its language.