DeFi protocols spent $18.8 billion on token buybacks and burns between January 2025 and July 2026, according to [Tokenomist](https://tokenomist.ai/research/buyback-and-burn-explained-what-they-are-who-is-doing-them-and-whether-they-actually-work). Only two of eleven major programs — BNB and RAY —...
"That era is over." — Matt Hougan, CIO, Bitwise Asset Management, on the shift from governance-only tokens to revenue-linked value accrual
DeFi protocols spent $18.8 billion on token buybacks and burns between January 2025 and July 2026, according to Tokenomist. Only two of eleven major programs — BNB and RAY — achieved net supply reduction after accounting for scheduled unlocks. The gap between buyback announcements and actual deflation defines the current state of token value return mechanisms.
Three developments in the past seven days accelerated the trend. Ethena opened a governance vote on August 27 to route 95% of foundation net revenue into programmatic ENA buybacks, contingent on USDe supply reaching $7.5 billion — roughly 85% above its current $4.04 billion, per The Block. Hyperliquid activated AQAv2 on August 26, creating a new yield-funded channel expected to add $135–160 million in annualized HYPE buyback capacity, according to Cryptopolitan. Maple Finance began executing MIP-021 buybacks in August with a tiered revenue-allocation framework, per MSB Intel.
The data reveals a widening divide: protocols with genuine, recurring revenue (Hyperliquid, Aave, Raydium) are compressing token supply, while those relying on treasury reserves or facing large unlock schedules (Jupiter, Optimism) are running buyback programs that amount to marginal offsets against dilution.
Development activity around buyback infrastructure has increased measurably in Q3 2026. A GitHub search for "buyback burn" repos sorted by update date returns five repositories updated within the past ten days, up from near-zero six months ago.
Notable signals:
NerdHerderDani/perma (updated Aug 25): An independent on-chain verification tool for Jito's JIP-38 — the JTX-to-JTO buyback pipeline. The repo implements static, zero-custody, bring-your-own-key verification, indicating community demand for transparent tracking of programmatic buybacks. One fork, zero stars — niche but functional.
dngr2/buyback-burner (updated Aug 21): A Foundry-tested smart contract implementing treasury-backed price floors with automatic burns to the 0xdEaD address. The contract includes unit and invariant tests for reserve conservation. This pattern — operator-funded reserves with permanent burns — represents a hardcoded alternative to governance-dependent buyback schedules.
davyjonesintern/token-buyback-hurdle (updated Aug 10): A TypeScript tool that tests whether annual buyback volume can clear a protocol's market-cap hurdle. The repo frames buybacks as a mathematical sufficiency question rather than a narrative one. Zero stars, but the framework mirrors the analytical approach institutional allocators are applying.
sentient-agi/CryptoAnalystBench (updated Aug 15, 12 stars, 4 forks): A benchmark for evaluating crypto AI agents on long-form analysis. The intersection of AI evaluation and crypto analytics infrastructure signals growing institutional tooling around token fundamental analysis.
The GitHub data suggests that buyback verification, not just execution, is becoming an infrastructure category. Independent tracking tools for Jito and generic buyback-burn contracts point to a market that increasingly demands auditability.
Tokenomist's comprehensive analysis of 27 tokens with active buyback or burn programs reveals a stark reality. Between January 2025 and July 2026, protocols executed $14.9 billion in burns and buybacks in 2025, followed by $3.9 billion in the first seven months of 2026, per Tokenomist.
The headline numbers mask a structural problem: most programs fail to offset token emissions.
| Protocol | Cumulative Buyback/Burn | Annual Inflation (Net of Buyback) | |----------|------------------------|-----------------------------------| | HYPE | $1.375B | +47% | | ASTER | $323.7M | +24% | | PUMP | $300.4M | +14% | | JUP | $275.8M | Offset by $1.2B in unlocks | | RAY | $216M | Net deflationary | | BNB | Quarterly burns | Net deflationary | | KAITO | Active program | +~100% |
Only BNB (scheduled quarterly destruction) and RAY (12% of trading fees automatically buy and burn) achieved genuine supply reduction when forward unlock schedules are factored in. HYPE's $1.375 billion in cumulative buybacks — the largest in DeFi — still resulted in 47% annual inflation because the token's emission schedule overwhelms repurchases.
The distinction that matters: a buyback funded by recurring protocol revenue is durable. A buyback funded by treasury reserves, or one that merely redistributes tokens to stakers rather than burning them, creates motion without permanent supply reduction.
On August 27, the Ethena Foundation opened a governance vote to implement a fee switch across all business lines — USDe savings, white-label stablecoins, and the forthcoming "Ethena [X]" product, according to crypto.news.
The proposal's structure:
Concurrent structural changes include the Foundation purchasing locked ENA tokens from certain seed investors who had been selling over the prior nine months, eliminating monthly venture investor token releases going forward, per Cryptowisser. ENA surged 23% on the announcement.
The corporate structure angle: this proposal channels revenue through the Ethena Foundation, not the DAO. Token holders vote to approve the mechanism but the Foundation executes buybacks and controls the growth allocation. The entity intermediation is explicit.
Hyperliquid activated its AQAv2 framework on August 26, creating a new yield-funded channel for HYPE buybacks, per Blockonomi. The mechanism directs yield earned from over $5 billion in USDC reserves into the Assistance Fund, which executes automated open-market HYPE purchases.
Cumulative figures: $1.3 billion in HYPE burned since launch in November 2024. The Assistance Fund already routes 97% of trading fees into continuous buybacks. AQAv2 adds a second revenue stream — reserve yield — estimated at $135–160 million annualized.
The critical nuance: despite $1.3 billion burned, HYPE still inflates 47% annually because emissions dwarf repurchases. Hyperliquid's $800 million+ in annualized revenue and 99% allocation to buybacks represent the most aggressive value-return mechanism in crypto, but the token's dilution schedule remains the binding constraint.
No foundation or corporate entity intermediates Hyperliquid's buyback. Revenue flows programmatically to the Assistance Fund. There are no equity investors, no venture token allocations. This structure eliminates the corporate-entity conflict present in most other protocols.
Maple Finance activated MIP-021 in August, a governance-approved buyback program that scales SYRUP repurchases based on monthly revenue tiers, according to MSB Intel:
Prior activity: the Syrup Strategic Fund bought back 8 million SYRUP in 2025 and 2.5 million in early 2026, deliberately pacing purchases. The protocol targets $2 billion in TVL for its permissionless Syrup.fi arm.
Maple's corporate structure: Maple Labs (the development company) is distinct from the DAO that governs SYRUP parameters. Revenue accrues to the protocol, and governance directs allocation. The tiered structure is notably conservative — at sub-$1.5M monthly revenue, only 10% reaches buybacks.
Uniswap activated its fee switch on December 28, 2025, redirecting approximately one-sixth of swap fees into TokenJar contracts that buy and burn UNI, per NewsBTC. Governance Proposal 100 expanded coverage to v4 pools across seven networks in July 2026, pushing daily protocol revenue from $114,000 to $325,000.
Eight months post-activation: $23 million in cumulative protocol revenue. Ark Invest estimates annualized burns at $90 million. On August 21, Uniswap burned approximately 150,000 UNI ($590,000) in a single day, a new record, according to Startup Fortune.
The uncomfortable data point: UNI fell to a cycle low despite functional fee burns. This contradicts the thesis that fee switches automatically translate to token price appreciation. LP yield compression (the fee switch takes from LPs) and broader market conditions have outweighed the deflationary mechanic.
Aave activated Aavenomics 3.0 on June 27, 2026, automating buyback execution through an on-chain engine that routes protocol revenue directly into open-market AAVE purchases, per Phemex. The mechanism removes approximately 292 AAVE from circulation daily, funded by roughly $400 million in annualized protocol revenue.
Prior to automation, the governance-approved buyback program (August 2024, $50M annual budget) acquired over 205,000 AAVE (1.28% of total supply) through February 2026. Weekly purchases ranged from $250,000 to $1.75 million.
Aave's corporate structure evolution: the "Aave Will Win" framework directs 100% of product revenue — protocol fees, front-end revenue, Aave Card income, institutional services — to the DAO treasury. Token holders, via governance, control allocation. Aave is one of the few protocols where only OKB and AAVE clearly beat Bitcoin 30 days after announcing buyback programs, per Tokenomist's analysis.
The Jito DAO passed JIP-38 on July 13, 2026, mandating that 100% of the DAO's share of JTX revenue (80% of platform fees) flows through a "Rev Splitter" for automated JTO buybacks and permanent burns, per Crypto Briefing. JTX launched on July 14.
The commitment runs through at least Q4 2027, when the DAO will review performance. The Rev Splitter is on-chain and verifiable — the PERMA terminal GitHub repo exists specifically to enable independent audit of the pipeline.
Jupiter allocated 50% of platform fees to JUP buybacks starting January 2025, executing approximately $275.8 million in total repurchases, per BeInCrypto. However, buybacks covered only 6% of unlocked tokens against $1.2 billion in scheduled unlocks, and JUP declined 89%.
Founder Siong publicly proposed halting buybacks and redirecting funds to user incentives, stating the $70 million spent "obviously didn't move much," according to Yellow. The 2026 Jupuary airdrop was reduced from 700 million to 200 million JUP to limit dilution.
Jupiter illustrates the buyback-unlock mismatch: when emissions exceed repurchases by 16x, the mechanism is cosmetic. The protocol is pivoting toward growth spending over supply reduction — a structural admission that buybacks without supply discipline do not work.
Pendle retired vePENDLE in late January 2026, replacing it with sPENDLE — a liquid staking governance token with a 14-day withdrawal period (or instant exit at 5% fee), per CoinMarketCap. The new model directs up to 80% of protocol revenue to PENDLE buybacks for sPENDLE holders and cuts emissions by approximately 30% through algorithmic allocation.
The shift from lock-based to liquid governance is significant: vePENDLE required up to 2-year commitments, while sPENDLE is transferable and composable across DeFi. Legacy holders received up to 4x boosted sPENDLE based on remaining lock duration.
The design trades governance commitment depth for capital efficiency — a bet that liquid staking produces more durable participation than forced lockups.
GMX transitioned from ETH-denominated rewards to a "Buyback and Distribute" model in late 2024. Since March 5, 2026, the DAO repurchased 313,650 GMX for approximately $1.965 million at an average price of $6.27, per Crypto Briefing. 30% of V1 fees and 27% of V2 fees fund purchases.
A governance proposal to increase fee coverage from 27% to 90% is under discussion. Reward distribution triggers only when GMX trades above $90 — a mechanism designed to prevent selling pressure at low prices.
Defi App DAO's Proposal DIP-004 allocates 80% of protocol revenue to systematic weekly HOME buybacks, per Tokenomist. The platform reports over 500,000 active users and $1.93 million in estimated daily buy pressure at current revenue levels. Four consecutive weekly buybacks totaled $330,000 in HOME purchased.
The model is structurally similar to Hyperliquid's approach — high revenue allocation to automated buybacks — but at significantly smaller scale and without the burn component that would make supply reduction permanent.
The data supports a three-tier classification of current value-return mechanisms:
Tier 1 — Net Deflationary (Revenue > Emissions):
Tier 2 — Large Buybacks, Still Inflationary:
Tier 3 — Conditional or Early-Stage:
The corporate structure determines who controls the surplus. Hyperliquid's programmatic, no-entity model routes value directly. Aave's DAO-controlled treasury gives token holders governance over allocation. Ethena's Foundation-mediated approach channels value through a centralized entity that executes at its discretion. Optimism's buyback is vulnerable to ecosystem departures — Base leaving stripped roughly 87% of buyback capacity.
Morpho represents an alternative model entirely: the June 2025 restructuring made Morpho Labs a subsidiary of the Morpho Association, legally owned by MORPHO token holders. Apollo Global Management committed to purchasing up to 90 million MORPHO tokens over 48 months. Rather than buybacks, Morpho aligns token value directly with equity — a TradFi approach to the token-equity gap.
$18.8B spent on buybacks and burns in 18 months, but only 2 of 11 major programs (BNB, RAY) achieved net supply reduction after accounting for unlock schedules. The gap between buyback volume and actual deflation is the defining metric.
Ethena's fee switch vote (Aug 27) proposes 95% of Foundation revenue to ENA buybacks, but requires USDe to nearly double from $4.04B to $7.5B before activation. The conditional structure delays value accrual indefinitely if growth stalls.
Hyperliquid's AQAv2 (Aug 26) adds $135–160M in annualized buyback capacity from reserve yield, supplementing $800M+ in trading revenue already directed to burns. Cumulative burns: $1.3B. Despite this, HYPE inflates 47% annually.
Uniswap's fee switch generates $325K/day across seven chains but UNI hit a cycle low regardless. Fee switches are necessary but not sufficient for price performance.
Jupiter's $70M buyback failure prompted the founder to propose redirecting funds to user incentives — a public admission that buybacks cannot compensate for 16x unlock ratios.
Pendle's sPENDLE migration trades governance lock commitment for capital efficiency, routing 80% of revenue to buybacks while cutting emissions 30%. The shift from ve-tokenomics to liquid staking governance may set a template.
Morpho's legal restructuring — making Labs a subsidiary owned by token holders via the Association — represents an alternative to buybacks: direct equity alignment rather than market repurchases.
Emission schedules overwhelm buybacks: Most programs cannot offset planned token unlocks. HYPE's 47% annual inflation despite $1.3B burned is the clearest example. Buyback announcements without emission reduction are incomplete.
Revenue dependency: Buybacks funded by genuine protocol revenue (Aave, Hyperliquid, Raydium) are sustainable. Those funded by treasury reserves are finite. Programs should be evaluated on revenue durability, not dollar volume.
Regulatory uncertainty: The SEC has not issued specific guidance on token buyback programs. The resemblance to equity share repurchases creates potential securities-law exposure. Ethena's Foundation-mediated structure and Jupiter's DAO-directed spending represent different regulatory risk profiles.
Corporate entity misalignment: Foundation-controlled buyback execution (Ethena, Optimism) introduces agency risk. Token holders vote on parameters but do not control execution timing or counterparty selection.
Conditional mechanisms may never activate: Ethena's $7.5B USDe threshold requires 85% supply growth. If market conditions prevent expansion, the buyback mechanism remains dormant indefinitely while the governance vote created positive price impact.
LP yield compression from fee switches: Uniswap's fee switch takes approximately one-sixth of swap fees from liquidity providers. If LP yields compress enough to drive liquidity migration, the revenue base that funds burns could erode.
Optimism's ecosystem fragility: Base's departure reduced OP buyback capacity by 87%. Revenue-dependent programs are only as stable as the ecosystems generating fees.
The crypto buyback movement has entered its accountability phase. After $18.8 billion in cumulative spending, the market has enough data to distinguish between programs that create genuine value accrual and those that provide narrative cover for token dilution.
The protocols that will define the next cycle of token value accrual share three characteristics: revenue-funded buybacks (not treasury-funded), permanent burns (not redistribution), and emission schedules that buyback volume can actually offset. By these criteria, the field narrows considerably. Raydium, BNB, and Aave meet all three. Hyperliquid meets the first two but fails the third. Most others meet one or none.
The structural trend is clear: DeFi is converging on revenue-linked value return mechanisms that resemble traditional equity buybacks. The gap between announcement and execution — between buying tokens and actually shrinking supply — is the metric that separates real value accrual from marketing.