Crypto protocols spent $1.4 billion on token buybacks in 2025, per CoinGecko data. That figure is set to accelerate in 2026 as Optimism, Pyth Network, Uniswap, and others launch or expand structured repurchase programs. Yet price performance across most buyback-executing tokens remains flat or ne...
"There is no role for an independent service provider if the largest budget recipient can influence its own approval without full disclosure." — Marc Zeller, Founder, Aave Chan Initiative
Crypto protocols spent $1.4 billion on token buybacks in 2025, per CoinGecko data. That figure is set to accelerate in 2026 as Optimism, Pyth Network, Uniswap, and others launch or expand structured repurchase programs. Yet price performance across most buyback-executing tokens remains flat or negative, raising a structural question: who actually benefits — token holders, or the corporate entities behind the protocols?
The buyback wave has reached a critical juncture. Hyperliquid directs 97% of trading fees to HYPE repurchases and burns, generating over $1.2 billion in annualized buy pressure. Uniswap expanded its fee switch to eight L2 chains in late February, targeting $61 million in annualized burn revenue. Jupiter, by contrast, spent $70 million on buybacks in 2025 and saw JUP decline 89% from its peak — prompting its co-founder to publicly question the entire strategy. Aave's governance is fracturing, with the Aave Chan Initiative exiting after a $51 million funding dispute exposed voting power concentration. The data shows buybacks work when funded by real revenue at scale; they fail when used as supply-side cosmetics against large unlock schedules.
Development activity around buyback infrastructure and treasury governance is measurably increasing. A Cairo library for autonomous token buybacks via Ekubo TWAMM (loothero/autonomous_buyback) received its v2 upgrade on January 23, 2026, adding enhanced features including ERC20 burn support and stream components for autonomous token distribution. The repo is early-stage (zero stars) but signals that on-chain buyback execution is becoming programmable infrastructure rather than ad hoc multisig operations.
The M0 Foundation's Two Token Governance (TTG) framework (m0-foundation/ttg, 11 stars) continues to operate as a reference implementation for dual-token governance systems — one token for voting, another for economic rights. Its frontend was updated as recently as February 15, 2026. The separation of governance power from economic claims is directly relevant to the buyback debate: protocols that merge these functions (vote-escrowed models) face different incentive structures than those that separate them.
A token buyback tracking repository (JINGEJINGE/tokenbuybacklong) was last updated February 20, 2026, while Castle Finance's curated list of DAO treasury management tools (castle-finance/awesome-dao-treasury-mgmt) saw activity in early February. These signal growing demand for transparency and tooling around protocol treasury operations. In the AI-agent space, repos like Hybrid (ian/hybrid) — a TypeScript framework for building crypto AI agents — continue to gain attention, though none yet integrate governance or treasury management functions.
DeFi protocols executed over $1.4 billion in token buybacks during 2025, according to CoinGecko data cited by The Block. The concentration is severe: 92% of that total flowed through the top 10 projects, per CWallet analysis. This is a winner-take-all dynamic where protocols with genuine revenue can afford meaningful buybacks, while smaller projects execute token-price cosmetics that are structurally overwhelmed by unlock schedules.
The philosophical split is documented. Four Pillars published "Rev Sharing is Dead. Long Live Buybacks & Burns," arguing that direct revenue sharing (dividends) can function as a liquidity faucet for insiders — foundations and large holders convert yield payouts to cash without visibly selling their core stake. Buyback-and-burn, by contrast, reduces supply proportionally, benefiting all holders equally. The counter-argument: buybacks give protocols discretion over timing and execution, which insiders can exploit through information asymmetry.
The data is mixed on effectiveness. According to DL News, DeFi teams spent $800 million on buyback programs by mid-2025, yet prices for many of those tokens remained flat or declined. Messari data, per MEXC, indicates that projects with regular buybacks failed to establish price floors, with many underperforming the broader market. Helium paused its buyback after seeing no market impact. The signal: buybacks without underlying revenue growth are capital destruction.
Hyperliquid operates the most aggressive buyback program in crypto. The protocol directs roughly 97% of trading fees to continuous HYPE token purchases and burns via its Assistance Fund, per Tokenomics.com. On February 5, 2026, the platform recorded $6.84 million in daily revenue — a new high — with $5.25 million flowing directly to buybacks, repurchasing 160,750 HYPE in a single day, according to AMBCrypto. Annualized protocol revenue approaches $1 billion.
A governance vote permanently burned 37.5 million tokens ($912 million at time of burn) from the Assistance Fund, per The Defiant. Monthly unlock schedules were reduced 88% in February 2026, per SpotedCrypto. The corporate structure is notable: no VC allocation, no exchange listings, a 3:7 team-to-community token ratio. The Hyper Foundation holds 6% of supply. This is the closest approximation to a "protocol as its own best buyer" model currently operating.
The UNIfication proposal passed on December 25, 2025, with 99.9% support (125 million tokens for, 742 against), per The Defiant. It activated the protocol fee switch and executed a retroactive burn of 100 million UNI (~$600 million), representing estimated value that would have accrued to holders had the fee switch been active since launch, according to DL News.
Since activation, Uniswap has burned over $5.5 million worth of UNI, implying an annualized pace of ~$34 million, per Coin Metrics. In late February 2026, a governance vote to expand the fee switch to eight additional L2 chains (Base, OP Mainnet, Arbitrum, Celo, Soneium, Worldchain, X Layer, Zora) gained momentum, pushing UNI up 15% in 24 hours, per CoinDesk. Estimates suggest the expansion could add ~$27 million in annualized revenue, bringing total burn capacity to ~$61 million annually.
The corporate structure question: Uniswap Labs, the private company, operates the frontend and mobile app and charges its own interface fee — separate from the protocol fee. Token holders benefit from the protocol fee switch; Uniswap Labs shareholders benefit from the interface fee. The two revenue streams are distinct.
Jupiter spent over $70 million on JUP buybacks in 2025, per crypto.news. JUP declined ~89% from its peak to trade near $0.20–$0.22 in early January 2026. Monthly unlocks of 53 million JUP through June 2026 increased circulating supply by roughly 150% since launch. The buybacks covered approximately 6% of unlocked tokens — structurally insufficient to offset dilution.
Jupiter co-founder Siong publicly asked the community whether stopping buybacks would be better for the project, according to Cryptopolitan. Solana co-founder Anatoly Yakovenko suggested an alternative: parking profits as protocol-owned assets redeemable through long-term staking locks, per BeInCrypto. Jupiter subsequently cut its planned 2026 airdrop from 700 million to 200 million JUP. The protocol currently distributes 50 million JUP (~$10 million) quarterly as staking rewards plus $10–$20 million in buybacks via 50% of revenue.
Optimism governance approved a 12-month pilot directing 50% of net Superchain sequencer revenue to monthly OP token buybacks, starting February 2026. The vote passed with 84.4% support, per The Block. The Foundation will partner with an OTC provider for monthly ETH-to-OP conversions, with a minimum threshold: the program pauses if monthly revenue falls below $200,000. Purchased tokens are held in the Collective treasury, not burned — a critical distinction. Token holders benefit from reduced sell pressure, but supply is not permanently reduced.
Sky Protocol projects $611.5 million in gross protocol revenue for 2026, an 81% year-over-year increase, with $157.8 million in protocol profits (198% YoY growth), per Sky Ecosystem data. Daily SKY buybacks run at approximately $250,000/day as of January 2026. Total buybacks through 2025 reached $96.8 million (combined exceeding $106 million), removing 1.12 billion SKY (~$75 million) from circulation, per CryptoNews.net.
Pendle is replacing its 2-year vote-escrow lock (vePENDLE) with liquid sPENDLE, featuring a 14-day unstaking period and algorithmic emissions, per Startup News. Under the new system, up to 80% of protocol revenue will fund PENDLE buybacks distributed to active sPENDLE holders. In 2025, Pendle generated over $37 million in protocol revenue, but the complex voting mechanics concentrated rewards among sophisticated vePENDLE holders. The transition to sPENDLE reduces total PENDLE emissions by ~30% while lowering the expertise barrier for fee participation. Governance shifts from weekly voting to periodic "critical" Pendle Protocol Proposals (PPP).
This is a structural evolution: moving from a governance-gated fee model (where only active voters earn) to a staking-gated fee model (where any staker earns). It broadens the distribution base but potentially weakens governance participation incentives.
Ethena activated its fee switch in September 2025 after meeting pre-set benchmarks, per Blockworks. Protocol revenues of $50–$60 million monthly now flow to sENA (staked ENA) holders, implying 4.5–15% annualized yield on approximately $750 million in staked ENA, per Cryptopolitan. Ethena surpassed Q4 2025 revenue totals in just 47 days of Q1 2026, according to FX Leaders. The fee switch could create up to $100 million in buying pressure based on current activity. Corporate structure: Ethena Labs (private company) builds the protocol; Ethena Foundation governs distribution parameters. Token holders receive direct yield; equity holders benefit from protocol growth.
Treehouse Protocol, with $294 million in deposits, launched a TREE token buyback program funded by 50% of Market Efficiency Yield (MEY) fees from its tETH product, per DL News. The program received 99.59% community support. Purchased TREE tokens are held in a DAO-controlled multisig, not burned, with future use determined by governance. TREE doubled in price following the announcement. The protocol exemplifies a smaller-scale, revenue-backed buyback where the ratio of buyback spending to circulating supply is more meaningful than at large-cap protocols.
Pyth Network launched its "PYTH Reserve" program, deploying 33% of DAO treasury balance monthly to open-market PYTH purchases, per The Block. Initial buybacks totaled $100,000–$200,000, with the DAO treasury holding ~$500,000 at announcement. The February 2026 purchases report is published on the Pyth DAO forum. Revenue source: Pyth Pro (premium data feeds) reached $1 million ARR, per FinanceFeeds. The scale is modest, but the transparency — monthly on-chain reports — sets a standard for buyback disclosure.
Ether.fi operates two distinct buyback streams: 100% of eETH withdrawal fee revenue on a weekly cadence, and a portion of broader protocol revenue monthly, per Ether.fi governance docs. The DAO also authorized a $50 million treasury buyback for when ETHFI trades below $3, per The Block. Repurchased tokens are distributed to staked ETHFI holders, not burned — meaning the supply impact is recycled rather than permanent.
Aave's buyback program has acquired over 205,000 AAVE (1.28% of total supply) in under a year. The Aave DAO proposed reducing the annual buyback budget from $50 million to $30 million in February 2026, citing a 25% decline in borrow fees from peak levels, per Aave governance forum.
But the more significant development is structural. On March 3, 2026, the Aave Chan Initiative (ACI) — which drove 61% of governance actions over three years and deployed $101 million in incentives — announced its exit from the DAO, per CoinDesk. The dispute centers on Aave Labs' "Aave Will Win" proposal requesting ~$51 million in stablecoins and 75,000 AAVE tokens. ACI alleged that addresses linked to Aave Labs voted on their own proposal, tipping the outcome. BGD Labs also announced plans to leave by April 2026, per The Defiant.
AAVE dropped 11% on the news. The implication for buyback analysis: a well-funded buyback program is irrelevant if the governance layer controlling treasury allocation is captured by a single entity. Aave Labs (the private company) benefits from DAO-funded development budgets; AAVE token holders benefit from buybacks. When the entity requesting funds can vote on its own proposals, the governance framework fails to protect token holder interests.
The buyback landscape separates into three tiers based on who actually captures value:
Tier 1: Token holders as primary beneficiaries. Hyperliquid (97% of fees to buyback/burn, no VC allocation), Pendle (80% of revenue to sPENDLE holders), Ethena (direct fee switch to stakers). In these protocols, the corporate entity (foundation or labs) takes a minority share of revenue. Token holder and protocol interests are structurally aligned.
Tier 2: Shared value, split incentives. Uniswap (protocol fee switch burns UNI, but Uniswap Labs collects separate interface fees), Aave (buybacks funded by DAO, but Aave Labs requests large development budgets from the same treasury), Sky (buybacks funded by revenue, but Foundation controls execution). In these protocols, token holders benefit, but a parallel corporate entity captures value through different channels.
Tier 3: Buybacks as supply management against dilution. Jupiter ($70M spent, 89% price decline, buybacks covered 6% of unlocks), Optimism (buybacks held in treasury, not burned, and subject to future governance decisions). Here, buybacks function as damage mitigation against aggressive token emission schedules rather than genuine value return.
The critical variable is the ratio of buyback spending to new token emissions. When buybacks cover less than 20% of concurrent unlocks, they are structurally insufficient to create value for existing holders. When they exceed 50% of concurrent emissions — as with Hyperliquid — the supply dynamic shifts meaningfully.
The token buyback wave of 2025–2026 is crypto's clearest attempt to adapt TradFi's shareholder return playbook. The data shows it works under specific conditions: real revenue at scale (Hyperliquid), a supply schedule where buybacks meaningfully exceed new emissions, and governance structures that prevent treasury capture by the protocol's own development entity.
It fails in the absence of these conditions. Jupiter's $70 million buyback against 150% supply inflation produced an 89% price decline. Aave's well-funded program is overshadowed by a governance crisis where the entity requesting $51 million in DAO funds allegedly voted on its own proposal. The distinction is not between protocols that buy back and those that do not — it is between protocols where buybacks represent genuine value return and those where they serve as narrative cover for continued insider value extraction.
The emerging standard is moving toward buyback-and-burn (Hyperliquid, Uniswap) rather than buyback-and-hold (Optimism, Treehouse), toward algorithmic execution (Pendle's sPENDLE) rather than discretionary multisig operations, and toward public monthly reports (Pyth) rather than opaque treasury management. The protocols that survive the next cycle will be those where token holders can verify, not just trust, that value is flowing back to them.