DeFi protocols generated more than $1.4 billion in token buybacks through the first five months of 2026, according to Keyrock data, a fivefold increase from the same period in 2024. The shift marks a structural transition: major protocols are routing protocol-level fees to token holders through b...
"Aave v4 is actually taking Aave into a new environment where we can start funding opportunities in the real world." — Stani Kulechov, Founder & CEO, Aave Labs
DeFi protocols generated more than $1.4 billion in token buybacks through the first five months of 2026, according to Keyrock data, a fivefold increase from the same period in 2024. The shift marks a structural transition: major protocols are routing protocol-level fees to token holders through buyback-and-burn mechanisms rather than relying on token inflation and liquidity mining to sustain valuations. Roughly 64% of revenue across leading protocols now flows back to token holders, a reversal from prior cycles where reinvestment and treasury accumulation were the default.
The five largest DeFi protocols by revenue — Hyperliquid, Sky (formerly MakerDAO), Aave, Uniswap, and Ethena — have each adopted distinct approaches to the same problem: how to convert protocol usage into token-holder value without triggering securities classification. Their combined annualized revenue run rate now exceeds $2.5 billion, built on trading fees, lending spreads, stablecoin interest margins, and staking commissions. This report compares their revenue models, buyback structures, capital efficiency, and risk profiles.
Hyperliquid leads all DeFi protocols by annualized revenue. The perpetual futures exchange recorded an $843 million annualized run rate in March 2026, with $255 million in year-to-date revenue as of May. The platform generated a record $6.84 million in single-day revenue on February 5, 2026. Revenue derives almost entirely from trading fees on leveraged perpetual contracts.
Sky (MakerDAO) reported $123.8 million in gross protocol revenue for Q1 2026 — its highest quarterly figure on record — and $46 million in net protocol surplus, approximately 92% of full-year 2025 net surplus achieved in a single quarter. Total protocol collateral reached $13 billion. Annualized projections point to $611.5 million in 2026 revenue, driven by institutional demand for USDS stablecoins and lending operations. Sky derives income from stability fees on collateralized debt positions, liquidation penalties, and interest earned on reserves.
Aave processed over $3.33 trillion in cumulative deposits and nearly $1 trillion in loans through its history, generating approximately $885 million in cumulative fee revenue. The protocol's TVL stands at approximately $24 billion across all versions. Aave V4, launched on Ethereum mainnet on March 30, 2026, introduced a hub-and-spoke architecture with three hub types — Core, Plus, and Prime — to segment risk and optimize capital allocation. Revenue comes from the spread between borrowing rates and deposit rates.
Uniswap activated its fee switch in December 2025, routing 17% of swap fees toward UNI buyback and burn operations. The protocol expanded the fee switch to eight Layer 2 networks (Arbitrum, Base, Celo, OP Mainnet, Soneium, X Layer, Worldchain, and Zora) in February 2026, adding an estimated $27 million in annualized revenue. The combined UNI burn revenue stack went from zero before December 2025 to approximately $61 million annualized. Base overtook Ethereum as Uniswap's largest fee-generating chain in 2026, with $55 million in trader fees since January 1.
Ethena operates USDe, the third-largest stablecoin at $5.92 billion in circulating supply as of March 2026. The protocol generated over $1.2 billion in annualized revenue in December 2024 through funding rates on delta-neutral positions, staking rewards, and minting fees. The staked version (sUSDe) offered a 7-day trailing average APY of 9.4% and a 90-day average of 11.8% as of April 25, 2026. Ethena's fee switch was approved by the foundation in September 2025 but remains unactivated pending governance decisions.
The protocols have adopted meaningfully different buyback architectures:
| Protocol | Mechanism | Annual Buyback Budget | % Revenue to Holders | Status | |----------|-----------|----------------------|----------------------|--------| | Hyperliquid | Assistance Fund auto-buys HYPE | ~$820M annualized | 97% | Active since late 2024 | | Sky | Treasury-funded SKY buybacks | Projected from $611.5M revenue | Surplus-directed | Active, building $150M reserve | | Aave | "Aave Will Win" DAO buyback | $50M annual allocation | 100% routed to DAO | Active since Q1 2026 | | Uniswap | Fee switch to UNI burn jar | ~$61M annualized | 17% of swap fees | Active, expanding to L2s | | Ethena | Fee switch approved, unactivated | TBD | 0% currently | Pending governance |
Hyperliquid operates the most aggressive buyback model. Ninety-seven percent of all trading fees flow to the Assistance Fund, which automatically purchases HYPE tokens from the open market. The fund has accumulated over $1 billion in HYPE and burned 40.5 million tokens since inception, reducing circulating supply. Only 3% of revenue goes to HLP liquidity providers.
Sky directs net protocol surplus toward building a $150 million reserve target (currently at $60.73 million) before distributing excess through buybacks. The protocol reported $46 million in net surplus in Q1 2026 alone.
Aave's "Aave Will Win" governance package routes 100% of product revenue to the DAO, funding active AAVE buybacks. Founder Stani Kulechov additionally purchased $9.8 million in AAVE from personal funds. The protocol's V4 hub-and-spoke design is intended to expand revenue sources into real-world credit markets, structured lending, and tokenized asset-backed credit.
Uniswap uses a token-jar mechanism: protocol fees are deposited into a smart contract where UNI holders can destroy their tokens to claim a proportional share of accumulated fees. The v3OpenFeeAdapter automates fee capture for new pools, removing the need for individual governance votes.
Lido, while not among the top five by revenue, provides a useful comparison. The protocol proposed a conditional $20 million LDO buyback that activates only when ETH exceeds $3,000 and annualized revenue surpasses $40 million. Revenue fell 23% to $40.5 million in 2025, with its DAO treasury at $157.5 million.
Revenue-to-TVL ratios reveal sharp differences in capital efficiency:
| Protocol | TVL | Annualized Revenue | Revenue/TVL Ratio | |----------|-----|-------------------|-------------------| | Hyperliquid | ~$2B | $843M | ~42% | | Sky | $13B | $611.5M (projected) | ~4.7% | | Aave | $24B | ~$350M (est. 2026) | ~1.5% | | Uniswap | $3.3B | $61M (protocol) | ~1.8% | | Ethena | $5.92B | >$500M (est.) | >8.4% |
Hyperliquid's ratio reflects the leverage-driven nature of perpetual trading — high fee generation per dollar locked. Ethena's model benefits from the spread between perpetual funding rates and stablecoin issuance costs. Aave's lower ratio is typical of collateralized lending, where utilization rates determine the spread.
Pendle merits mention as an adjacent case. The yield tokenization protocol reached $5.8 billion in TVL and $40 million in annualized revenue in 2025, with $47.8 billion in trading volume. Its Boros product for perpetual funding-rate exposure attracted $80 million in initial open interest. Pendle has become the default fixed-rate venue for institutional yield products from Apollo, Paxos, and Ethena.
The buyback wave introduces structural risks that earlier DeFi cycles did not face.
Governance concentration. When protocol revenue flows to token holders proportionally, large holders benefit disproportionately and have stronger incentives to entrench governance positions. Aave's V4 launch was accompanied by internal DAO tensions — a governance vote rejected an IP transfer proposal from Aave Labs, and the second-largest AAVE holder began selling positions. The Aave DAO's decision to route 100% of revenue to itself creates a dynamic where governance power directly translates to economic return.
Regulatory reclassification. Token buybacks funded by protocol revenue bear structural resemblance to stock buybacks or dividend distributions. Under the GENIUS Act framework now being implemented by six federal bodies (OCC, FDIC, NCUA, FinCEN, Treasury, OFAC), regulators are building the infrastructure to classify and regulate crypto-native financial instruments. The OCC comment period on stablecoin rulemaking closed May 1, 2026. If regulators treat buyback tokens as securities, the entire model faces existential legal risk.
Bad debt contagion. Aave's Q2 2026 illustrates how lending protocols remain vulnerable. The $292 million KelpDAO exploit in April left Aave with approximately $200 million in bad debt, triggering heavy withdrawals and a confidence crisis despite the V4 launch. Protocols that distribute revenue rather than building reserves may lack buffers for such events.
Revenue volatility. Hyperliquid's $843 million annualized run rate is a function of leveraged trading volumes that correlate with market volatility. A sustained low-volatility period would compress revenue and buyback capacity, potentially creating a reflexive downward cycle in token price.
The GENIUS Act, enacted July 18, 2025, provides the most immediate regulatory framework affecting DeFi protocol treasuries. Six federal agencies have issued proposed rules between December 2025 and May 2026, with most regulations required by July 18, 2026.
Key implications for protocol revenue models:
The AICPA has urged regulators to adopt its stablecoin attestation criteria in GENIUS Act rulemaking, per a May 2026 submission. This would impose accounting standards on reserve verification that currently do not apply to most DeFi protocols.
DeFi's revenue layer has matured faster than its governance and risk infrastructure. The five largest protocols now generate billions in annualized fees and have adopted mechanisms to return that value to token holders. The economic logic is sound: protocols with demonstrable cash flows trade at higher valuations than those sustained by inflationary incentives.
The unresolved questions are structural. Buyback mechanisms tie token value to protocol revenue, creating reflexive dynamics in both directions. Regulatory frameworks are advancing rapidly — the July 2026 GENIUS Act implementation deadline will force clarity on whether revenue-generating tokens are securities, payment instruments, or something new. And as the KelpDAO-Aave episode illustrates, protocols that prioritize distribution over reserves may find their earnings models tested by the next exploit or liquidity crisis.
The DeFi sector is no longer competing on narrative alone. It is competing on earnings, and that competition will intensify as regulatory and architectural constraints take shape through the second half of 2026.