How much revenue reaches the token

Nearly all protocol fees route to the Assistance Fund, which continuously buys HYPE on the open market and burns it. Holder revenue is effectively the whole take.

One of the strongest value-accrual designs in crypto. Token Terminal shows about 16.7x, so the low multiple is genuinely not misleading here.

Trading fees flow to veAERO lockers who vote for the fee pools, a direct dividend. The team keeps nothing from fees; passive holders get nothing, only lockers who vote.

The earlier 7.9x was stale. On trailing 30-day revenue Aerodrome is nearer 43x, since fee revenue cooled while the veAERO dividend still takes essentially all of it.

Since the December 2025 UNIfication, protocol fees accumulate on-chain and can only be claimed by burning UNI, the first sustained burn in Uniswap history. Votes are underway to extend it to v4 and Robinhood Chain.

Fees run past $5M a day, but UNI captures only the roughly 17% protocol slice, near $48M a year, so it trades about 47x that. Live DEX share is near 36%, not the 45 to 55% often quoted.

Essentially all protocol revenue funds open-market LIT buybacks, with a burn leg added July 1, 2026.

Recent run-rate passthrough is ~97%, but the full year is nearer 41% since buybacks ramped after TGE, and the burn is brand new with the first burn still pending.

12% of all trading fees buy back RAY. Of the revenue Raydium keeps (buyback plus a 4% treasury cut), roughly 80% goes to the buyback.

The 80.5% is the buyback share of protocol revenue, not of total fees. About $196M cumulative buybacks, ~26% of supply.

About 65% of protocol revenue converts to open-market CAKE buybacks that are burned, a deflationary passthrough. The rest goes to treasury.

DefiLlama puts the buyback share at 65.0%, matching the claim. On trailing-year revenue the multiple is nearer 4.4x.

50% of protocol revenue buys back JUP into the Litterbox trust, locked for three years. The other 50% funds the team treasury in SOL and USDC.

Over 113M JUP accumulated in the Litterbox. A governance proposal to lift the buyback to 70% is under debate.

50% of net revenue from the core products auto-routes to open-market PUMP buybacks that are burned immediately through a locked contract.

A rare case where a very low multiple is genuinely backed by revenue reaching the token. Over $218M bought back, about 36% of supply burned.

Aavenomics 3.0 went live in 2026 as an immutable, non-discretionary buyback that routes protocol and GHO revenue to AAVE. The budget was cut from about $50M to $30M a year.

The share reaching the token is low, near 14%, but the mechanism is now continuous and automatic, so the post's non-continuous label is out of date.

A buyback plus staking rewards route about 8.8% of protocol revenue to SKY holders. DefiLlama confirms holders revenue of $1.11M against $12.66M revenue over 30 days.

Honest framing. The 8.1x looks cheap, but with under 9% reaching the token, most of the revenue never touches SKY.

Around 80% of Pendle V2 fees buy back PENDLE, with up to all of the repurchased tokens paid to active sPENDLE stakers.

A genuine high-passthrough case, so the multiple is not misleadingly low. Passive holders capture less than sPENDLE stakers, and revenue has been softening.

The Chainlink Reserve uses Payment Abstraction to convert onchain and enterprise revenue into LINK, then holds it in a strategic reserve rather than burning or paying it out.

Real revenue capture, but the multiple is the opposite of cheap. DefiLlama shows about 100x price-to-onchain-revenue, so the token is priced well ahead of the cash flow it captures.

Half of Curve's trading fees plus most crvUSD interest go to veCRV lockers as a weekly crvUSD dividend. The other half of trading fees is LP revenue, so gross fees overstate what the token captures.

You have to lock CRV into veCRV to earn, and unlocked CRV earns nothing. The multiple is on the roughly 50% net revenue, not gross fees.

About 60% of participating dApp fees fund a weekly burn auction where bidders pay INJ, and the winning INJ is burned. Real deflation, but on a small fee base.

Only about 60% of dApp fees enter the auction, so gross fees overstate it. Protocol revenue is near $3.3M a year against a roughly $482M cap, so the multiple stays high.

Liquity pays 100% of fees to users, but in V2 most of that goes to Stability Pool depositors and PIL, not directly to LQTY stakers. Direct LQTY holder revenue is now small.

The 100%-to-users line is true, but it does not all go to LQTY stakers. The direct LQTY fee stream is a residual V1 flow, about $115K a year.

EIP-1559 burns the base fee, which benefits every holder, but issuance to stakers currently exceeds the burn. Since Dencun moved activity to L2s, ETH turned net inflationary.

ETH is not deflationary ultrasound money right now. The daily base-fee burn is far below staking issuance, so net supply is rising.

Avalanche burns 100% of C-Chain base and priority fees, real deflation, but the burn is small next to issuance, so net supply still grows.

The full fee burn is genuine, but only about $56M a year against a roughly $2.9B cap, and below issuance, so net value capture is still negative.

Convex takes a 17% cut of the CRV rewards its users earn. Most of that goes to cvxCRV and vlCVX stakers, and vote-locked CVX also collects external vote bribes.

The 11x is on Convex's DefiLlama revenue and leaves out Votium vote bribes, which are often the larger flow to vlCVX, so real value to the token is understated.

Since YIP-88 in October 2025, 90% of Yearn revenue is auctioned into yield-bearing stablecoins and streamed to stYFI stakers as a continuous dividend. 10% goes to the DAO.

The old buyback-and-burn is retired. Value now reaches the token as a real dividend, but revenue is small, so the multiple is high near 165x.

Protocol yield from Frax's products goes to veFXS lockers and buybacks, so nearly all captured revenue reaches the token.

Note the ticker. FXS was renamed to FRAX one-to-one in January 2026 under the North Star hardfork, so this is now the single FRAX token. Revenue is small against the cap, so about 85x.

Rocket Pool was pure node-operator collateral until the Saturn One upgrade in February 2026 added a Voter Share, routing about 9% of ETH commission to staked RPL.

The share is new and small, and DefiLlama still logs $0 protocol revenue for the token because its fee adapter does not track it yet, so treat the flow as early.

Users lock CETUS into xCETUS to earn a weekly fee dividend paid in SUI, CETUS and USDC. About half of net protocol profit goes to xCETUS, the other half services a Sui Foundation loan.

The other half of profit repays the loan taken after the 2025 exploit, so the token's share is temporarily capped. Gross swap fees are far larger than the protocol take.

Under DragonFi 2.0, QuickSwap uses about 40% of protocol revenue to buy back and burn QUICK. The rest funds LP farming and treasury.

The old dQUICK fee share is largely deprecated. On the roughly 40% the token captures, the multiple is near 5x.

Buybacks are active, not paused. 100% of eETH withdrawal fees buy ETHFI weekly and pay it to sETHFI stakers, plus a share of broader revenue.

Per ether.fi docs the sETHFI buyback is active, but DefiLlama has logged no holders revenue over the past 30 days, so recent flow to the token is not visible on-chain even though revenue runs near $33M a year.

Discretionary treasury buybacks from a single wallet. A $10M buyback took about 2.3% of supply and ran to 88% of quarterly revenue.

Active for now, but discretionary and revenue is softening, so the pace may not hold. The 2.3% per quarter roughly offsets tokens that vest.

USDe yield goes to sUSDe stakers and the reserve fund, not to ENA. A fee switch was approved, but the split is still undecided and nothing flows to sENA yet.

The large 2025 buybacks were treasury-funded one-offs. As of mid-2026 no continuous protocol revenue reaches ENA, and Ethena Labs raised venture equity above the token.

Since a November 2025 vote, 75% of net protocol fees buy DYDX on the market and stake it. Most value reaches the token, but total revenue is small, near $4M a year.

High pass-through on a tiny revenue base, so the multiple stays high even though most of the revenue reaches the token.

Synthetix dropped fee dividends for a buyback in 2026, but revenue has collapsed to near zero and sUSD was retired after depegging, so almost nothing reaches SNX today.

The buyback exists on paper, but DefiLlama shows revenue near $0 over recent months, and SNX is down about 99% from its high.

BNB has no emissions and a fixed target supply, so all value comes from burns. A slice of BSC gas is burned in real time, and Binance funds large quarterly auto-burns.

The headline billion-dollar quarterly burns are funded by Binance, not by BSC gas revenue. BNB is tied to a private exchange company, so equity sits above the token.

A 12-month pilot that started in February 2026 uses about half of net Superchain sequencer revenue to buy OP on the market. The rest funds the Collective and public-goods programs.

New and time-limited. Before 2026 OP captured nothing, and OP Labs raised large venture equity, so equity sits above the token.

After reserving treasury, Spark uses 10% of monthly surplus to buy SPK on the market. A modest but real share of protocol surplus reaches the token.

SPK here is Sky's SparkLend, a different project from the Lido-adjacent Spark. About $3.9M a year of buyback against roughly $25M revenue, so near 16% reaches SPK.

Fluid's on-chain buyback used up to 100% of protocol revenue to buy FLUID into a reserve, where governance decides what to do with it. It is not a direct dividend.

The buyback has been paused since a March 2026 exploit left the DAO rebuilding its treasury, so revenue currently goes to the treasury, not the token. On DefiLlama revenue the multiple is about 13x.

The DAO can direct swap and withdrawal fees to on-chain EKUBO buybacks, then optionally hand the tokens to stakers. Real buybacks have happened, but the flow is discretionary.

Ekubo Inc currently favors spending revenue on growth until the protocol is self-sustaining, so live flow to the token can run below the buyback share. Gross fees are far larger than the protocol take.

Euler's Fee Flow auctions accrued fees for EUL, which lands in the DAO treasury, not with holders. Governance then decides whether to burn or spend it.

DefiLlama shows $0 holders revenue over 30 days, about $2.5M over a year, so the flow is intermittent. The widely quoted $33M annual EUL buyback is a projection, not actuals. Euler Labs raised venture equity.

Lido's 10% staking fee splits roughly evenly between node operators and the DAO treasury. Only a conditional buyback capped at $10M a year, and only when ETH is above $3,000 and revenue above $40M, reaches LDO, and it lands in a DAO-owned LP rather than being burned.

DefiLlama logs about $2.5M a month of holders revenue, but that is the conditional buyback landing in a DAO-owned LP, not a direct distribution to LDO. Most Lido value still goes to stakers and node operators.

A CIP-38 buyback converts fees to COW, but it is sized to offset solver emissions rather than to distribute value. DefiLlama logs zero holders revenue.

The buyback is real but roughly neutralizes emissions, so net value reaching holders is near zero today. The 4.7x on market cap looks cheap only if you ignore that.

A governance-approved buyback-and-burn routes protocol-owned-liquidity treasury fees into WLFI. Over $19M has been spent and burned.

A mechanism does exist, so the no-mechanism claim is wrong, but it only routes POL fees. Grayscale puts the multiple nearer 17x than 13.7x.

A discretionary CARDS buyback-and-burn started in June 2026, but it is small, about 3.4% of net revenue to date.

CARDS here is Collector Crypt, not Cardstarter. Both post numbers are off. The 0.4x uses inflated gross revenue; on net revenue the FDV multiple is about 7x, and a small buyback does exist.

The fee switch is off, so the protocol takes no revenue for itself. MORPHO earns nothing today, even though the markets generate about $21M a month in interest that all goes to lenders.

DefiLlama shows $0 protocol revenue and $0 holders revenue. Morpho Labs folded 100% into a nonprofit in 2025, so no equity class sits above the token, but with the fee switch off the token still captures nothing. A rare case where the structure is clean and the accrual is zero.

The roughly $67M a year in management fees from USDY and OUSG go to Ondo Inc and the DAO treasury. ONDO is a governance token with no buyback or dividend.

The 23x price-to-fees figure that circulates is on gross fees the token never receives. ONDO captures nothing today, and Ondo Inc raised venture equity.

All Arbitrum sequencer and Timeboost fees go to the Arbitrum DAO treasury in ETH. ARB is a governance token and captures no protocol revenue.

Arbitrum earns real sequencer revenue, over $100M a year, but it accrues to the DAO treasury, not to ARB. Offchain Labs is a venture-backed private company.

Solana burns only half of base fees, which is negligible, and priority fees now go entirely to validators with no burn. Roughly 4% annual issuance means net value capture is dilutive.

Headline network revenue is large, but almost none of it reduces SOL supply. Net issuance is positive, so holders are diluted rather than paid.

Borrower interest funds each market's reserves, a governance-controlled backstop. None of it reaches COMP holders. COMP is a pure governance token.

A stCOMP staking product that would share reserves is proposed but not confirmed live. The DefiLlama revenue figure is reserves retained by the DAO, not value to the token. Compound Labs raised venture equity.

1inch charges no protocol fee on standard aggregator swaps, so there is no ongoing revenue routed to the token. Staking is governance only, through Unicorn Power.

DefiLlama has no fees module for 1inch. Resolver rewards go to whitelisted resolvers, not to passive 1INCH holders.

GFI is a governance and staking-incentive token. Lending interest goes to liquidity providers and reserve fees to the treasury, with nothing routed to GFI holders.

GFI does not capture RWA lending revenue, a common misread. Warbler Labs raised venture equity above the token.

Huma earns borrower fees on receivable financing, but value accrual to the token is still being explored and is not live.

Explainer sites present a 50% fee buy-and-burn as if it were active. Huma's own blog says it is not confirmed. Huma raised venture equity.

EigenLayer's fees are earned by stakers, operators and AVSs. DefiLlama logs protocol revenue of $0, so nothing reaches the EIGEN token today.

A big gross-versus-net trap. Fees run near $90M a year, but the protocol take to the token is zero, and Eigen Labs raised large venture equity.