The market
The protocols in this study have collected $86.5bn in fees since 2018. The busiest single month was January 2025 at $3.3bn; the most recent complete month, September 2026, came in at $2.5bn, 76% of that peak.
Against its own history the market fee multiple sits at the 38th percentile of the last 208 weeks. The index is chain-linked, its constituent count rising from 10 at the start of the window to 60 today. The range over that window runs from 82 to 336, against 176 today. That is a statement about a ratio and its own past, and about nothing else.
Against its own history the market fee multiple sits at the 38th percentile of the last 208 weeks. The index is chain-linked, its constituent count rising from 10 at the start of the window to 60 today. The range over that window runs from 82 to 336, against 176 today. That is a statement about a ratio and its own past, and about nothing else.
Multiples by category
Across 507 tokens with a computable multiple, the median price-to-fees is 25.4. The mean is 61,692, which is 2,430 times the median, and that is why neither it nor a standard deviation appears anywhere in this report. What follows is the middle half of each category and its median.
The spread between categories is the first thing to read. Dexs sits at a median of 18.3 across 134 projects, while Chain sits at 7,464.6. No single number describes the crypto multiple, and a benchmark drawn from the wrong category is worse than none.
The spread between categories is the first thing to read. Dexs sits at a median of 18.3 across 134 projects, while Chain sits at 7,464.6. No single number describes the crypto multiple, and a benchmark drawn from the wrong category is worse than none.
Where the fees actually go
A protocol does not keep the fees it collects. Of 507 projects collecting fees, 133 keep less than a tenth of them and 164 keep more than nine tenths. The middle is thin. This is not a spectrum with most of the market in the centre; it is two business models sharing one metric, and which one a project is running changes what its fee line means.
The consequence for valuation is direct. Price-to-fees and price-to-revenue are the same number for a protocol that retains everything, and differ by an order of magnitude for one that passes its fees to liquidity providers. Comparing the two across that divide compares nothing.
The consequence for valuation is direct. Price-to-fees and price-to-revenue are the same number for a protocol that retains everything, and differ by an order of magnitude for one that passes its fees to liquidity providers. Comparing the two across that divide compares nothing.
The outliers
Every project on the chart below is measured against the median of its own category, so a chain is judged against chains. The distances are the finding: the dearest project in most categories trades at hundreds or thousands of times its peer median, which is another way of saying that for much of this market the fee line is not what sets the price.
Credits and Disclaimers
Data Source: Protocol fees and revenue are sourced from DefiLlama; token prices, market capitalisation and supply from CoinGecko. Both are read through their public APIs. Historical diluted value is reconstructed as price multiplied by maximum supply, which does not change, so no supply history is required.
Methodology: Price-to-fees is fully diluted market capitalisation divided by gross fees over the last 30 days, annualised. Price-to-revenue uses the portion of those fees the protocol retains. A protocol with no token has no multiple and is excluded rather than counted as zero. Categories are DefiLlama's own, taken from the parent protocol so that a project is not filed under one of its adapters.
On medians rather than averages: the distribution of multiples is not one a mean describes. Across the set the mean is several thousand times the median and the standard deviation is larger than the mean, so both are reported nowhere in this document. Medians and interquartile ranges are used throughout, and the number of projects behind every figure is stated.
On the chain-linked index: each week's move is measured only over the tokens present in both that week and the one before, and those moves are multiplied together. Membership grows as protocols accumulate history, and a token joining never moves the level, so a change in the index is a change in the market rather than a change in who is being measured. The constituent count is stated with the chart and on every row of the underlying series. The whole history is rebuilt each month from the sixty protocols earning the most fees at that time, and scaled so the latest week equals that week's own aggregate. Past weeks are therefore restated from one issue to the next, and a figure in this issue is the series as it stood when the issue was produced.
Accuracy Disclaimer: While every effort has been made to ensure the accuracy of the information presented in this report, FinDaS does not guarantee the completeness, accuracy, or reliability of the data. Users are advised to independently verify the information before making any business decisions based on the content of this report.
Not investment advice: This report describes ratios and their history. Nothing in it is a recommendation to buy, sell or hold any asset, and it does not take account of any reader's circumstances.
Copyright Notice: This report is the intellectual property of FinDaS and is protected by copyright law. Unauthorized reproduction, distribution, or sharing of this report, in whole or in part, without the explicit permission of FinDaS is strictly prohibited. All rights reserved.
Limitation of Liability: FinDaS shall not be held liable for any errors, omissions, or inaccuracies in the data presented in this report, nor for any actions taken in reliance thereon.
