What the fee-generating assets look like on 03 September 2026
Fee-stream data as of 03 September 2026 reveals wide dispersion in market-cap-to-fees multiples across fourteen digital assets, with methodological caveats demanding careful interpretation.
Spread of Market-Cap-to-Fees Ratios
Across the fourteen assets observed on 03 September 2026, the market-cap-to-fees ratio (annualized from the trailing 30 days) spans from 2.46x for Uniswap to 19,338x for Bitcoin—a range of roughly four orders of magnitude. Canton sits at 7.23x, TRON at 99.07x, and Chainlink at 143.12x, while Ethereum registers 2,315.81x and Gram (prev. Toncoin) 3,582.98x. A wide ratio indicates that market capitalization is large relative to the fee base annualized from the trailing 30 days; it does not by itself indicate anything about whether that relationship is appropriate, sustainable, or anomalous. The trailing-30-day annualization window means a single atypical month can compress or inflate the denominator substantially, so the ratio is sensitive to the observation period chosen.
Fee Retention: Take Rate Divergence
Assets differ sharply in how much of collected fees they retain as protocol revenue. TRON, Canton, Hyperliquid, and Avalanche each show a take rate of 100%—fees annualized from the trailing 30 days equal reported revenue annualized from the trailing 30 days—meaning no measurable fee share is passed to external parties in the accounting basis used. Chainlink retains 92.96% and Rain 60.67%. At the other end, Uniswap retains only 7.92% of its fee stream as revenue ($114.96 million annualized from the trailing 30 days against $1.45 billion fees annualized from the trailing 30 days), with the remainder flowing to liquidity providers. Solana retains 9.58% and BNB 10.00%. Bitcoin and Gram carry no revenue figure in the dataset. This distinction matters: comparing market-cap-to-fees across high- and low-retention assets conflates structurally different economic models.
Chain-Basis vs. Application-Basis: Non-Comparability
The dataset explicitly labels each asset's accounting basis. Layer-1 chains such as Bitcoin (chain:Bitcoin), Solana (chain:Solana), and Canton (chain:Canton) capture fees at the settlement layer—every transaction on that network contributes. Application-layer entries such as Uniswap (application:Uniswap V3), Aave (application:Aave V3), and Rain (application:Rain) capture only fees flowing through a specific protocol deployed on top of some chain. A chain's fee total therefore includes activity from many applications simultaneously, whereas an application's fee total is a subset of its host chain's activity. Placing these side by side in a ranked table can suggest false equivalence; the figures answer different questions about different scopes of economic activity.
Why These Are Not Equity Multiples
The single largest reason market-cap-to-fees figures cannot be read as equity price-to-sales or price-to-earnings multiples is that token holders do not possess a standardized legal claim on fee cash flows. In equity markets, the numerator (market capitalization) and denominator (revenue or earnings) are linked by enforceable ownership rights and accounting standards. In digital assets, the relationship between holding a token and receiving any portion of protocol fees depends entirely on each protocol's specific—and variable—governance and distribution mechanics, none of which are harmonized across this dataset. The multiple therefore describes a ratio of two observable numbers; it does not describe a valuation in the equity-finance sense.