What the fee-generating assets look like on 06 September 2026

Spread of Market-Cap-to-Fees Ratios

Across the fourteen assets observed on 06 September 2026, the market-cap-to-fees ratio (annualized from the trailing 30 days) spans from 2.56× for Uniswap to 19,715.9× for Bitcoin. Between those poles sit Canton at 7.3×, TRON at 104.9×, Chainlink at 158.2×, and Ethereum at 2,341.6×, among others. A wide ratio means the market-cap figure is large relative to the fee run-rate annualized from the trailing 30 days; it does not by itself indicate anything about quality, growth trajectory, or appropriate valuation. Equally, a narrow ratio does not signal distress or efficiency — it is a arithmetic relationship between two point-in-time or annualized figures, nothing more.

Fee Retention: Take Rate as a Structural Descriptor

Several assets in this dataset pass nearly all fees through to the protocol itself. TRON, Hyperliquid, Canton, and Avalanche each report a take rate of 100% (annualized from the trailing 30 days), meaning fees and revenue figures are identical. At the other end, Uniswap retains only 7.8% of its $1.72 billion in annualized fees as revenue ($135.0 million), with the remainder flowing to liquidity providers. Ethereum retains 23.8%, BNB retains 10.0%, and Chainlink retains 93.7%. These differences reflect structural design choices — whether fee flows are directed to token holders, validators, external service providers, or liquidity suppliers — and make raw fee figures non-equivalent across protocols even when the dollar amounts are similar.

Chain-Basis vs. Application-Basis: A Methodological Boundary

The accounting basis column draws a necessary distinction. Assets marked chain:X — Bitcoin, Ethereum, BNB (BSC), Solana, TRON, Hyperliquid, Canton, TON, Avalanche, and Sui — capture fees paid to the underlying network layer. Assets marked application:X — Uniswap V3, Aave V3, CCIP (Chainlink), and Rain — capture fees generated by a specific application that runs on top of some chain. An application's fee stream does not include the gas costs users also pay to the host chain, and a chain's fee stream does not disaggregate which applications generated which transactions. Placing Canton's $595.8 million in annualized chain fees alongside Uniswap's $1.72 billion in annualized application fees in a single ranked list is arithmetically valid but structurally misleading without this context; the two figures measure activity at different layers of the stack.

Why These Are Not Equity Multiples

The single most important reason not to read market-cap-to-fees as an analog to a price-to-sales or price-to-earnings equity multiple is that token holders have no standardized legal claim on fee cash flows. In equities, revenue and earnings flow through defined corporate structures to shareholders by law or contract. In this dataset, whether fees accrue to token holders, validators, stakers, liquidity providers, or are burned depends entirely on each protocol's own rules — rules that can change through governance. The multiple therefore describes a ratio between two numbers; it does not describe a yield, a claim, or an entitlement.