What the fee-generating assets look like on 22 September 2026

The Spread in Market-Cap-to-Fees Ratios

As of 22 September 2026, the market-cap-to-fees multiples across the fourteen assets in this dataset span from 2.34x (Uniswap, annualized from the trailing 30 days) to 21,145x (Bitcoin, annualized from the trailing 30 days)—a range of roughly four orders of magnitude. A wide multiple indicates that a large amount of market capitalization is supported by a relatively small annualized fee base; a narrow multiple indicates the opposite. What such spreads do tell a reader is the arithmetic relationship between two point-in-time or period-averaged figures. What they do not tell a reader is anything about causation, sustainability, or whether either figure is appropriate given the asset's design, security model, or token supply mechanics.

Fee Retention: Take-Rate Variation

Several assets in this dataset retain all fees they generate. TRON, Avalanche, Canton, and Hyperliquid each record a take rate of 100% (annualized from the trailing 30 days), meaning fees paid on-chain are also recognized as protocol revenue. At the other end, Uniswap's take rate stands at 8.1% (annualized from the trailing 30 days), with the large majority of the $2.36 billion annualized fees (annualized from the trailing 30 days) flowing to liquidity providers rather than the protocol itself. Chainlink sits at 95.4% (annualized from the trailing 30 days), while Ethereum retains 23.2% (annualized from the trailing 30 days) of its fee base. These differences reflect fundamentally different economic architectures: a low take rate is not inherently preferable or detrimental, but it does mean that protocol-level revenue and gross fee volume must be read as distinct figures rather than proxies for one another.

Chain-Basis Versus Application-Basis Figures

Four entries—Chainlink, Rain, Uniswap, and Aave—are measured on an application basis, while the remaining ten are measured on a chain basis. This distinction is methodologically significant. A chain-basis figure aggregates all transaction fees paid to validators or block producers on that network; an application-basis figure captures only the fees attributable to one protocol deployed across one or more chains. Comparing, for example, Canton's $607 million annualized revenue (annualized from the trailing 30 days) on a chain basis to Aave's $61.9 million annualized revenue (annualized from the trailing 30 days) on an application basis conflates two different scopes of economic activity. Readers should treat these two sub-groups separately and resist aggregating or ranking them in a single fee-efficiency table without first adjusting for scope.

Why These Are Not Equity Multiples

The single most important reason market-cap-to-fees ratios cannot be read as equity price-to-earnings multiples is that token holders do not possess a standardized legal claim on fee cash flows. In equities, earnings per share represents a defined residual interest; in digital assets, whether, how, and to whom fees accrue is governed by protocol rules that vary by asset, can be altered by governance, and do not constitute ownership in any jurisdictional sense recognized uniformly across markets. The arithmetic of the ratio is identical in form to a P/E calculation, but the underlying legal and economic substance is not.