What the fee-generating assets look like on 02 September 2026
Fee-stream data as of 02 September 2026 reveal wide variation in how fourteen digital assets generate and retain on-chain revenue.
Market-Cap-to-Fees: The Spread
Across the fourteen assets observed on 02 September 2026, the market-cap-to-fees ratio (annualized from the trailing 30 days) ranges from 2.89x for Uniswap to 19,246x for Bitcoin. That is a spread of roughly four orders of magnitude within a single table. A wide ratio indicates that, relative to its point-in-time market cap, an asset generates comparatively little in annualized fees; a narrow ratio indicates the opposite. What the ratio does not imply is anything about trajectory, sustainability, or whether fee generation is likely to expand or contract. It is a snapshot descriptor, not a forecast.
Fee Retention: Pass-Through vs. Capture
Several assets in this dataset retain all fees they collect. TRON, Canton, Hyperliquid, and Avalanche each show a take rate of 100% (annualized from the trailing 30 days), meaning annualized revenue equals annualized fees in every case. At the other end, Uniswap retains only 8.26% of its $1.35 billion in annualized fees, passing the remainder to liquidity providers. Solana retains 9.45% and BNB retains 10.00%. Chainlink, by contrast, retains 92.96% of its annualized fees. These differences reflect structural design choices — how a protocol allocates fee flows between its treasury or token holders and the external parties who provide services on its network. Comparing retained-revenue figures across assets with very different retention architectures without accounting for that structure can produce misleading impressions.
Chain Basis vs. Application Basis: A Critical Distinction
The accounting basis column separates chain-level figures (e.g., chain:Bitcoin, chain:Ethereum, chain:Solana) from application-level figures (e.g., application:Uniswap V3, application:Aave V3, application:CCIP). Chain-basis fees aggregate all economic activity settled on that network; application-basis fees capture only what flows through a specific protocol running on top of some chain. An application's fees are therefore a subset of — and already counted within — the host chain's fee total. Placing Uniswap's $1.35 billion annualized fees alongside Ethereum's $126.5 million annualized fees in the same column does not mean Uniswap generates more network value than Ethereum in an equivalent sense; it means two different scopes of measurement appear in the same field. Readers must treat the two groups separately and resist cross-group ratio comparisons.
Why These Are Not Equity Multiples
The single most important reason not to read market-cap-to-fees as an equity-style price-to-earnings multiple is that fees paid to a network are not net income accruing to token holders. Token holders in most of these networks do not hold a residual legal claim on fee cash flows in the way equity shareholders hold a claim on after-tax earnings. Fee revenue may be burned, redistributed to validators, held in protocol treasuries, or directed to liquidity providers — each with different economic implications for token value. Without knowing the precise flow of fees to and from the token, applying an earnings-multiple interpretive framework to these ratios introduces a category error that the figures themselves cannot resolve.