What the fee-generating assets look like on 08 September 2026
Fee-stream data as of 08 September 2026 reveals wide variation in how digital assets generate, retain, and relate fees to market capitalisation.
Spread of Market-Cap-to-Fees Across the Group
Over the 30 days to 08 September 2026, the market-cap-to-fees multiple (annualized from the trailing 30 days) ranges from 2.4× for Uniswap to 19,513× for Bitcoin, with Chainlink at 162×, TRON at 108×, and Solana at 222× sitting at the compressed end, while Rain at 5,912×, Gram (prev. Toncoin) at 3,665×, and Avalanche at 2,491× occupy the wider end. A wide multiple indicates that a large market capitalisation (point-in-time) corresponds to a comparatively small annualized fee base; it does not by itself indicate mispricing, inefficiency, or any particular trajectory, because market capitalisation prices many factors—monetary premium, optionality, network effects—that fee flows do not capture.
Fee Retention: Pass-Through Versus Capture
The take-rate figure (revenue as a share of fees, annualized from the trailing 30 days) separates assets that retain most fees from those that distribute them to other parties. TRON, Hyperliquid, Canton, and Avalanche each show a take rate of 100%, meaning their reported revenue equals their reported fees over this period. Chainlink records a take rate of 93.8%. At the other end, Uniswap retains 8.0% of its $1.85 billion annualized fee base as revenue, passing the remainder to liquidity providers. Aave retains 13.9% of its $404 million annualized fees. Rain retains 60.7%. Bitcoin and Ethereum's take-rate figures reflect structurally different dynamics: Bitcoin's revenue figure is listed as null, while Ethereum's take rate stands at 23.3%, reflecting the share of fees burned or otherwise captured versus paid to validators. These differences mean that comparing raw fee figures across assets without accounting for retention can be misleading.
Chain-Basis Versus Application-Basis: Non-Comparability
Several entries in this dataset are accounted on a chain basis (Bitcoin, Ethereum, BNB, Solana, TRON, Hyperliquid L1, Canton, TON, Avalanche, Sui), while others are accounted on an application basis (Uniswap V3, Aave V3, CCIP, Rain). Application-basis figures capture only the fees flowing through a specific protocol deployed on a chain, not the underlying chain's total activity. Chain-basis figures aggregate all activity settled on that chain. Comparing Uniswap's $1.85 billion annualized application-basis fees directly against Ethereum's $134 million annualized chain-basis fees, for instance, does not imply that the application is larger in an equivalent sense; the accounting scopes are categorically different and must be read within their own columns.
Why These Multiples Are Not Equity Multiples
The single largest reason a reader should not treat market-cap-to-fees ratios as equity price-to-revenue multiples is that token holders in most of these networks have no enforceable legal claim on fee flows. Fees accrue to validators, liquidity providers, stakers, or are burned—depending on each protocol's design—and token market capitalisation reflects a bundle of rights, uses, and expectations that differs fundamentally from a residual equity interest. The accounting basis (annualized from the trailing 30 days) further means the fee figure extrapolates a short window, amplifying any transient activity spikes or lulls into the annual figure.
periods{'market_cap': 'point-in-time', 'fees_and_revenue': 'annualized from the trailing 30 days'}
observation date08 September 2026
Written by an AI language model from the figures listed above and nothing else. It contains no price view and no recommendation. Figures carry the reference periods shown on their own pages.
· claude-sonnet-4-6