Where DeFi's Harder Assets Trade: Q3 2026 DEX Liquidity Report

Depth and trading do not move together on decentralized exchanges. In the asset classes we measure, the typical major trades through its ±1% depth many times a day. ETH liquid staking tokens and the largest wrapped bitcoin book hold depth that is a multiple of a day’s trading. For majors, liquidity is used constantly. For those books it is parked and waits for a counterparty.

This report reads that split across eight asset classes, then looks at a second pattern: books whose depth sits mostly on one side of the price. Every figure is in a table with its date.

How to read these tables

• Catalog snapshot. Table 1 comes from the catalog as it stood on 17 August 2026.
• Depth. The depth values are the API’s aggregated ±1% depth in USD for the newest collector observation of each pair. They are delayed. Reads that come back far below a pair’s normal level are left out, never corrected. The observations fall between 27 August and 8 September 2026.
• Bid side and ask side. The two halves of ±1% depth, one on each side of the mid price.
• Volume. Seven-day on-chain volume of the measured pools, not exchange-wide volume. Turnover divides two-sided ±1% depth by that volume averaged per day. Books with no reported seven-day volume (a value of zero in the data) are left out of turnover, while the volume medians in Table 2 include them.
• Aliases. Some books appear under two names with identical bid and ask: wrapped natives, bridged copies and renamed tokens. They are shown once, folded into the base book.
• Overlap. On-chain BTC/USD is read from wrapped bitcoin pools, so its book overlaps the wrapper books in Table 3.
• Medians. A class with fewer than three USD books gets no median in Table 2.

A “book” here means the measured on-chain liquidity for one asset priced in USD. The methodology page describes how the underlying readings are produced.

What the catalog covers

Table 1 sets the frame. The classes differ a lot in how many pairs they hold. They also differ in how many of those pairs have live pool data on the snapshot date.

Table 1. The catalog by asset class (snapshot 17 August 2026)

Base assets count each asset once, whatever the spelling of its symbol, as the site’s category pages count them.

Dollar stablecoins make up the largest group by a wide margin. The stablecoin prices hub lists them. Most classes have live pool data for the large majority of their pairs. The gap between the two columns is the pairs without live pool data on that date.

Depth and turnover by class

Table 2 reports the median USD book in each class. Read the columns together: the depth tells you how much a book can absorb near the price, the volume tells you how much trading it sees and the turnover relates the two.

Table 2. Depth and turnover by asset class (observations 27 August to 8 September 2026)

Three groups show up.

Traded through. The majors class has the highest median volume of any class and a median turnover far below one. A day of trading is many times the depth. DeFi tokens also sit below one, though for a different reason: both their depth and their volume are at the low end of the table. Wrapped bitcoin and dollar stablecoins sit close to one, so a day of trading and the standing depth are of similar size.

Parked. ETH liquid staking has a median turnover several times above one. Its median volume is among the lowest in the table. Depth there is a multiple of what trades in a day. Yield-bearing and RWA assets and Solana liquid staking sit between the two groups, with turnover a few times above one.

Median versus headline books. The median book in a class and the books that make headlines are different objects. The majors median depth is lower than the dollar stablecoin median, yet ETH/USD and BTC/USD both carry eight-figure depth on each side (Table 3). The class median describes the typical USD book in the class, which includes many tokens that are not ETH or BTC.

The same two numbers matter for anyone sizing a position against a book. Depth tells you what the book absorbs at a given moment. Turnover tells you whether that depth is replenished by trading or just waits. Price impact and slippage explains how depth converts into the cost of a trade.

The largest books in each class

Table 3 lists the three USD books with the most ±1% depth in each class. Where a class has fewer than three books, all of them appear.

Table 3. The three USD books with the most ±1% depth in each class (observations 27 August to 8 September 2026)

Class by class:

• Wrapped and staked BTC. All three books carry eight-figure depth on each side, yet they sit at opposite ends of the turnover scale. TBTC/USD has the most depth and the least volume of the three, so its depth is many times a day’s trading. cbBTC/USD is the reverse: the least depth and by far the most volume. Compare the books on the wrapped bitcoin prices hub.
• Euro stablecoins. The class has two USD books, so there is no class median. EURC/USD has seven-figure depth on both sides and a day’s trading larger than that depth. EURA/USD has six-figure depth, weekly volume of a similar order of magnitude and a turnover many times above one.
• Dollar stablecoins. The two largest books carry nine-figure depth and volume among the highest in the table, with turnover close to one for USDT/USD and lower for USDC/USD. AUSD/USD has far less volume and a turnover several times higher.
• ETH liquid staking. All three books carry eight-figure depth on the bid side. In each, depth is a multiple of a day’s volume. OETH/USD sits at the extreme: depth is a very large multiple of what trades in a day. The liquid staking token prices hub covers the class.
• Yield-bearing and RWA. Turnover runs from just above one to several times above it. The sUSDe/USD book is lopsided, a pattern Table 4 picks up.
• Solana liquid staking. JITOSOL/USD has eight-figure depth, nearly all of it on the bid side. The other two books have six-figure depth.
• Majors and L1/L2 tokens. ETH/USD trades through its depth several times a day and SOL/USD more than ten times, BTC/USD about once. BTC/USD overlaps WBTC/USD because the on-chain BTC price is read from wrapped bitcoin pools, which is why the two rows look alike.
• DeFi tokens. Two of the three books have no reported volume, so they have no turnover. JLP/USD trades through its depth many times a day.

One-sided books

A book can show healthy two-sided depth in total and still offer very different amounts on each side. Table 4 lists the most lopsided books among those with at least $1M of two-sided depth.

Table 4. The most one-sided books with at least $1M two-sided depth (observations 27 August to 8 September 2026)

Half of these books lean to the bid side and half to the ask side. The yield-bearing and staking books at the top lean heavily toward the bid side. Among the dollar stablecoins the direction varies: GHO/USD leans to the bid side, while USDf/USD, USDS/USD, JUPUSD/USD and USX/USD lean to the ask side.

A lean can reflect pool design, or a redemption path that gives holders a route out of the token other than the pool. The data does not say which applies to any of these books. A lean carries no judgement of the asset, but it does change what a trade of a given size costs: the same two-sided total gives a different outcome depending on which side you need. The feed pages show the bid and ask sides of ±1% depth.

What to take from the numbers

Two questions get mixed up: how much a book can absorb and how much it trades. A risk desk sizing a liquidation, or a protocol choosing which market a price should be read from, needs both, plus the side. Turnover alone does not rank books. A high value can mean parked liquidity or a quiet market. A low value can mean busy trading. Depth, volume and side together separate those cases.

For the method behind the readings, DEX-native price feeds, explained covers what an on-chain price measures. The coverage page shows which pairs are in the catalog today.