Base strengthened its position by winning trading from other networks, not by following a general rise, and that win settled in fewer venues and fewer versions, leaving anyone swapping tokens there dependent on a small set of routes.
Base: daily trading on exchanges, last 90 days ($ million)
Trading volumeSeven-day average
Data: DefiLlama.Base: exchanges by share of the latest day ($ million)
Aero84956%
Uniswap39526%
PancakeSwap18112%
Hanji Protocol18.71.2%
Kipseli15.61.0%
Other55.43.7%
Data: DefiLlama.
DefiLlama put the reading at 14 November 2026. The aggregator tracks daily dollar trading across the decentralised exchanges it covers on Base, the layer-2 network from Coinbase. A decentralised exchange lets people swap tokens through software instead of a central broker. Trading volume is the dollar value of those swaps, and share is Base's portion of all tracked trading of that kind.
A larger slice of tracked trading
The network held 15.5% of tracked decentralised-exchange trading, against a 12.2% level across the preceding thirty days. Trading through its venues reached $1.51 billion, 48% above a thirty-day average of $1.02 billion; the seven-day average stood at the same level. Thirty days earlier, trading had reached $1.30 billion, so the latest figure was 16.6% higher.
That pairing separates a shift in location from a general increase in activity. If trading had risen everywhere at the same pace, Base's portion would have stayed close to its recent level. Instead the portion rose while the dollar total also rose, which points to counterparties gathering on Base on that occasion rather than simply trading more across every network.
Three venues absorbed the gain
Aero led Base trading with $848.58 million, or 56.0% of the total, against 54.9% across the preceding thirty days. Its preceding seven days had risen 28.2% on the seven before them. Uniswap followed with $394.75 million, or 26.1%, against 22.3% in the thirty-day comparison, after a 19.4% weekly rise. PancakeSwap added $181.12 million, or 12.0%, against 9.5% for the thirty days before it.
Base: share of trading, latest day against thirty days (%)Data: DefiLlama.
Together those three accounted for almost all of the day's trading, and each gained share against its recent level. The weekly rises before the spike show the movement was already under way at different speeds, rather than appearing from a single print. A week remains short, and the smaller venues show that momentum did not reach every exchange.
Hanji Protocol recorded $18.73 million, with its preceding week down 7.1% on the seven before it. Kipseli recorded $15.61 million, after rising 28.9% week on week. Metric V1 recorded $11.14 million, after falling 65.8% week on week. Scale fell sharply below the top three, and direction split.
One version inside each house took over
Inside Aero, Aero Slipstream took 86.1% of the project's Base trading, up from 62.7% over thirty days. Aerodrome Slipstream fell to 12.9% from 36.0%. Aero Basic held 0.6%, against 0.5% before, and Aerodrome V1 held 0.4%, against 0.8% before.
Uniswap V3 took 92.2% of Uniswap trading on Base, up from 82.2% across thirty days. Uniswap V4 dropped to 7.8% from 17.7%. The remaining versions together accounted for almost nothing, which leaves the parent total dominated by a single version.
PancakeSwap AMM V3 took 99.9% of PancakeSwap trading on Base, up from 99.3% over the preceding thirty days. PancakeSwap Infinity fell to 0.1% from 0.7%. The other versions stood at 0.0% in both periods, so trading inside that house had effectively settled on one version.
Such internal concentration can deepen liquidity in one place while tying a project's Base result to that version. A holder following Aero, Uniswap or PancakeSwap therefore needs the split inside each house, not only the parent total, because the parent total hides how far trading moved.
A strong day inside a wide year
Against the past year, the latest figure was strong without setting a peak. The highest day in that span reached $2.28 billion, while the lowest reached $235.19 million. Base has already traded in far quieter and far busier conditions within the same year, so one busy day cannot be read as a settled level.
Coverage sets a limit on precision. DefiLlama counts only the exchanges it tracks on Base, counts each version separately, and may revise a day's figure as its data completes. That limit applies evenly to the comparisons above, so it qualifies precision without reversing direction. The test now is the next weekly reading: whether share and the internal splits hold after the spike, or whether trading spreads back across venues and versions.
Concentration on two levels defines the day. Base took a larger slice of tracked trading, the largest venues took more of Base, and one version took more inside each venue. For someone trading or holding tokens there, that means liquidity was deeper where it gathered and thinner elsewhere, and the next reading will show whether the gathering lasts.<p data-vantemio-partner-disclosure="v1"><em>Disclosure: links to Coinbase on Vantemio are partner links.</em></p>