Two clusters of automated-looking wallets generated 65% of decentralized-exchange volume on Avalanche over an eight-day period, according to a new Bitquery investigation that raises questions about how much headline blockchain trading activity reflects broad user participation.
The research, published October 6 using on-chain data verified through October 5, traced 366 wallets back to just two addresses that originally supplied them with AVAX for transaction fees.
From September 27 through October 4, one cluster accounted for 36% of Avalanche DEX volume and the other generated another 29%, Bitquery found.
The researchers say the wallets behave like arbitrage bots, repeatedly buying and selling in response to movements in AVAX on Binance. But the blockchain does not reveal who operates the clusters or confirm what, if anything, they trade on Binance, making “arbitrage” an interpretation of the observed pattern rather than a proven identity or strategy.
The bigger finding is the concentration underneath Avalanche’s headline volume. Other wallets classified by Bitquery as bots generated another 27% of trading during the period. Wallets that did not meet the researcher’s bot criteria were responsible for just 8% of volume, equivalent to roughly $12 million per day.
That means as much as 92% of the measured DEX turnover came from wallets Bitquery classified as automated.
Two Funding Addresses Led to 366 Trading Wallets
Bitquery identified the clusters by tracing where high-frequency trading wallets initially received AVAX for gas.
One funding address supplied 141 wallets between August and October 2025. A second address funded 225 wallets and has made more than 21,500 AVAX transfers to them since February 2024.
The clusters were also highly concentrated at the smart-contract level. Bitquery found that 99.6% of transactions from the older group went through a single contract, while the newer group relied on a small number of separate trading contracts.
That strengthens the case that the wallets form coordinated systems rather than hundreds of independent traders.
It does not establish who controls them.
The distinction is important. Blockchain clustering can show common funding, shared contracts and synchronized behavior, but it does not automatically reveal the legal entity or person behind an address.
A similar limitation appears in other forms of wallet-clustering analysis, where unusual relationships can identify strong investigative leads without establishing misconduct or common ownership by themselves.
The Bots Appear to Follow Binance Prices
Bitquery found the two Avalanche clusters trading primarily WAVAX against USDC.
Approximately 83% of their activity involved WAVAX-USDC, while most of the remaining volume involved wrapped bitcoin or ether against WAVAX.
The timing produced another clue.
In roughly eight out of every 10 minutes in which the clusters traded, they bought on Avalanche when AVAX was rising on Binance and sold when Binance prices were falling.
During one hour on September 29, for example, AVAX climbed from approximately $10.95 to $11.50 on Binance. Over the same period, the clusters bought around $3.9 million of WAVAX and sold about $2.1 million on Avalanche, with the heaviest buying occurring around the price rise.
Bitquery calculated same-minute correlations of 0.58 and 0.61 between the clusters’ trading direction and Binance price changes, while correlations with Binance’s following minute were close to zero.
That pattern is consistent with algorithms reacting to centralized-exchange price discovery and trading against temporary price differences on Avalanche.
But the researchers cannot see whether the clusters actually execute an offsetting trade on Binance. Any centralized-exchange leg would occur outside the public blockchain.
That limitation resembles the problem investors face when interpreting large visible on-chain positions: transparency can reveal one side of a strategy without necessarily revealing the complete economic exposure.
Pharaoh Captured 81% of Avalanche DEX Volume in September
The concentration extends beyond the wallets themselves.
Pharaoh became the dominant Avalanche decentralized exchange during 2026, according to Bitquery’s reconstruction of C-Chain swaps.
Its share of Avalanche DEX volume increased from approximately 27% in April to 81% in September. Two WAVAX-USDC Pharaoh pools alone generated 62% of the chain’s DEX volume during September.
The two identified bot clusters were heavily concentrated there as well. Bitquery found that 83% of one cluster’s trading and 89% of the other’s went through Pharaoh.
In Pharaoh’s busiest pool, the two groups paid approximately $347,000 in pool fees during the eight-day study period—roughly seven times what they paid in network gas.
That creates an important economic distinction.
Automated turnover may inflate the impression of how many independent traders are using Avalanche, but it can still generate real fee income for the infrastructure processing that activity.
The volume is therefore not necessarily “fake.” Tokens genuinely change hands, transactions consume block space and fees are paid.
The question is what the volume actually measures.
The Two Clusters Paid Nearly a Third of Avalanche’s Gas
Bitquery estimates the two clusters were responsible for approximately 32% of all Avalanche gas fees during the post-price-floor period.
On a typical day, the groups paid around 580 AVAX in network fees, worth roughly $6,400 at the prices used in the investigation.
One cluster’s daily gas expenditure more than tripled after Avalanche’s new minimum gas-price mechanism began affecting transaction costs, increasing from around 84 AVAX to 297 AVAX on a typical day.
The change followed Avalanche’s Helicon upgrade and the implementation of a validator-controlled minimum gas-price mechanism designed partly to make ultra-cheap automated traffic less economical.
The higher floor dramatically reduced some other automated activity on Avalanche, but it did not eliminate the DEX clusters.
Instead, the number of swaps fell while average trade sizes increased.
High Volume Does Not Necessarily Mean Broad Liquidity
This is the most important investor takeaway from the research.
Crypto dashboards frequently use transaction counts, active addresses and DEX volume as shorthand for network adoption.
Those numbers can be technically correct while telling a misleading story about participation.
If one algorithm splits itself across 200 wallets and trades the same capital repeatedly, the blockchain records hundreds of active addresses and potentially billions of dollars in volume. That is materially different from hundreds of independent investors each bringing new capital onto the network.
The issue recently appeared in a different market when repeating crypto-perpetual trades on Kalshi showed how a relatively small number of professional liquidity participants can generate a disproportionately large share of headline turnover without the trades necessarily being fraudulent.
Avalanche’s data pushes the same lesson much further.
Bitquery found that the two clusters alone have generated between 40% and 67% of Avalanche DEX volume every month since September 2025.
This is therefore not an eight-day anomaly.
The 92% Bot Figure Needs One Important Caveat
Bitquery’s classification is analytical, not absolute.
A wallet was classified as a bot under its broadest test if it sent at least 50 transactions on a typical active day or 1,000 transactions on any single day.
A sufficiently active human trader could therefore be classified as automated, while a slower algorithm could avoid the threshold entirely.
The 8% “non-bot” figure should consequently be read as wallets that did not exhibit Bitquery’s chosen high-frequency behavior, not as a perfect count of human trading.
More importantly, arbitrage itself is legitimate market activity.
Arbitrage bots can improve price alignment between exchanges, provide counterparties to other traders and make fragmented crypto markets more efficient.
Nothing in Bitquery’s investigation establishes wash trading, market manipulation or fraudulent volume generation by the two clusters.
The concern is concentration, not necessarily misconduct.
Avalanche May Have a Measurement Problem More Than a Trading Problem
For Avalanche, the findings create an awkward but useful distinction.
The network clearly has trading activity. Hundreds of millions of dollars of tokens are actually moving through its DEXs, contracts are being executed and fees are being paid.
What the figures may not demonstrate is equally broad retail or institutional adoption.
If two automated systems generate nearly two-thirds of volume and all wallets outside Bitquery’s bot categories generate only around $12 million per day, then investors evaluating Avalanche based primarily on aggregate DEX turnover could be overestimating the diversity of its trading ecosystem.
The better metrics now are concentration-adjusted ones: how many independent funding sources exist, how much volume survives after the largest clusters are removed, how much liquidity remains if incentives change, and whether unique recurring users continue growing.
That is ultimately what makes the October 6 Bitquery investigation more important than another bot-wallet alert.
It does not show that Avalanche’s billions in DEX volume are fictitious.
It shows that blockchain volume can be real while the apparent breadth of participation behind it is much smaller than the headline number suggests.
Johan Shamshad is a financial markets writer at Dave Finances covering cryptocurrencies, trading platforms, brokers, fintech, financial regulation, and developments across global markets. He previously worked at Gulf News, adding newsroom experience to his coverage of fast-moving financial and digital-asset markets.
His work focuses on identifying market-moving events, company developments, regulatory changes, product launches, and shifts in trading and financial infrastructure.
Johan contributes news and analysis designed to help readers understand not only what happened, but why a development matters and how it may affect the wider financial landscape.