Most whale-tracking dashboards show a green "bought" or a red "sold" next to every large move. We wanted to know how many of those moves were actually trades.
So we classified every move made by our tracked Ethereum whale wallets between September 6 and September 30, 2026 — 25 full days, 31,523 moves, $9.74B in value. Each one landed in one of two buckets:
- Exchange transfer — coins moving onto or off a centralized exchange (a deposit or a withdrawal).
- DEX swap — an actual trade on a decentralized exchange like Uniswap.
The headline: 97.4% of the dollars were exchange transfers
| Asset group | Moves | Value | Exchange share of $ | DEX share of $ |
|---|---|---|---|---|
| ETH / WETH / wBTC | 14,832 | $4.00B | 93.8% | 6.2% |
| Directional tokens (UNI, LINK, ONDO…) | 11,637 | $1.04B | 99.1% | 0.9% |
| Stablecoins | 5,054 | $4.70B | 100% | 0% |
| All | 31,523 | $9.74B | 97.4% | 2.6% |
For tokens like UNI, LINK and ONDO, 99.1% of the whale dollar volume never touched a DEX. It was coins being withdrawn from exchanges or deposited to them.
That matters because a withdrawal is not a purchase. Moving $5M of LINK from Binance to a cold wallet tells you where the coins went. It does not tell you the wallet bought anything — the coins might have been bought months earlier on the exchange, or not bought at all.
Size changes everything
The DEX share collapses as moves get bigger:
| Move size | Moves | DEX share of moves |
|---|---|---|
| Under $10K | 6,502 | 29.3% |
| $10K–$100K | 3,325 | 3.8% |
| $100K–$1M | 1,638 | 0.2% |
| $1M–$10M | 172 | 0.0% |
Small whale moves are often real swaps. Once a move passes $100K, it is almost always an exchange transfer. In 25 days we did not see a single directional-token move of $1M or more happen on a DEX.
What the exchange flow did show
Labelled honestly, the exchange data is still informative. Across directional tokens, tracked whales withdrew $688.8M from exchanges and deposited $340.2M — net $348.6M leaving exchanges, a 66.9% withdrawal share. On-chain analysts conventionally read coins leaving exchanges as an accumulation-side signal and coins arriving as a distribution-side signal. Those are descriptive labels for observed flow, not predictions.
The tokens with the largest net exchange outflow in the window:
- LINK: +$45.1M net leaving exchanges (69.4% withdrawal share), 126 wallets
- UNI: +$41.8M (59.7%), 132 wallets
- QNT: +$38.2M (70.5%), 69 wallets
- ONDO: +$37.7M (73.1%), 72 wallets
- ENA: +$23.6M (67.1%), 67 wallets
Every one of those had a DEX share of 0.3% or less.
How to read whale data without fooling yourself
- Check the move type before the direction. "Withdrawn from Binance" and "bought on Uniswap" are different events. Treat them differently.
- Use size-aware thresholds. A $500K stablecoin move is routine; a $500K small-cap token move is rare.
- Separate the two flows when you aggregate. Blending exchange transfers and swaps into one "buy ratio" hides what actually happened.
We changed our own labels after running this study: on Deep Blue Alpha, exchange moves are now shown as withdrawals and deposits, and only real swaps are called buys and sells. You can see the split live on the whale feed and per token on pages like LINK.
The full study, with methodology and the complete token table, is here: 97% of Whale Flow Is Exchange Transfers, Not Trades.
Data: Deep Blue Alpha, whale-scoped moves from tracked Ethereum wallets, September 6–30, 2026 (UTC). Exchange-transfer rows are kept for 30 days, so the window starts September 6.
Deep Blue Alpha is an Ethereum whale intelligence platform tracking 20,000+ whale wallets in real time. This article is for informational purposes only and does not constitute financial advice. NFA/DYOR.
Track whale activity for free at deepbluealpha.io
Top comments (0)