AAVE and UNI, in Layers: When the Last Week Disagrees With the Month
Over the 30 days to October 10, 2026, AAVE's tracked whale flow leaned clearly toward coins leaving exchanges: a 63.2% withdrawal share, $18.12M net out. Over the final seven days of that same window, the lean had nearly evaporated. The withdrawal share was 56.8% and the net was $3.44M.
UNI did the opposite. Its 30-day withdrawal share was a near-balanced 57.9%, and its final week tilted harder, to 65.0%.
Same tokens, same data, same window. The month said one thing and the most recent week said another, and in opposite directions for the two tokens. This is the normal state of whale data, not an anomaly, and it is the reason a single time window is never enough to describe a token's flow.
This piece walks through AAVE and UNI at three scales: the month, the week and the day.
The month, split into weeks
| Week (UTC) | UNI $ moved | UNI withdrawal share | AAVE $ moved | AAVE withdrawal share |
|---|---|---|---|---|
| Sep 13–19 | $58.56M | 65.5% | $4.64M | 71.3% |
| Sep 20–26 | $87.56M | 49.4% | $6.48M | 58.5% |
| Sep 27–Oct 3 | $78.55M | 56.7% | $31.63M | 69.0% |
| Oct 4–10 | $48.76M | 65.0% | $25.33M | 56.8% |
(September 11–12, the first two days of the window, are left out; they held 2.5% of UNI's whale dollars and 1.1% of AAVE's.)
A withdrawal share is withdrawals divided by withdrawals plus deposits. Above 50%, more coins left exchanges than arrived, the conventional accumulation-side read. Below 50%, more arrived, the distribution-side read. Both describe observed movement, not intent. And because DEX swaps were 0.0% of UNI whale dollars and 0.1% of AAVE's, all of this is exchange transfers, not purchases or sales.
Read down each column and neither token held a steady line. UNI swung from 65.5% to 49.4% and back up to 65.0%. AAVE alternated between the high 60s and the high 50s. A monthly ratio is the dollar-weighted average of those swings, and dollar weighting matters a great deal here, as the next section shows.
AAVE's month happened late
The single most important fact about AAVE's window is when the dollars arrived.
| AAVE whale $ moved | Amount | Share of 30D |
|---|---|---|
| Sep 11–26 (16 days) | $11.84M | 17.2% |
| Sep 27–Oct 10 (14 days) | $56.95M | 82.8% |
For the first 16 days, AAVE whale flow was light. September 14 saw $53.3K moved; September 26 saw $67.8K. Then activity jumped. September 28 moved $5.86M, September 30 moved $8.01M and October 1 moved $8.67M from 40 moves by 19 wallets, AAVE's busiest day of the window.
UNI's month was spread far more evenly. The same final 14 days held 45.4% of UNI's dollars, which means UNI's big September carried more of its month than AAVE's quiet September did.
The practical consequence: AAVE's 30-day numbers are, in effect, a two-week story wearing a one-month label. Anyone reading "AAVE: $68.80M of whale flow over 30 days" would picture steady activity. The daily data shows almost none for half the window.
AAVE's five heaviest days
| Date (UTC) | AAVE whale $ moved | Share of 30D | Net leaving exchanges |
|---|---|---|---|
| Oct 1 | $8.67M | 12.6% | $3.91M |
| Sep 30 | $8.01M | 11.6% | $2.66M |
| Oct 5 | $6.66M | 9.7% | $3.81M |
| Oct 6 | $6.42M | 9.3% | $4.04M |
| Sep 28 | $5.86M | 8.5% | $1.71M |
Five days carried 51.8% of AAVE's month, and every one of them fell between September 28 and October 6. All five finished net leaving exchanges. Nine of AAVE's 30 days moved less than $500K.
UNI's heaviest days were spread across the whole window instead: October 1 ($29.96M), September 15 ($22.23M), September 21 ($20.04M), October 5 ($17.97M) and September 23 ($16.56M). UNI's five busiest days carried 38.1% of its month, and only two of its 30 days moved less than $500K.
That difference in shape is easy to miss in a monthly table and obvious in a daily one. AAVE had a burst. UNI had a steady, heavy month.
The day the streak ended
Inside those busy two weeks, AAVE had a clear run and then a clear break.
From September 30 through October 6, AAVE finished seven straight days with more coins leaving exchanges than arriving. Then the direction flipped. October 7 finished $3.31M net onto exchanges and October 8 another $2.64M, a combined $5.95M arriving in two days. October 7 included the largest AAVE move of the latest week: a $2.95M deposit to an exchange by a shortened wallet, 0xb830ba…3c64, on its own 11.7% of that week's AAVE whale dollars.
That two-day reversal is most of the reason the latest week's net shrank to $3.44M from $12.01M the week before. Deposits rose from $9.78M to $10.94M while withdrawals fell, and the withdrawal share dropped from 69.0% to 56.8%.
Over the whole window, AAVE had 11 days that leaned toward deposits out of 30. UNI had 8.
UNI's mirror image
UNI's latest week moved the other way. Whale dollars fell 37.9%, from $78.55M to $48.76M, but deposits fell faster than withdrawals: $34.02M deposited in the prior week against $17.05M in the latest. The withdrawal share rose from 56.7% to 65.0% and net leaving exchanges rose from $10.51M to $14.60M.
The daily view shows the run that produced it. From September 28 through October 6, UNI finished nine straight days net leaving exchanges, $21.71M in total. UNI's lone deposit-leaning week came earlier, September 20 to 26, at a 49.4% withdrawal share, and it contained the two largest single UNI moves of the month, both deposits.
Participation shrank in both tokens even as the ratios moved apart. UNI went from 73 active whale wallets and 296 moves in the prior week to 51 wallets and 152 moves in the latest. AAVE went from 54 wallets and 179 moves to 47 wallets and 128. Fewer wallets doing less is the common thread; what differed was which side of the exchange boundary the remaining activity favoured.
So the two tokens crossed. Over the month, AAVE was the more one-sided; in the final week, UNI was.
Price did not follow either line
| Oct 4–10 week | UNI | AAVE |
|---|---|---|
| Withdrawal share | 65.0% | 56.8% |
| Net leaving exchanges | $14.60M | $3.44M |
| Price change, 7D to Oct 11 | −13.3% | −5.4% |
| Price change, 30D to Oct 11 | +26.5% | +34.2% |
In the latest week, UNI's flow tilted further toward coins leaving exchanges while its price fell more steeply. AAVE's flow moved back toward balance while its price fell less. Over the full month, AAVE rose more than UNI while its whale flow was the smaller of the two.
None of that is a contradiction. Exchange flow records where large holders moved coins; price reflects the whole market. Over a week or a month, the two series often move independently, and this window is a clear example. Treating exchange flow as a price explanation, or price as a flow explanation, would mislead in both tokens.
How to read whale flow at more than one scale
A method that holds up across tokens:
- Start with the month, then split it into weeks. If the weekly ratios swing more than ten points, the monthly ratio is an average of different behaviour, and the description needs to say so.
- Check when the dollars arrived. If one half of the window holds most of the dollars, as AAVE's final 14 days held 82.8%, say so. A 30-day label on a two-week story misleads.
- Find the streaks and the breaks. A run of same-direction days and the day it ends are usually more informative than the total. AAVE's October 7–8 deposits explain its latest week better than any ratio.
- Report both windows when they disagree. "AAVE leaned toward the accumulation side over 30 days; the latest week moved back toward balance" is accurate. Either half alone is not.
- Keep price in its own column. Show it alongside the flow, never as its cause or result.
Writing a two-window sentence
When the month and the week disagree, the sentence has to carry both. Some rewrites from this window:
- Instead of "AAVE whales are pulling coins off exchanges," write "AAVE leaned toward coins leaving exchanges over 30 days (63.2%), but the latest week moved back toward balance (56.8%) after two deposit-heavy days."
- Instead of "UNI whale flow is balanced," write "UNI's 30-day flow was close to balanced (57.9%), with a latest week that tilted further toward withdrawals (65.0%)."
- Instead of "AAVE saw $68.80M of whale activity this month," write "AAVE saw $68.80M of whale activity, 82.8% of it in the last 14 days."
Each version is past tense, names the window, and leaves the reader with the shape of the flow rather than one number.
How the windows were cut
All windows are UTC. The 30-day window runs from September 11, 00:00 to October 10, 23:59, the full span for which exchange-flow rows are retained. The weeks are fixed seven-day blocks ending on October 10, so the latest week is October 4 to 10 and the prior week September 27 to October 3; the first two days of the month sit outside the weekly table. Daily figures sum each day's exchange withdrawals, exchange deposits and DEX swaps by tracked wallets. Fixed, stated windows matter: shifting a week by a single day would move AAVE's October 7 deposit into a different block and change the comparison.
Checking a specific day yourself
- Open AAVE on Deep Blue Alpha and find the October 7 deposit, noting the shortened wallet.
- On Etherscan, search the full address and open Token Transfers (ERC-20), filtered by the AAVE contract address.
- Find the October 7 transfer and check the receiving address. If Etherscan labels it as an exchange wallet, the move was a deposit.
- Confirm the sending side is not also an exchange; exchange-to-exchange transfers are internal shuffles, not holder decisions.
- Scroll the wallet's history before and after the date. A deposit from a wallet that withdrew the same token a week earlier is a round trip, which changes how much weight it deserves.
What the layers still cannot tell you
They cannot identify wallet owners; wallets are shortened and not attributed. They cannot explain why AAVE's whale activity arrived late or why UNI's deposits slowed. They cover tracked wallets on Ethereum only. And no number of time scales turns past exchange flow into a statement about future prices.
Takeaway
AAVE and UNI crossed over in this window. Over 30 days, AAVE was the more one-sided token; in the final week, UNI was, while AAVE's lean shrank after two deposit-heavy days. AAVE's month was also mostly two weeks long, with 82.8% of its whale dollars in the last 14 days. A monthly figure is a starting point. The weekly split, the timing of the dollars and the streak breaks are where the actual description lives.
The full comparison, with gross flows, wallet overlap and the largest moves, is at UNI vs AAVE: Whale Flow Compared. Live moves are on the UNI and AAVE token pages, the wider sector is on the DeFi whale board, and DeFi blue-chip whale activity covers the other governance tokens.
Data: Deep Blue Alpha tracked-wallet data, whale-scoped, September 11 – October 10, 2026 (UTC), Ethereum only. Market data from CoinGecko, October 11, 2026, 18:30 UTC.
Read the full analysis with interactive charts: AAVE and UNI, in Layers: When the Last Week Disagrees With the Month
Track smart money in real time at deepbluealpha.io — free tier, no signup required.
More research: deepbluealpha.io/research | Live feed: deepbluealpha.io/feed | Wallet leaderboard: deepbluealpha.io/wallets
Comments
Post a Comment