ETH on Exchanges Climbed 7.6% in 27 Days. Here's What I Do With That Number

I pull a wallet-tracking job every six hours that sums the ETH balance sitting across 14 known wallets on 6 exchanges — Binance, Bitfinex, Kraken, Gemini, Upbit and Crypto.com (Coinbase has no single publicly identifiable address I could find, so it's excluded). On September 24, 2026, that tracked balance bottomed at 5,564,589 ETH. By October 21, it had climbed to 5,987,765 ETH. That's 423,176 more ETH parked on exchanges in 27 days — a 7.6% increase. As of today it's pulled back slightly to 5,932,493 ETH, still up 6.3% over the trailing 30 days.

I want to be upfront about what this data actually is before I say anything about what it means: it's a best-effort sum of 14 wallets I can identify, not an audited exchange-by-exchange reserve figure, and exchanges rotate cold-wallet addresses over time without announcing it. Binance alone accounts for 2,593,828 ETH of the 5,932,493 tracked total — 43.7% of everything I'm counting. If one of Binance's wallets got relabeled mid-month, the whole curve shifts. I re-check the wallet list periodically for exactly this reason.

Why exchange balance is a supply number, not a price prediction

ETH held in a self-custody wallet isn't available to sell into an order book in the next five minutes. ETH sitting on an exchange is. So when the tracked balance rises nearly 424,000 coins in under a month, what's actually growing is the pool of coins that could hit the bid without a bridge transaction or a withdrawal delay in the way. That's not the same claim as "price is about to drop" — deposits also happen because people are moving coins to trade, to stake through an exchange product, or to collateralize a derivatives position long, not just to sell spot. I've seen balances climb into rallies before. The number tells you about available liquidity, not about direction.

Where it does matter directly is leverage. More available sell-side supply sitting on exchanges is one more input into how fast a cascade can move once it starts, because the coins needed to fill market sells on the way down are already sitting in exchange wallets instead of needing to be bridged in first. It's the same mechanic I wrote about when ETH's 8.3% squeeze wiped out $255M in shorts in a single session — the leverage band that gets caught depends on how far the move actually travels, and available liquidity is part of what decides how far a move can travel once it's underway.

What I actually check now

I added the 50-day history to what I look at before sizing anything leveraged on ETH — not as a trigger, as context. If the balance is near a multi-week high, like it was on October 21st, I treat that as one more reason to run my position through the ETH liquidation calculator at a leverage a notch lower than I'd otherwise use, the same way I'd tighten up before a scheduled options expiry or a thin-liquidity weekend. It changes the size I'd run, not whether I'd enter. I check the leverage number itself against the safe leverage calculator and size the position with the position size calculator rather than trusting an exchange's default slider — the same habit I described after the ETF outflow week where $645M left the ETFs while perp positioning stayed net long anyway. Both are supply/demand reads that don't tell you what happens next on their own — they tell you how much margin for error you actually have if it goes the wrong way.

The honest summary: 423,176 ETH landed on exchanges in less than a month, the balance is still elevated versus where it sat in September, and I'm not making a directional call out of that. I'm just running smaller leverage until it normalizes. If you want the underlying math on what any given move does to a leveraged position, that's what the Learn section and the liquidation calculators are for.