Coinbase Took Nearly Twice Binance's Bitcoin Inflow Share. The Number That Matters Is 2.5%
Bitcoin deposits split almost two to one in Coinbase's favor against Binance last month, and the striking detail is that neither venue moved much off its normal share. The split is a custody story dressed up as a sentiment story. Meanwhile, only 2.5% of recent buyers' exchange deposits arrived in profit, and that's the figure with real teeth.
Here's a number that should stop you cold: 2.5 percent.
That's the share of recent bitcoin buyers whose exchange deposits arrived in profit. Which means the other 97.5 percent, the overwhelming majority of coins hitting order books from that cohort, got there underwater. People don't move coins to an exchange at a loss for entertainment. They move them because a margin call is due, a tax bill is owed, or the position has gone so wrong that sitting still feels worse than selling.
Now hold that next to the flow data. Bitcoin deposits split nearly two to one in Coinbase's favor against Binance over the past month. That gap has been written up as a verdict on the two exchanges, a shifting of the guard, a migration of conviction.
It isn't any of those things. Or at least, the data doesn't support the drama.
The Split, and Why It Looks Bigger Than it's
The raw ratio is real enough. For every dollar of bitcoin that landed on Binance, close to two landed on Coinbase. That's a wide gap between the two largest venues in the market, and on its own it reads like a statement.
But here's the part that complicates the story. Each exchange's share of inflows stayed close to its usual level. Read that again. Coinbase didn't suddenly eat Binance's lunch, and Binance didn't hemorrhage deposits to a US rival. The two-to-one split looks like a shift because we tend to compare exchanges against each other instead of against their own baselines, and those baselines have been lopsided for a long while.
So the composition of the month explains most of it. When large holders move coins, they move them to venues with deep custody arrangements, prime brokerage desks, and the settlement plumbing that sits behind spot bitcoin fund creation and redemption. That's Coinbase's home turf. When smaller holders move coins, they move them to venues with tighter spreads and looser onboarding. That's Binance's.
Change the mix of who is moving coins in a given month and the ratio swings, even if nothing about either exchange's competitive position has changed at all. That's not a story about momentum. It's arithmetic.
Spokespeople for both venues didn't immediately respond to a request for comment.
What the Flows Actually Tell You
Here's my read, and it's a cautious one. Exchange inflow share is a custody indicator wearing a sentiment indicator's clothes, and conflating the two has become one of the more reliable ways to be wrong about this market.
Think about what generates a Coinbase deposit. A fund rebalancing. An over-the-counter desk settling with a client. A family office moving coins out of cold storage into a qualified custodian because a compliance officer finally got the signature they needed. None of that's a directional bet on price. All of it shows up as an inflow.
Now think about what generates a Binance deposit. A trader topping up margin. A retail buyer taking a first position. A market maker repositioning inventory. Some of that's directional, but a lot of it's just the ordinary churn of a high-volume spot venue.
The two streams aren't measuring the same thing. Comparing them and calling the result a sentiment reading is a category error, and it's one that gets repeated every month because the ratio is easy to quote.
Who wins from the split as it stands? Coinbase, plainly, and not because of anything it did this month. The institutional plumbing it built through the last cycle is now the default rail for large-balance movement, and that position compounds incrementally with every fund launch and every custody mandate. Custody revenue is dull, recurring, and far less cyclical than trading revenue. Dull wins.
Who loses? Offshore venues that spent years arguing they'd capture US institutional flow once the rules got clearer. The rules got clearer. The flow went to the licensed venue anyway. That's the fault line now, and it isn't really about fees.
But the louder number in all of this is still the 2.5 percent, and it's the one I'd put on a wall.
When nearly every recent buyer who moves coins to an exchange is doing so at a loss, the supply hitting the market isn't profit-taking. It's distress, or something close to it. That changes the calculus for anyone modeling the next leg. A market where holders are sitting on gains produces sellers who can afford to be patient and price-sensitive. A market where holders are underwater produces sellers who take what's bid.
So why hasn't price fallen out of bed? Because the deposits are small relative to total supply, and because the buyers above them are still holding. That's the honest answer, and it's not a comfortable one. It means the cushion is thinner than the headline numbers suggest.
What to Watch Next
Two numbers, and neither one is the ratio everybody's quoting.
The first is whether the profit share of exchange deposits climbs back above 10 percent. If it does, the overhang is clearing and the sellers become choosier. If it stays pinned near the floor for another month, the marginal seller keeps capitulating on price, and every rally meets supply from people who just want out.
The second is whether the Coinbase-to-Binance split holds near two to one once the composition of flows normalizes. If it reverts, this month was noise. If it holds, then what we're watching is a structural migration of large-balance custody, and that has nothing to do with who's bullish.
Reading the tea leaves here doesn't require a strong view on bitcoin's direction. It requires keeping two unrelated datasets in separate boxes. Coinbase won the flow month. That's a fact about plumbing. The 2.5 percent is a fact about pain, and pain is what moves markets.
The exchanges will keep publishing their splits. The interesting question is whether anyone stops mistaking one for the other.