A Uniswap Range Is A Job, Not A Setting.

Your token launched, and somebody had to decide where its liquidity would sit. Maybe it was you. Maybe it was a market maker's analyst, or an engineer who picked two numbers that looked sensible on the day. Either way, two prices got typed into Uniswap, and the team's tokens went in between them.

Then the price moved, the way prices do. Some mornings it's still inside the range you drew. Some mornings it isn't, and nobody on the team could tell you which kind of morning today is without going to look.

We went and looked at every one on Ethereum.

On Uniswap you don't have to spread your tokens across every possible price. You pick two prices, and your tokens are only available to trade while the price is between them. That's the whole idea, and it's a good one: while the price stays inside, the same money earns far more from every trade than it would spread thinly from zero to infinity.

When the price leaves, two things happen at once. Your money stops earning anything. And it has all turned into one of the two tokens, whichever one people were selling on the way out.

That second part is the one people forget. A range the price has fallen out of isn't sitting there safely in cash. It's sitting in the token that fell.

We read every position ever opened on Uniswap v3 on Ethereum, 1,224,442 of them, as they stood at one moment on the morning of September 15. Most are long closed. Of the ones still holding money, we kept the ones in pools where both tokens are dollars, ETH or bitcoin, so nobody has to take our word for what any of it is worth.

At $100,000 and up, which is where a treasury lives, that's 356 positions holding $217.9 million between them. $65.9 million of it was outside its range, earning nothing. That's 30.2%, about three dollars in every ten, and it doesn't get better with size.

Whatever floor you pick, from a dollar to a hundred thousand, the share out of range sits within a point of thirty percent. The people with the most money in these ranges are not tending them any more closely than everyone else.

It isn't one odd morning, either. We ran the same count two days earlier and got 29.7%. And widening it to every pool we could put a fair price on, project tokens paired with ETH or dollars included, doesn't change the answer: 30.1% of $529.6 million.

If you've run a pool, you already have an answer ready for that number. Some people put money in a range on purpose where the price isn't, as a way of saying "sell my ETH if it gets to this price." Uniswap calls those range orders, and they are out of range by design.

They also have a shape. A range order is as narrow as Uniswap allows, one step wide, sitting just past the price, so it fills the moment the price walks through it. So we looked at the shape of the money that's out of range.

None of the $65.9 million sits in a range one step wide. 89.2% of it, almost nine dollars in ten, is in ranges whose top price is more than 5% above the bottom one. On 78.2% of it, the price would have to move more than 10% just to get back to the edge. Those are ranges drawn to earn around a price, and the price left.

Some of it will still be deliberate. A wide range on one side of the price is also a patient way to sell out of a token over time, and a count can't read anybody's intentions. But almost none of it is shaped like a limit order, and "they meant to do that" now has a lot more to explain.

98.7% of the out-of-range money at $100,000 and up is held directly in ordinary accounts. Not in a vault, not in a company's shared wallet that needs several people to sign, not inside any program. That doesn't tell you a person rather than a script is behind every one of them. It tells you that while it's been sitting out of range, nobody has redrawn it.

And it's almost all in the pools you'd think were the easy ones. 93% of the out-of-range money is in pools that trade ETH against dollars, the busiest pairs there are, and in those pools 30.5% of everything in a range is out.

None of this is news to the people who sell the fix. Arrakis, which runs liquidity for token teams, says it plainly in its own docs: doing this in house "generally requires a full-time quant function," and "without active management, concentrated positions quickly fall out of range, stop earning fees, and leak value." They're right, and the count above is what it looks like when they are.

So here's what's actually on the table.

Arrakis does real work: you deposit into its vault and pick a strategy, and it rebalances as the price moves and adjusts the fee to the market, which is real skill and worth paying for. Revert's Auto-Range leaves the position in your own account, and when the price drifts out by an amount you choose, its bots take the money out and put it back in a range of the same width around the new price, for 0.15% of the position each time. A market maker brings exchange listings and depth no pool will give you. And full range never leaves its range, because it doesn't have one. It spreads your money across every price there is and earns the least for it.

If your pool looks like everybody else's, three tenths of what the team has in its range is doing nothing on an ordinary morning. Two million dollars in a range is six hundred thousand dollars sitting in whichever token people were selling, earning nothing. And if a bot is redrawing it for you, 0.15% of two million dollars is three thousand dollars, every time it does.

Look at what every one of those options has in common. The range itself never changes. It's two prices, typed once, and nothing the market does moves them. The only thing that ever redraws a range is somebody sending a new instruction: you, when you remember, a manager you've handed the tokens to, or a bot running its own rule.

So the choice everyone puts in front of you, which range to draw and who should draw it, is really a choice about who keeps redrawing it for as long as the money is in the pool.

That's the job we built Plug for. Plug is programmable Ethereum transactions without smart contracts. Instead of typing two prices, you write down how the range should be drawn: say, ten percent either side of wherever the price is. You sign that once. When it runs, it reads the pool's price at that moment, works the range out from it, takes the money out of the old range and puts it into the new one, swapping some into the other token first if the price left it all on one side. It runs from the team's own account, on the team's own rule.

It doesn't write the rule for you, and a plug doesn't know your token better than a desk of quants does. What it changes is where the rule lives: in a sentence about the price, worked out each time from the price that's actually there, instead of two prices somebody typed at launch.

On the morning we looked, $65.9 million of the $217.9 million in treasury-sized ranges was sitting outside them, almost none of it shaped like an order and 98.7% of it held in ordinary accounts that had not redrawn it. Every one of those ranges was drawn from a price that has since moved.

We read every position ever opened on Uniswap v3 on Ethereum, 1,224,442 of them, at a single moment (Ethereum block 25,983,329, the morning of September 15, 2026), with no failed reads. 1,102,517 had been emptied, 9,477 were full range and can't leave it, and 5,242 were in pools with no dollar, ETH or bitcoin token on either side, so there was no honest way to price them. The headline uses only pools where both tokens are dollars, ETH or bitcoin, so no pool's own price enters the value: 18,185 positions worth $271.3 million, 356 of them worth $100,000 or more. The wider figure counts a project token at no more than what its pool actually holds in dollars, ETH or bitcoin, because pricing thinly traded tokens at their own pools' prices puts a trillion dollars in these positions. Fees earned but not yet collected aren't counted. Managers that put money into a pool straight from their own vault, which includes Arrakis, aren't in these positions at all. The same count at block 25,972,572, two days earlier, gave 29.7%. Each count describes one moment: how much is out of range, not for how long.