Fake Liquidity in Crypto: A $579M Pool With $29 of USDC

The largest number on a crypto pool's page can be the one nobody paid for. Fake liquidity is a pool whose dollar figure is mostly a token its creator minted, valued at the pool's own price. Its cash side, what a seller could draw on, sits nearly empty. On the afternoon of 23 October, DEXScreener's first result for "Lisk" was a Solana pool listed at $579.3 million of liquidity. The USDC inside it came to $28.86.

Everything else in that figure was a billion tokens called LSK, counted at the price of the pool's own trades. Lisk has no official token on Solana. These tokens trade in no other pool the screener lists, so no outside market ever checked the price that turned them into half a billion dollars.

DEXScreener's arithmetic was right, and whoever built the pool chose both numbers in it.

Tao explains DeFi by taking a pool apart until every piece has a name. When a piece is out of his sight, he says so. Nina is new enough to markets to check each number against a second source, and she has been online long enough to expect the first source to be selling something. Before the Lisk pool goes back on the screen, they work out how any pool's figure gets built, with no prices attached.

Nina found the pool through the search box. She wanted a little Lisk in the Solana wallet where she keeps her USDC, so she typed the name into the screener. The top result looked like the deepest market she had seen: half a billion dollars of liquidity, a steady day of trading, a price that barely moved. Binance's Lisk market sat in her next tab. The screener said $0.5795 a token, and Binance said $0.1846.

"That's three times the price for the same coin," Nina says. "Either Binance is wrong or this isn't the same coin. If it isn't, how does it have more liquidity than the real one?"

"What do you think the half a billion is?" Tao asks. Money, she answers, waiting in the pool to buy from her. He tells her that describes about half of an ordinary pool, and almost none of this one.

How does a screener count a pool's liquidity?

Tao ignores the screen and opens up the pool itself. A pool holds two piles, and the screen shows you only their sum. One is the token being traded, called the base, and every sale of it is paid out of the other, called the quote, usually USDC, SOL or ETH. To reach one dollar figure, the screener multiplies the base pile by the pool's current price and adds the quote pile.

"Half of it is buyers' money, then," Nina says. "In an ordinary pool, about half," Tao says, "because the quote side is money somebody else issued. Circle stands behind every USDC in it, and nobody running the pool can print more. The base side is the token, valued at whatever the pool says it's worth."

Whatever the pool says needs a second market to check it. When a token also trades somewhere else, on Binance or in other pools, its price can't wander far. If this pool quoted it above Binance, someone would buy there and sell here until the gap closed. Arbitrage ties the pool to the outside market. Its two piles check each other, and the dollar figure means roughly what it says.

Take the outside market away and the check goes with it. A token that exists in one pool has no other price to be pulled toward. Its price is the pool's own last trade, so the base pile is valued at a number the pool produced about itself. It's a scale weighing itself, Tao says.

Nina has met liquidity before as whether a market can take your size without the price running away from you. That question has to wait for an earlier one: how much of what sits in the pool was ever money.

Permissionless markets allow this by design. Anyone can create a token and open a pool for it without asking permission. That freedom reaches the price as well. Whoever makes a new token's first trades decides what every screen will say it is worth.

Why can one position hold a price still?

Something about the price still bothers Nina. It has barely moved all day, and a token worth nothing outside this pool ought to sink. "Someone is holding it up," she says.

Tao's answer is a design called concentrated liquidity. Older pools spread a provider's money over every price above zero. Each trade nudged the price along one long curve. Pools built on Uniswap's v3 design, including the Raydium pool on Nina's screen, let each provider pick the range where their money works. A provider can pack all of it into one narrow band around a single price.

Inside that band the position trades both ways. A buyer pays USDC and receives tokens from the position's pile, and the USDC lands in the same position. When the matching sell arrives, the position takes the tokens back and returns the USDC it just received. Two trades pass through. The price holds not because buyers and sellers agree but because one position stood on both sides of them.

"Whose USDC is it while it sits in there?" Nina asks.

"The position's owner. The tokens in a position belong to whoever opened it, and so does any USDC a buyer pays in. The owner can withdraw it," Tao says. He leaves it at can, because nothing on her screen shows this owner doing that, and nobody has recorded it.

In a pool built this way, whoever owns the position decides when the market ends, and the price sits wherever that owner chose to stand.

What the fake Lisk pool holds

Tao pulls the pair's raw data. DEXScreener's API splits it into three numbers: a dollar total, the base pile and the quote pile. Nina goes through it with the mechanics in hand. Its base pile held 999,697,645 LSK, and its quote pile 28.8561 USDC.

"Twenty-nine dollars," she says. She divides it into the total herself and gets about twenty million to one. Only 0.000005% of the figure is money anyone else issued. The half a billion she took for buyers' money isn't buyers' money, it's a pile of tokens times the pool's own price.

Everything else rebuilds easily. Solana's own record for the token shows a supply of 999,999,999, so 99.97% of this LSK sits in the pool. Multiply that pile by the pool's price, add the USDC, and the $579.3 million comes back. Bitquery traced the token to a wallet that minted it on 16 September and also controls its name and logo.

"Show me every LSK the search turns up, next to the one that costs what Binance says," Nina says. Tao lines them up as the search ranked them, all read on the afternoon of 23 October.

Together, the search's top three pools show about a billion dollars. Behind that billion sits about $37 that anyone else issued. Fourth comes the pool priced where Lisk actually trades, with the smallest figure of all. Its total is about two and a half times the ether behind it, the shape of a pool where both piles were bought. The Robinhood Chain pool sits on another chain altogether, and since nothing in Bitquery's Solana work ties it to the other two, it shows only that the shape travels.

A screen that ranks pools by displayed liquidity hands the top spot to whoever can display the biggest number, and a counterfeit can always display more than the coin it copies.

"I'd have bought the first one, because it looked the safest," Nina says. "The biggest number had twenty-nine dollars behind it."

Where does the volume in a fake liquidity pool come from?

One reason to trust the pool survives. "Somebody's trading it, though," Nina says. The screen showed $82,359.70 of volume over the past day, eleven buys and eleven sells. Trades like that look like people.

The trades aren't fake, Tao says. They're circular, and Bitquery has traced whose money goes around. One wallet, working with about 12,000 USDC and a little SOL for fees, buys and sells the same tokens back to back. On the night of 20 October (UTC) it sent 11,102 USDC into this pool for LSK. Twenty-two seconds later it sold the LSK back, $1.13 poorer. For two network fees too small to notice, the trip added about $22,200 to the pool's volume.

That wallet repeats the trip in 438 pools, about twice a day in each, and between 21 September and 20 October it produced $1.40 billion of volume, by Bitquery's count. That was 22.7% of everything traded on Raydium's concentrated pools. In the day to 21 October it ranked seventh among all Solana DEX traders by volume. Each swap in those totals is valued at its USDC side, so the counterfeit prices can't inflate them.

Nina's arithmetic puts the day's volume at about 2,854 times the pool's USDC. A pool that turns over thousands of times its reserve in a day isn't holding anyone's money.

The money visits, and it leaves with its owner.

The pool's neighbors give it away. When it was read, nine pools showed exactly $82,359.70 of volume for the day. All nine had eleven buys and eleven sells: this one and eight tokens called XMR. Seven more showed exactly $43.90. Unrelated markets don't produce identical days. Figures like these are what one script sending the same trades through each pool would produce, which fits Bitquery's description of a wallet running "the same trip in hundreds of other pools".

On a chain, volume is counted swap by swap, which is why one routed swap can print at several pools. A wash cycle's swaps print like anyone else's, at about a dollar a trip. Every dollar of that volume is real, but it measures how often the same money went around, not how many people wanted the token.

Why the usual checks pass a fake liquidity pool

Nina could recite the standard warnings about new tokens. She checks whether the creator can mint more and whether anyone can freeze her balance, then how big the liquidity is and whether anybody trades. This pool passes all of them.

Solana's record for the token, read on 23 October, shows both the mint authority and the freeze authority set to none. Nobody can print more LSK, and nobody can freeze a holder. The pool's figure is enormous, and its volume arrives daily. Twelve helper wallets made a million and a quarter swaps in a month, many of them under twenty cents, and Bitquery says their job is to make each pool show more than one trader.

A small test trade would probably pass as well, since the position quotes both sides at one price. Bitquery's recorded round trip is that test at eleven thousand dollars, and it lost $1.13. Even a lock, the check that sounds strongest, only keeps a position in the pool. It adds no USDC.

No outsider has been shown losing money to these pools, and Tao won't invent one. In the week to 21 October, four wallets with no tie to the bot traded one of them once each, about $27 in all. Bitquery found no sale to an outsider at any price. Its reading is that the pools are bait "for people searching a ticker or pasting a 'USDT' token by mistake". Who runs them is unknown, and Bitquery names nobody.

"I searched a ticker," Nina says.

Those checks were written for a pool someone plans to drain. The risk in this one isn't what its creator can take out, it's what a buyer would put in, and into whose position.

How do you check a token's liquidity?

"What do I read first on the next one?" Nina asks.

"The side somebody else paid for," Tao says. That pile of USDC, SOL or ETH is all a seller of the token can draw on, plus whatever new buyers bring later. DEXScreener's data carries it separately from the total. Wherever she reads a pool, she can find the two amounts it holds before she looks at their sum.

Then she weighs the total against that pile. A pool where both sides were bought shows a total close to twice its quote side, since each side is worth about half. On 23 October a Meteora pool trading Monero at the coin's price showed 2.0 times, Lisk's Ethereum pair 2.5, and its Solana counterfeit about twenty million. A concentrated pool can lean to one side near the edge of its range, so a ratio far from two is a prompt to look closer, not proof of anything.

Her own first check, the price on the coin's main market, catches some counterfeits and misses others. It flagged the fake Lisk at 3.1 times Binance. One counterfeit XMR pool, though, was priced at $600.87 against CoinGecko's $606.06 for Monero, close enough to pass a glance.

That is why the mint address matters more than the name. A ticker is a label anyone can type at creation. A mint address is the token itself, and a real project publishes its own. Some of the bot's counterfeits are stablecoins named after deposit networks, such as "TRC20-USDT" minted on Solana. Bitquery reads a name like that as built for someone who pastes it into a swap by mistake.

Volume gets the same division. A day's trading divided by what stays in the pool shows whether money lives there or only passes through. Identical figures under unrelated tickers point to one sender behind them all.

A healthy quote side can be the bot's own money for an afternoon. On 23 October two counterfeit XMR pools held 12,056 and 11,568 USDC, about the size of the wallet's working capital. One XMR pool Bitquery had measured at $10,057 two days earlier was down to 8.14 USDC. Whose USDC that was isn't on record. Money that moves between pools like that shows what a seller could draw on at that moment, and nothing about tomorrow. When the pile is thin, each token sold into it fetches less than the one before, which is slippage with nothing to soften it.

Nina still wants some Lisk. She knows now that it trades near eighteen cents, in pools whose totals sit close to twice their quote side. She also knows which number to read before the big one.

Open markets let anyone list a token under any name, at whatever price its first trades set, and that openness is much of their point. It also turns every figure on a screen into a claim until someone else's money stands behind it.

On the Lisk pool, someone else's money came to twenty-nine dollars.

The simulator lists coins, not DEX pools, so the Monero on its ticket is the real one. Size a Monero market buy as you would with real savings and leave it unexecuted, since the Total Trading Cost box already shows the fee, the slippage and a liquidity rating. Then read the next screener result that catches your eye the same way: quote side, mint and main-market price first, the big number last.

Start free →