The 7.5 ETH rug template: 30+ Ethereum scams ran the exact same script in May 2026

If you only watch the headline rugs, you miss the assembly line.

The top two Ethereum rugs of the past 30 days, Arc Network (ARC) at 121 ETH profit and Inkon Chain (INK) at 119 ETH profit, are outliers. Above 50 ETH lifted in a single LP yank, you are looking at someone who actually marketed their pool, brought in real exit liquidity, and chose their ticker carefully.

Everything below that line tells a very different story. Pull the next 30 rugs from our rug metrics endpoint sorted by ETH profit descending, and you find a number that should not be possible by accident:

17.5 ETH withdrawn, 10 ETH initial deposit, 7.5 ETH net profit.

Not 7.4. Not 7.6. Exactly 7.5, repeated across 30+ different contracts deployed by what looks like 30+ different wallets, with names ranging from FIFA to Hermes Agent to Bankcoin to Baby Wojak. Same number. Different ticker. Different lifetime. Different scoring profile. Same rug.

Here is what the leaderboard looks like once you sort by profit and skip the first two outliers (data: May 1 to May 27, 2026, source: our analyzer’s /api/analytics/rug-metrics?since=30d):

Token	Symbol	Profit (ETH)	Initial LP	Withdrawn	Rugged at
Inkon Chain	INK	119.32	20.0	139.32	May 22
Zest Protocol	ZEST	11.65	10.0	21.65	May 26
Hermes Agent	HERMES	7.58	10.0	17.58	May 26
FIFACoin	FIFA	7.58	10.0	17.58	May 22
FIFA	FIFA	7.53	10.0	17.53	May 23
Bankcoin	Bank	7.53	10.0	17.53	May 22
fotbol	fotbol	7.53	10.0	17.53	May 24
Boomerang	Boomerang	7.52	10.0	17.52	May 24
Snowy	SNOWY	7.50	10.0	17.50	May 23
WIKI CAT	WKC	7.50	10.0	17.50	May 21
World Cup Coin	WORLDCUP	7.49	10.0	17.49	May 22
Elien Musk	ELIEN	7.56	10.0	17.56	May 12
SpaceX XAI	SpaceXAI	7.77	10.0	17.77	May 8
Giggle Cat	GICAT	7.53	10.0	17.53	May 13
Purple Ethereum	PETH	7.53	10.0	17.53	May 19

That’s 13 contracts from a single 30-day window. Sort by profit further down and you get another 20+ in the 7.40-7.60 ETH band, all with the same 10 ETH initial deposit. We stopped pasting after 15 rows because the picture is already clear.

Why 7.5? It’s the gas-floor of viable rug pulls

The 17.5 ETH withdrawal target is not random. It’s what you get when you reverse-engineer the minimum viable rug on a Uniswap V2 or V3 pool with 10 ETH of seed liquidity.

A scammer doing the napkin math has three constraints:

Initial bait - you need enough ETH in the pool so DEX aggregators (Uniswap UI, 1inch, CowSwap) actually surface your pair. Below ~5 ETH the slippage is so brutal that no organic buyer touches it. Below 8 ETH the bot scrapers (us included) flag the pool as low-priority. 10 ETH is the smallest “looks alive” seed.
Exit pressure ceiling - at 17.5 ETH in pool, the price has roughly 1.75x’d from the seed price (Uniswap V2 invariant x * y = k, when x goes from 10 to 17.5, y drops by 43%). That’s the headline gain that draws buyers in over a few hours. Push for higher and you either need real marketing (Telegram pump groups, paid shills, X engagement bait), or you start getting flagged by scam-watching bots that look for lp_withdrawal_pct > 90% on contracts under a week old.
Gas + deployer cost floor - deploying the contract, seeding the pool, the LP yank transaction, and the wallet hop to launder costs the deployer roughly 0.05-0.15 ETH in gas on a normal mainnet day, plus the 0.5 ETH they typically pay to fund the deployer wallet from a mixer. Net profit below 5 ETH stops being worth the operator’s time.

7.5 ETH net is the sweet spot. It clears the gas floor by ~10x, it doesn’t require any marketing effort, and it terminates the contract before the scam-watchers escalate the risk score above 70 and start tweeting about it.

It’s the break-even rug at scale. You can run it 50 times in a month with one deployer-funder loop and clear 375 ETH (~$940K at $2,500 per ETH) without ever needing to come up with a creative angle.

The naming is decorative, the numbers are the script

What we want to make explicit: the ticker doesn’t matter to the script.

FIFA (May 22 and May 23 - two separate contracts), fotbol (May 24), World Cup Coin (May 22) - World Cup 2026 starts in 13 days. They picked the hype window. Same exit price as the others.
Elien Musk (May 12), SpaceX XAI (May 8) - Elon brand-jacks, evergreen. Same exit price.
Hermes Agent (May 26), Giggle Cat (May 13), Snowy (May 23) - generic AI/meme tickers, no brand attached. Same exit price.
Bankcoin, Purple Ethereum, WIKI CAT, Boomerang - fully generic, no hype tie-in. Same exit price.

If you were doing the rug pulling, would you spend hours on the ticker if it didn’t change the outcome? Probably not. That’s why the wave is so wide and so cheap to ship. The deployer just runs the script with whatever string is on top of yesterday’s Twitter trending tab and lets the numerical machine do the rest.

The output is so uniform that you can use it backwards as a detection signal: if you see a fresh Uniswap V2 pair with 10 ETH initial liquidity and no LP lock, the expected exit-rug profit is 7.5 ETH within 1-2 days. Whether the ticker is $FIFA or $RANDOMMEME, the operator clearing it expects roughly the same payday.

This is also why we don’t run “per-token” alerts on our signals page - the only thing worth alerting on is the template, not the name attached to the template.

How long does it run before it pulls?

Lifetime varies in a wider band than profit does:

Fastest in the 7.5 ETH cohort: Yotsuba (May 4) at 34 blocks (~7 minutes). Deployer didn’t even wait for organic flow; pulled immediately after a sandwich-bot bought in.
Slowest in the cohort: Chainalysis (May 20) at 9,999 blocks (~33 hours). Took a day and a half to fill the bag, but the exit price was identical.
Median lifetime: roughly 1,200-2,500 blocks (4-8 hours).

The blocks-to-rug gap doesn’t change the takeaway. The deployer doesn’t choose when to rug based on time - they choose based on reaching the 17.5 ETH pool target. Once the invariant says the LP is worth the threshold, the script fires.

This explains the data spread in our earlier post on why scams still work when 80+ on-chain signals exist: the median rug fires within seconds of the first sell, not because the operator is watching, but because the script is watching getReserves() and triggers on the invariant.

The 17.5 ETH out / 10 ETH in template as a detection filter

We added this exact pattern to our analyzer two months ago as one of the 52 signals. The flag fires when a contract has all of:

Initial pool deposit between 8 and 12 ETH
Pool age under 7 days
LP withdrawal percentage above 80%
No LP lock (no Unicrypt / TeamFinance / PinkLock / etc. lockup)

That filter alone, run against the past 30 days, would have caught 31 of the 45+ rugs we counted with this signature. The remaining 14 had cosmetic differences (8 ETH initial instead of 10, partial withdrawal that resumed later, fresh pair re-deployment) but still cleared in the 5-9 ETH profit band.

If you paste any of the 30+ contract addresses above into app.rektradar.io, you’ll see the analyzer flagged the template signature in the risk breakdown - usually under low_liquidity + lp_unlocked + mass_funder if the deployer was seeded from one of the 11 known funder clusters.

What this means for buyers

Three operational takeaways from the data:

Initial liquidity around 10 ETH is the modal scam configuration. It’s the median budget for a scam operator running the assembly line. If a pool has exactly 10 ETH initial liquidity, no other context (good team, real KYC, locked LP) - assume rug template until proven otherwise.
Lifetime is not a safety signal. A scam that’s been live for 33 hours is still on script. The script doesn’t care how long it has to wait; it only cares about hitting the 17.5 ETH invariant.
The ticker is misleading by design. The operator is running 1 of 5 templates and renaming the output. $FIFA and $BABYWOJAK and $ELIEN are the same product to the operator. Stop trading any of them as if the name matters to the contract logic.

The rebuttal we get when we publish this kind of data is “well I’m careful, I only buy from people I trust, I always check the chart”. The 7.5 ETH template doesn’t need the chart - it needs the next person who saw the ticker on a Twitter timeline at 2am and figured it might be early. There are 30+ of those people in the past 30 days, each one converted into roughly 0.25 ETH of operator profit (assuming the typical 30 buyers per rug split).

Paste any contract address into app.rektradar.io for the full flag breakdown, or use the rug pull checker for the no-signup version.