Weekly rug roundup #2 (May 25-31, 2026): 780 rugs, 1,476 ETH extracted, two whales ate half the week

Recap of every rug pull on Ethereum mainnet from May 25 to May 31, 2026. 780 contracts drained 2,418 ETH (operators kept 1,476 ETH ≈ $3.7M). Two outliers - OpenTrade and Cycles - extracted 680 ETH between them, accounting for 46% of the week. The other 778 rugs ran the 7.5 ETH assembly-line template we documented last Wednesday.

Second issue of the roundup. Every Saturday we publish what rugged on Ethereum mainnet over the past 7 days - names, addresses, dominant flags, and short case studies on the most telling deployments. Goal: a stable weekly receipt of the rug-pull pipeline so the data compounds.

The week in numbers (May 25-31, 2026)

DayRugs flaggedETH withdrawn (operator)Net profit
Sun May 25105199.4 ETH105.4 ETH
Mon May 26117275.9 ETH162.0 ETH
Tue May 27127223.0 ETH128.2 ETH
Wed May 281441,006.4 ETH550.5 ETH
Thu May 2989129.2 ETH60.9 ETH
Fri May 30103506.0 ETH385.6 ETH
Sat May 318078.1 ETH68.3 ETH
Week total7802,418 ETH1,476 ETH

1,476 ETH net = roughly $3.7M at the May 31 ETH/USD price. Down from week #1 in count (1,469 high-risk contracts last week) but up in extracted value, driven entirely by two outlier days: May 28 alone saw 550 ETH withdrawn - more than the previous three days combined.

That spike has a single cause.

Two whales ate 46% of the week

Of the 1,476 ETH net profit, 680 ETH (46%) came from just two contracts:

TokenSymbolProfit (ETH)ScoreRugged
OpenTrade$OPNTR388.9870/100May 28
Cycles$CCLS291.6280/100May 30

These are both brand-jacks of real projects. OpenTrade is a legit institutional credit protocol; the rugger registered an ERC-20 with the same name to catch retail buyers searching the brand. Cycles mirrors the same playbook against a different real project. Together they account for almost half the week’s total operator profit - without them, the median rug of the week was the same 7.5 ETH template we documented in Wednesday’s 7.5 ETH rug template article.

Pull these two outliers out of the math and the remaining 778 rugs averaged 1.02 ETH net profit per contract. Add them back in, the average jumps to 1.89 ETH - pure outlier bias from two whales.

The standard rugs after the outliers

Sorting the remaining contracts by ETH profit shows the script-driven cluster the previous article identified - the 7.5 ETH gas-floor template:

RankTokenSymbolProfit (ETH)Score
3Zest Protocol$ZEST37.6370
4Moon Pay$MOON34.6370
5Stablehouse$HOUSE7.7670
6Memecoin Packs$PACKS7.6570
7Hermes Agent$HERMES7.5880
8grail$GRAIL7.4960
9Bamboo$BAM6.7270
10Sushi Roll$SUSHI22.2580

The cluster around 7.5 ETH ± 0.5 is exactly the gas-floor of viable rug pulls that we showed in Wednesday’s piece. The arithmetic is unchanged: 10 ETH initial LP, 17.5 ETH withdrawn on exit, 7.5 ETH net once gas and the seed deposit clear. Different ticker, same operators iterating on what is profitable enough to ship at scale.

Five brand-jacks in the top 10 alone: Zest Protocol (real Stacks BTC stablecoin), MoonPay (the fiat onramp), Stablehouse (institutional yield), Memecoin Packs (real card-trading project), Sushi Roll (riffing on SushiSwap). The pattern is the same one we documented in the top 10 brand-jacked tickers: a known name + an unverified contract + 10 ETH of seed liquidity, every week, on rotation.

The emoji-symbol distribution trick

Three contracts this week shipped tickers that begin with an emoji character:

This isn’t decorative. DEX aggregators sort search results alphabetically and Unicode emoji code points sit far before standard ASCII letters in the sort order. Putting 🧊 in front of a ticker pushes that token to the top of the search results when a retail buyer searches anything close to the name. We added a flag for it (emoji_symbol) two weeks ago after spotting the pattern on a $UMI emoji-variant. Three more deployments this week confirms it’s a script-level technique, not a one-off.

What this means for buyers

The numbers compound in a way that’s worth restating from last week:

  1. Brand-jacks make up half the top 10 every week. $OPNTR / $MOON / $HOUSE / $PACKS / $SUSHI2 this week. $WILMA / $TEMPO / Golden Tempo last week. The script doesn’t care which name is trending - it just rotates onto the next one that has Google search volume. The defence is the same: never buy a token because you read the ticker on Twitter; paste the contract address into a scam detector first.

  2. The 7.5 ETH template is the floor. Most operators ship contracts they expect to clear roughly 7-10 ETH on, full stop. The 388 ETH whale rugs like OpenTrade are statistical noise - they happen when a brand-jack accidentally catches a real influencer mention. Most of what’s deploying is mechanical: 10 ETH bait, 17.5 ETH exit. Knowing that lets you predict, when you see a fresh 10 ETH pool, where it’s going to end up.

  3. Emoji-prefix symbols are an active surface. If you see a token whose ticker starts with a non-ASCII character on a DEX aggregator, that placement was engineered. Skip it.

Paste any contract address into app.rektradar.io for a free score, or use the rug pull checker for the no-signup version.

See also

Next issue lands Saturday June 7. Follow @mik3fly__ on X for the heads-up.