How much ETH is stolen by scam tokens? $9.5M in 90 days -- our data

Real on-chain numbers: 2,220 rug pulls, 4,157 ETH net stolen from victims, worst day $1M+. Data from RektRadar's 69,000+ analyzed contracts on Ethereum mainnet.

Every week, articles claim “millions stolen by crypto scams” — but the numbers are usually vague estimates from security firms aggregating every type of incident: hacks, bridge exploits, phishing, and simple scam tokens all lumped together.

This article is different. These numbers come from RektRadar’s own on-chain tracking pipeline: real pool deposits, real withdrawals, real ETH amounts extracted from Uniswap V2 and V3 pools by scam token deployers.

The 90-day summary (Feb — May 2026)

MetricValue
Rug pulls tracked2,220
ETH deposited by scammers (liquidity seed)5,076 ETH
ETH withdrawn when rugging9,233 ETH
Net ETH stolen from victims4,157 ETH (~$9.5M)
Average per day46 rug pulls / 277 ETH stolen
Average per week1,900 ETH ($4.3M)

At the current ETH price ($2,279), 4,157 ETH = approximately $9.47 million extracted from retail buyers in 90 days — just from the rug pulls RektRadar has tracked on Ethereum mainnet.

The real figure is higher. Our pipeline catches pools that go through Uniswap V2/V3 factory events. Scams deployed on other AMMs (Sushiswap, Camelot, etc.) or that never create a pool are not counted.

How the math works

In a typical rug pull:

  1. Scammer deploys an ERC-20 token and adds initial liquidity (e.g. 1 ETH + 1 billion tokens)
  2. Victims buy the token on Uniswap — they send ETH into the pool, receive tokens back
  3. Scammer removes liquidity — they withdraw their original 1 ETH plus all the ETH victims contributed

The “ETH deposited” in our data is the scammer’s initial seed. The “ETH withdrawn” is everything they pull out. The difference — 4,157 ETH — is the net ETH that went from victims’ wallets to scammers’ wallets.

The worst days

Bar chart of the 5 worst days for rug pulls on Ethereum in April 2026. April 25 peaked at 469 ETH stolen across 216 rugs. April 27 reached 411 ETH, April 24 reached 394 ETH, April 18 reached 391 ETH, April 23 reached 373 ETH.
ETH extracted from rug-pull victims, worst 5 days. Source: RektRadar, Uniswap V2/V3, Apr 2026.
DateRug pullsETH stolenUSD equivalent
2026-04-25216469 ETHapprox. $1.07M
2026-04-27224411 ETHapprox. $937k
2026-04-24217394 ETHapprox. $898k
2026-04-23200373 ETHapprox. $851k
2026-04-1862391 ETHapprox. $891k

April 24-27 was a peak period: 4 days back-to-back with 200+ rug pulls per day, totalling over 1,600 ETH extracted (~$3.6M in 4 days). This coincides with a memecoin hype cycle — scam factories front-run narrative-driven buying.

Scale: 69,000+ contracts analyzed

MetricValue
Total contracts analyzed69,172
Liquidity pools monitored94,156
Unique deployer wallets mapped33,311
Deployer clusters (connected networks)491
Blacklisted scam clusters10
Honeypot rate14.94%
Contracts with risk score 61-10011,951

1 in 6 new ERC-20 tokens on Uniswap is a honeypot or high-risk rug pull. The median risk score across all analyzed tokens is 13/100 — the majority of deployments are legitimate — but the malicious tail accounts for the bulk of victim losses.

Scam factories: one bytecode, hundreds of thousands of deploys

The single most replicated drain contract blueprint in our bytecode database has been matched against 301,311 on-chain deployments. One scam template, hundreds of thousands of tokens.

This is how organized scam operations work: a factory operator develops one functional honeypot contract, then mass-deploys it under different names and tickers for each narrative cycle ($WAR, $PEPE, $AI, $TRUMP…). The bytecode is essentially identical — only the metadata changes.

Top scam bytecode patterns tracked:

Pattern typeOn-chain deploymentsTop flags
Drain blueprint301,311scam_factory_name, hardcoded_blacklist, misspelled_drain_function
New wallet rug194,358new_wallet, liquidity_at_creation
Backdoor mint96,031hidden_owner, unrestricted_mint

Connected scam networks

We track wallet clusters — groups of addresses connected by on-chain funding patterns. Our graph-crawler maps the ETH flow between wallets: when the same funder seeds multiple deployer wallets across different scam tokens, they appear in the same cluster.

Of 491 clusters identified, 10 are fully blacklisted as confirmed multi-token scam networks. A single blacklisted cluster is often responsible for 20-50+ rug pulls, all from fresh wallets funded from the same origin address.

Typical pattern: origin wallet → 5-10 deployer wallets → each deploys 1-3 tokens → rug within hours → proceeds flow back to origin. The whole cycle takes under 24 hours.

Extrapolated annual impact

TimeframeRug pullsETH stolenUSD (at $2,279/ETH)
90 days (tracked)2,2204,157 ETH~$9.5M
Estimated annual (×4)~8,880~16,600 ETH~$37.8M/year

This is a conservative lower bound for ETH lost on Uniswap V2/V3 alone from scam tokens RektRadar has detected. Add undetected scams, other DEXes, and non-pool scams (fake airdrops, approval drains) and the real annual figure is likely 2-3x higher.

How to protect yourself

  1. Always check the contract address before buying — never trust the token name or ticker
  2. Paste the address into app.rektradar.io — risk score in seconds
  3. Red flags to never ignore: honeypot, sell_failed, hidden_owner, scam_factory_name
  4. Check the deployer: has the same wallet deployed other flagged contracts?
  5. If a token is trending on Telegram or Discord and has no audit — assume it is a scam until proven otherwise

RektRadar runs three independent signals: on-chain buy/sell simulation, static bytecode analysis, and deployer graph traversal. A token that passes all three is significantly safer than one that has never been scanned.

The live stats are updated in real time at rektradar.io.