Need help? Support
BITCOIN
Tether Dominance USDT.D

Single attacker drains $2M from AI tokens

Published 678 words 4 min read

TLDR

A single attacker has compromised shared infrastructure around several AI tokens, draining and minting roughly $2 million in assets and hitting some AI token markets hard.

  1. Security firms say the same wallet drained about $1.56 million in FET and minted roughly $452,000 worth of NTX on NuNet, for around $2.01 million in combined losses.
  2. NuNet (NTX) crashed more than 70 percent to an all time low, while FET dropped about 5 to 8 percent, exposing how thin liquidity magnifies cross project security failures.
  3. The incidents are linked to compromised signing keys and weak contract checks, with the same actor later abusing SingularityNET infrastructure, highlighting systemic risk across AI token ecosystems.

Deep Dive

1. One Attacker, Two AI Projects

On 20 September 2026, a single exploiter hit both Fetch.ai (FET) and NuNet (NTX) within roughly the same minute, with total losses near $2.01 million.

According to Blockaid and PeckShield, the attacker used a valid conversion authorizer signature to call the conversionIn function on Fetch.ais TokenConversionManagerV3 converter on Ethereum, draining about 8.72 million FET, valued near $1.55 to $1.56 million at the time. A BeInCrypto report notes that the same wallet, identified as 0x1572...c362, also received a large NTX mint from the NuNet deployer, worth around $452,000, bringing the combined impact to about $2 million in assets.

Investigators describe this as a trust and key management failure rather than a pure code math bug, since the contract accepted a valid signature without enforcing sufficient onchain checks on limits or backing. A detailed breakdown appears in a Bitcoin.com security writeup on the attacker who hits two crypto projects and walks with $2 million.

2. Market Damage To FET And NTX

The NTX side of the exploit was especially destructive because the attacker minted about 408.5 million new NTX, roughly 42 percent of supply, which collapsed the market. Reporting from BeInCrypto notes that NTX fell more than 70 percent in 24 hours, hitting an all time low around $0.000328, with some trackers showing even steeper percentage drops due to thin liquidity.

FET, backed by deeper liquidity and a larger market cap, dropped a more modest 5 to 8 percent across major feeds, still enough to shake confidence in AI narratives that had recently been rallying. The same BeInCrypto coverage points out that this exploit pushed Septembers total crypto hack losses above $333 million, after a string of DeFi incidents earlier in the month.

What this means

AI theme tokens can behave like high beta DeFi assets; when security breaks, small caps can effectively implode, while larger names bend but still feel the shock.

3. Shared Infrastructure And Ongoing Risk

The Bitcoin.com analysis notes that both attacks relied on compromised credentials, including a backend signing key used by SingularityNETs cross chain infrastructure, which tied the Fetch.ai converter and NuNet minting permissions to the same trust layer. Fetch.ai has said treasury funds, exchange wallets, and user self custody coins remained safe, but the converter and NTX supply are clear weak points.

Importantly, many of the newly minted NTX have not yet been sold, leaving a large overhang that could pressure the market if the attacker or any future controller of those keys begins to dump. A separate Binance News item, distributed via TradingView, reports that the same attacker later used a SingularityNET vulnerability to mint 260 million AGIX and 53.838 million WMTX on Ethereum, with holdings around $16.77 million in AGIX, ETH, and WMTX, extending the campaign beyond just FET and NTX.

What this means

If you follow AI tokens, the bigger risk is not only one projects bug but shared converters, bridges, and signing keys that link multiple AI ecosystems; a single compromised trust layer can ripple across several tokens.

Conclusion

A single actor exploiting shared AI token infrastructure turned what might have been a contained project bug into a multi asset event, wiping out much of NTXs value and denting FET.

The episode underlines that AI narratives do not reduce basic smart contract and key management risk; if anything, rapid experimentation increases attack surface.

For crypto users, the practical takeaway is to treat bridges, converters, and cross project AI stacks as critical risk points and to watch for official post mortems and key rotation plans before assuming the danger has passed.

Educational information only. Crypto markets are volatile and this is not financial advice.


Top