Coldcard Attack Tracking Shifts to Chinese AI as US LLM Restrictions Hinder Analysis

Galaxy Research pivots to Chinese open-source AI for on-chain tracking due to US LLM limits. Coldcard exploit attacks persist, with post-2021 single-sig funds at high risk of drainage.

Woofun AI reports that Alex Thorn, head of research at Galaxy Research, disclosed that restrictions on large language models in the United States impeded the tracking of stolen Bitcoin funds. Consequently, the team utilized open-source AI models from China to assist in on-chain analysis and asset protection. Thorn noted that the average time before stolen Bitcoin was moved was 3.18 years, with a median of 3.55 years, affecting mostly long-term holders. He described the regulatory constraints as "incredible" and intends to raise the issue at government and industry levels.

Thorn emphasized that exploit attacks targeting Coldcard keys remain active. Funds from previous attack waves stay untouched in attacker addresses, displaying systematic transaction patterns.

However, opportunistic attackers are now using complex splitting mechanisms to launder funds within hours. He warned that all single-signature Coldcard addresses generated after the March 2021 firmware upgrade will eventually be emptied, urging users to transfer their assets immediately.