Bitget Hack: $387 Million Stolen in Hours, AI Tracing Speeds Up Investigation
The Bitget hack, which occurred on September 24, resulted in the theft of $387 million. The funds were quickly moved across multiple networks, making it a challenging task for investigators to track the stolen money. Chainalysis investigators used in-house AI and custom automation to speed up the process, compressing over 20 hours of manual bridge reconciliation into under 10 minutes.
The stolen funds were initially split across four chains, with Ethereum accounting for 49.7%, XRP 40.8%, Zcash 7.6%, and Tron 1.8%. The attackers used various laundering mechanisms, including cross-chain liquidity protocols, messaging protocols, instant swaps, and laundering services. The stolen XRP created a particularly messy trail, with the attackers moving it through a cross-chain liquidity protocol to receive Bitcoin on another network.
Chainalysis stated that its use of AI did not replace investigators, but rather, experts defined the logic, reviewed outputs, and directed the investigation while automation handled repetitive tracing work. The attack is attributed to North Korean actors, which pushed the total crypto stolen by such actors in 2026 past $1 billion.