
Chainalysis Used AI to Trace the $387M Bitget Hack Back to North Korea
Chainalysis published an attribution report tying the September 24 $387M Bitget exchange hack to North Korea-linked actors, pushing DPRK's 2026 crypto theft total past $1 billion. The report details the attackers' laundering path: $387M moved in 23 transfers within three hours across Ethereum (49.7%), XRP (40.8%), Zcash (7.6%) and Tron (1.8%), with stolen XRP routed through a cross-chain liquidity protocol into Bitcoin rather than an exchange. Chainalysis also disclosed it used in-house AI automation to compress over 20 hours of manual bridge-transaction reconciliation into under 10 minutes, with humans still directing the investigation.
What it is
Chainalysis's attribution report on the Sept. 24, 2026 Bitget hack ($387M), the detailed post-theft laundering trail, and the specific response actions taken by other protocols and stablecoin issuers as the funds moved.
How it works
After the initial three-hour, 23-transfer split across Ethereum, XRP, Zcash and Tron, the attackers routed the stolen XRP (40.8% of the haul) through a cross-chain liquidity protocol rather than an exchange, converting it to Bitcoin on the other side. Tens of millions moved this way over roughly a day and a half before settling into attacker-controlled Bitcoin addresses that are now under watch. Separately, the attacker began shielding a portion of funds in Zcash's shielded pool to break traceability.
flowchart TD
A[Bitget hack: $387M, 23 transfers] --> B[XRP: 40.8%]
B --> C[Cross-chain liquidity protocol]
C --> D[Swapped to Bitcoin]
D --> E[Attacker-controlled BTC addresses, now monitored]
A --> F[Zcash: 7.6%]
F --> G[Shielded pool - trail obscured]Preconditions / Affected
Relevant to teams monitoring DPRK-linked laundering flows or evaluating cross-chain bridge/swap exposure: the laundering used instant swaps, messaging protocols, and liquidity bridges that don't require KYC'd exchange on/off-ramps, which is why the XRP-to-BTC leg was harder to interdict than a direct exchange deposit would have been.
What to do
The page names concrete real-world interdiction outcomes worth noting as reference points: Near Intents rejected over $50 million in swaps tied to the hacker (and was itself hacked days later); Thorchain continued processing the attacker's swaps; Circle and Tether froze roughly $318,000 in stablecoins tied to the theft. Teams building transaction-screening or swap-rejection logic now have a concrete precedent (Near Intents' $50M+ rejection) and a counterexample (Thorchain) to compare policies against.
Caveats
The attacker-controlled Bitcoin addresses are described as "now under watch" but the page does not state any funds have been frozen or recovered there. The Near Intents hack that followed its rejection of attacker swaps is reported without a stated causal link to the rejection itself.
Organisations named here
- Gracy Chen
Technologies named here
- Chainalysis in-house AI / custom automation
- Bitget exchange
- Chainalysis
- Elliptic
- NEAR Intents
- THORChain
- Ethereum
- Bitcoin
Techniques named here
- Rapid multi-chain fund dispersal
- Cross-chain laundering via liquidity/messaging protocols
- Cross-chain conversion to obscure XRP proceeds
- Use of shielded privacy pools
Threat actors named here
- North Korea-linked actors
Related in Blockchain

THORChain Exploit Report #2 — GG20 root-cause disclosure

Black Thursday revisited: MakerDAO keeper exploited for $500K
.jpg)