Bitget CEO says $388M exploit came from third-party security product credential leak
The exchange says private keys and cold wallets were not compromised, but recovery totals and attribution remain unverified.
Bitget CEO Gracy Chen said the exchange’s reported $388 million exploit was enabled by a vulnerability in a third-party security product that exposed “high-level internal credentials.” Bitget says the attacker used those credentials to send fraudulent withdrawal commands, while recovery totals and a possible North Korea link remain under investigation.
Key Takeaways
- Bitget’s CEO tied the reported $388 million exploit to a vulnerability in a third-party security product that exposed “high-level internal credentials.”
- The attacker used those credentials to issue fraudulent withdrawal commands, while Bitget says private keys were not compromised and cold wallets were not affected.
- Unauthorized transfers were detected from several hot wallets on Sept. 24, prompting a temporary withdrawal pause and an initial estimate of about $352 million affected.
- Verified recovery and attribution are still pending, with Mandiant and SlowMist supporting an ongoing forensic investigation and North Korea indicators described as preliminary.
Bitget Pins the $388M Exploit on a Third-Party Security Product Flaw
Bitget CEO Gracy Chen said the exchange’s reported $388 million exploit traced back to a vulnerability in a third-party security product, which she said allowed an attacker to obtain “high-level internal credentials.” The claim matters because it frames the incident as an access-control failure upstream of Bitget’s wallet infrastructure, rather than a direct compromise of wallet private keys.
Chen said the attacker used the exposed credentials to issue fraudulent withdrawal commands. In practical terms, that implies the attacker did not need to break cryptography to move funds. They needed the kind of internal permissions that let withdrawals be initiated and approved through normal operational pathways.
Bitget has not identified the third-party security product or described the vulnerability beyond the credential exposure. That leaves a key gap for counterparties trying to map residual risk, because “third-party product flaw” can mean anything from a misconfiguration to a vendor-side breach, and the remediation path looks different depending on which it was.
Hot Wallet Drain, Withdrawal Pause, and the $352M vs. $388M Gap
Bitget detected the incident on Sept. 24 after unauthorized transfers from several of its hot wallets, then temporarily suspended withdrawals. A hot wallet is connected to the internet and typically used to service day-to-day withdrawals, which makes it operationally useful and structurally more exposed than offline storage.
At the time, Bitget initially estimated that about $352 million in assets were affected. The later figure referenced by Chen is $388 million. The source material does not explain what changed between those numbers, whether the delta reflects additional wallets identified after the first estimate, price movements during the drain, or a revised accounting of what “affected” means.
For traders, that gap is not cosmetic. Loss estimates are often revised in the first days after an exploit as internal ledgers are reconciled against on-chain flows and as exchanges separate attempted withdrawals from completed ones. Until Bitget explains the $352 million versus $388 million discrepancy, the market is left to treat both figures as provisional labels for the same event rather than a cleanly audited total.
Keyless Breach: What Bitget Says Wasn’t Compromised—and the Controls It Tightened
Chen said Bitget’s private keys were not compromised and its cold wallets were not affected. Private keys are the secret cryptographic material that controls the ability to move funds from a wallet. Cold wallets are kept offline to reduce hacking risk and are typically where exchanges store a large share of customer assets.
That distinction is the mechanical heart of Bitget’s narrative. If private keys were stolen, the attacker can usually move funds directly from the wallets those keys control, and the exchange’s main defense becomes wallet architecture and key management. If keys were not stolen, the failure mode shifts to internal access paths, including who or what can initiate withdrawals, how approvals are enforced, and what monitoring catches anomalous behavior.
Bitget said it addressed the security flaw and tightened withdrawal controls after the incident. The exchange described three specific changes: restricting internal access, adding independent verification for withdrawals, and increasing monitoring for unusual activity. Those are workflow controls, not wallet primitives, and they are designed to reduce the blast radius of credential exposure by forcing additional checks before funds leave.
The cross-chain angle remains a live routing risk. Bitget previously called on THORChain, a decentralized protocol that enables swapping assets between different blockchains, to refuse services to addresses linked to the attack as Bitget tried to prevent stolen assets from being moved. THORChain has said it cannot selectively blacklist individual addresses. Chen acknowledged that constraint, saying: “We understand that THORChain operates as a decentralized protocol and has said that it cannot selectively blacklist individual addresses. We respect the technical constraints of different networks and are not asking any protocol to take actions that are not technically possible.”
What Traders Should Monitor Next: Verification, Forensics, and Attribution
The next concrete signal is Bitget’s release of verified totals for assets recovered or frozen. The exchange has not disclosed how much stolen crypto has been recovered or frozen, and Chen said some assets have been frozen with help from other industry participants, but totals will be published only after verification.
A second milestone is any update from the independent forensic investigation Chen said is being supported by Mandiant and SlowMist. A forensic investigation is a technical review of systems and transaction traces meant to determine how the attack occurred and, when possible, who was behind it. Chen said earlier suspicion that North Korea may have been involved was based on “preliminary indicators,” adding: “What was shared previously was based on preliminary indicators identified during the investigation,” and “Those indicators are still being assessed. Mandiant and SlowMist are supporting the independent forensic investigation, and that work is ongoing. We will share further findings as they are verified.”
Two other open questions are still gating clarity. Bitget has not named the third-party security product implicated in the credential exposure or detailed remediation beyond the tightened withdrawal controls. The exchange also has not explained why the affected-asset figure shifted from an initial estimate of about $352 million to the later $388 million figure.
My Read: Counterparty Risk Isn’t Just About Keys—It’s About Internal Access Paths
The filing-equivalent detail here is Chen’s insistence that private keys and cold wallets were not compromised, because it pushes the incident into a category traders often underweight: operational access paths that can still produce valid-looking withdrawals. If the attacker really issued fraudulent withdrawal commands using “high-level internal credentials,” then the risk model is less about wallet architecture and more about who can authorize movement, how many independent checks exist, and whether monitoring is tuned to catch “legitimate” actions that are illegitimate in intent.
The threshold that matters is verification. If Bitget publishes audited recovery or freeze totals and the Mandiant and SlowMist forensics converge on a clear root cause and attribution, this shifts from a headline-driven hack narrative into a measurable counterparty-risk repricing tied to internal controls and cross-chain exit routes.
