Bybit Crypto Exchange Hack: Technical Breakdown and Operational Implications

In February 2025, Bybit suffered one of the largest crypto exchange breaches on record when attackers drained approximately $1.5 billion in ETH from a compromised Ethereum safe multisig wallet. The incident exposed critical vulnerabilities in how centralized exchanges manage cold storage transition points and multisig authorization flows. This article examines the attack vector, operational failures that enabled it, and the structural changes practitioners should implement when evaluating exchange custody models.

Attack Mechanics and Entry Point

The breach targeted a Safe (formerly Gnosis Safe) multisig wallet during a routine transfer operation. Attackers compromised one or more signing keys, likely through targeted social engineering or supply chain infiltration of developer workstations rather than a cryptographic break. Once threshold signatures were obtained, the attackers initiated a withdrawal transaction that appeared legitimate to the remaining signers and automated monitoring systems.

The timing of the attack during a known operational window suggests reconnaissance. Bybit’s cold storage procedures involved periodic consolidation transfers where assets move from deeper cold storage to intermediate wallets before reaching hot wallets for customer withdrawals. This transition point represents the highest risk window because keys must be briefly exposed and multiple parties coordinate signing. The attackers exploited this predictable pattern.

Safe multisig contracts rely on offchain signature collection followed by onchain execution. The compromise occurred offchain, meaning no smart contract vulnerability was exploited. The Ethereum contract functioned as designed: once valid signatures met the threshold, the transaction executed. This distinction matters for assessing whether the risk was protocol level or operational security failure.

Cold Storage Architecture Weaknesses

Traditional exchange cold storage assumes physical isolation protects assets. In practice, cold wallets must periodically interact with networked systems to authorize transfers. Bybit’s model required threshold signatures from a known set of key holders, but lacked sufficient runtime verification of transaction intent.

The compromised wallet held an unusually large balance for a transitional custody point. Best practice calls for limiting intermediate wallet exposure to the minimum required for near term operational needs, typically no more than 24 to 48 hours of projected withdrawal volume. Concentrating $1.5 billion in a single multisig during a routine consolidation suggests either poor balance distribution or infrequent cold storage cycling.

Hardware security modules (HSMs) and cold signing devices do not prevent this attack class. If the operator authorizing the HSM signature is compromised or deceived, the HSM will sign a malicious transaction. The security model depends entirely on human verification of transaction parameters before signing. At scale, this creates fatigue and process shortcuts.

Authorization Flow Breakdown

Examine what failed at the approval stage. Multisig security assumes independent verification by each signer. In reality, many implementations use:

• A single administrator who prepares transactions and distributes them to signers
• Automated or semi automated signing by some threshold parties
• Time pressure during operational windows that discourages thorough verification

If Bybit’s process allowed a compromised administrator to present a malicious transaction as routine, and signers relied on that administrator’s representation rather than independently verifying the destination address and amount against expected parameters, the multisig threshold becomes a false security claim.

Robust signing flows require each signer to independently reconstruct expected transaction parameters from source of truth data (customer withdrawal queues, known operational addresses, scheduled consolidation amounts) and flag discrepancies. This demands tooling that surfaces mismatches automatically, not manual address comparison.

Customer Fund Segregation Reality

Exchanges typically claim customer funds are segregated and identifiable. The Bybit incident demonstrates the practical limits of this claim. While individual customer account balances appear segregated in the exchange database, the underlying blockchain custody is pooled. When a pooled cold wallet is drained, customer segregation exists only as database entries with no backing assets.

The speed and completeness of customer recovery depends entirely on the exchange’s capital reserves, insurance arrangements, and willingness to socialize losses. Bybit publicly committed to covering the loss, reportedly through a combination of reserves and potential external financing. This commitment, while commendable, highlights that custody at centralized exchanges ultimately depends on corporate solvency rather than cryptographic guarantees.

Practitioners evaluating exchange risk should treat custody claims as credit risk. The relevant questions become: what is the exchange’s capital ratio, what percentage of customer assets could they cover if a major wallet is compromised, and what legal jurisdiction governs insolvency procedures?

Post Breach Transaction Analysis

Blockchain transparency allowed realtime tracking of stolen funds. The attackers immediately began laundering operations through:

• Ethereum mixers to break onchain linkage
• Crosschain bridges to move funds to networks with different compliance frameworks
• DEX swaps to convert ETH into less traceable assets
• Distribution across thousands of addresses to complicate freezing efforts

The effectiveness of these laundering techniques varies by jurisdiction and asset. USDT and USDC issuer freezes blocked some laundered stablecoins, but ETH and decentralized assets cannot be frozen without validator coordination. Law enforcement recovery in crypto hacks typically captures less than 10% of stolen funds, concentrated in cases where attackers made operational security mistakes.

For users holding funds on Bybit at the time of the hack, the practical outcome was a temporary withdrawal suspension followed by resumed operations once replacement capital was secured. However, the incident triggered sustained withdrawals as users reassessed custodial risk.

Worked Example: Evaluating a Multisig Custody Claim

An exchange advertises “institutional grade 5 of 8 multisig cold storage.” You want to assess actual security.

Request or infer:

1. Signer independence: Are the 8 key holders truly independent entities, or are 6 of them employees reporting to the same executive?
2. Authorization workflow: Does each signer independently verify transaction parameters against a source of truth, or do they trust a transaction prepared by operations staff?
3. Transaction frequency: How often do cold storage transactions occur? Daily operations indicate higher key exposure than monthly.
4. Balance concentration: What percentage of total customer assets sit in any single multisig address? Lower concentration limits per incident loss.
5. Verification tooling: Do signers use software that automatically flags unexpected destinations or amounts, or rely on manual review?

If the exchange cannot or will not answer these questions with specifics, assume the multisig provides minimal additional security over a standard hot wallet with access controls.

Common Mistakes and Misconfigurations

• Treating multisig threshold as sufficient security without verifying signer independence and transaction verification processes. A 4 of 6 multisig where all 6 signers work in the same office under a compromised operations manager provides little defense.
• Assuming hardware wallets prevent social engineering. Hardware wallets secure private keys but do not validate transaction legitimacy. A deceived operator will authorize a malicious transaction even with a hardware wallet.
• Overlooking consolidation windows. Automated monitoring often focuses on hot wallet thresholds but may not alert on unusual cold storage movements during known operational windows when such transfers are expected.
• Conflating proof of reserves with custody security. Proof of reserves confirms an exchange controls specific addresses at a point in time but says nothing about key management practices or susceptibility to compromise.
• Relying on insurance claims without reading policy terms. Crypto insurance typically covers narrow scenarios (employee theft, specific hack types) with sublimits far below total customer assets. Policies often exclude social engineering.
• Ignoring settlement risk during recovery. After a major hack, an exchange may continue operating on fractional reserves during recovery. Customer withdrawals become a coordination game where late movers may face delays or haircuts.

What to Verify Before Relying on Exchange Custody

• Current proof of reserves attestation, ideally with Merkle tree inclusion proof for your specific account balance against the attested onchain addresses.
• Exchange capital adequacy ratio: what percentage of customer liabilities could be covered by equity and reserves if largest cold wallet is compromised?
• Wallet address rotation frequency. Exchanges that use the same cold storage addresses for extended periods signal poor operational security hygiene.
• Insurance policy details: coverage limits, exclusions, claims process, and insurer credit rating. Most exchange insurance covers a small fraction of assets.
• Jurisdiction and insolvency framework. Where is the legal entity holding your assets domiciled, and what customer protections exist in bankruptcy?
• Withdrawal processing time during normal operations. Slow withdrawals may indicate fractional reserve practices or operational dysfunction.
• Historical incident response. How did the exchange handle previous security incidents or service disruptions? Did they socialize losses or cover them from reserves?
• Multisig implementation specifics for any public addresses you can link to the exchange. Check current threshold, signer addresses, and recent transaction patterns.
• Third party security audit recency and scope. Audits older than 12 months have limited value given infrastructure changes.
• Regulatory compliance status in your jurisdiction. Licensed exchanges face higher operational standards and supervision, though this does not eliminate custody risk.

Next Steps

• Migrate high value holdings to self custody using hardware wallets with verified firmware, accepting the operational burden of key management. Centralized exchange custody is fundamentally corporate credit risk.
• Distribute remaining exchange balances across multiple platforms to limit per incident exposure. No single exchange balance should represent funds you cannot afford to lose access to for 30 plus days.
• Monitor your exchange exposure against published proof of reserves. Set alerts for withdrawal delays or unusual platform behavior that may signal liquidity stress following an undisclosed incident.

Category: Crypto Security