Binance Hack: Withdrawals, Frozen Funds and DeFi Risks12 min read2,354 words

Binance Hack: Withdrawals, Frozen Funds and DeFi Risks

Explore Binance’s withdrawal data, the response to its $6.3B hack, and the THORChain dispute over freezing funds, decentralization and DeFi risk.

Binance hackcrypto exchange securityTHORChaincrypto regulationDeFi risks
Binance Hack: Withdrawals, Frozen Funds and DeFi Risks

Binance's $6.3B Hack Tests Exchange Trust and DeFi Neutrality

Binance reopened withdrawals across 15 blockchain networks after attackers stole an estimated $6.3 billion from its hot wallets. The immediate question was simple: would customers stay or flee? Binance said it processed about $4.65 billion in withdrawals and received $1 billion in deposits after reopening. Founder Changpeng Zhao said the first 12 hours brought a net outflow of just 2.1%. But DeFi Llama's wallet tracker showed $5.8 billion in outflows over 24 hours - a bigger number that may include internal transfers or movements between wallets.

Those figures don't automatically contradict each other. They underline a more annoying reality: during a crisis, people can look at the same on-chain activity and come away with very different conclusions - customers, analysts, and regulators included. And the stolen-funds response raised a tougher question for decentralized finance: when law enforcement asks, should protocol operators block transactions?

This article breaks down what the withdrawal numbers can - and can't - tell you, how the stolen funds were frozen and moved, why THORChain's position matters, and what the incident signals for exchange security and crypto regulation.

What do Binance's withdrawal figures actually show?

Binance's reopening was a stress test for customer confidence, but there isn't one clean, public number that answers how much money "fled" the exchange. Binance reported roughly $4.65 billion in withdrawals and $1 billion in deposits after reopening. Zhao said net outflows in the first 12 hours were 2.1%. DeFi Llama, meanwhile, reported $5.8 billion in outflows over 24 hours.

The distinction is basic: gross withdrawals aren't the same thing as net outflows. If users withdraw $4.65 billion and others deposit $1 billion, that's not the same story as a $4.65 billion net loss. Also, wallet trackers can log assets moving between wallets controlled by the same exchange. New hot wallets, internal rebalancing, and transfers between custody arrangements can make movement look like a customer exit even when it isn't.

Why the numbers diverge

Exchange disclosures and blockchain tracking use different inputs. Binance can pull from its own account and deposit records. External trackers infer activity from wallet addresses they believe belong to the exchange. That inference can be helpful, but it's not perfect - and it doesn't reveal what users meant to do.

A customer might shift assets to self-custody as a risk-management move, not because they're abandoning Binance. Some flows could be operational rather than user-driven. And a small net outflow doesn't prove that everyone remained comfortable with the exchange's controls.

So the real questions are more specific than "How many billions left?" For example:

  • Is the number gross or net, and what time window does it cover?
  • Does it count deposits, internal transfers, or newly identified wallets?
  • Do withdrawals run across the networks customers actually use?
  • Can the exchange show how it's meeting withdrawal demand and replacing what was lost?

Binance said it replenished about 95% of the stolen crypto using corporate reserves, insurance, and half of its Secure Asset Fund for Users (SAFU). That matters for short-term liquidity and confidence - but it doesn't, on its own, answer how the attack happened, whether the failing controls were actually fixed, or whether the same playbook would work next time.

Key takeaway: treat exchange flow numbers as clues, not verdicts. The method and context matter as much as the headline amount.

How were stolen funds frozen and moved?

The FBI attributed the attack to TraderTraitor, a group linked to North Korea's Lazarus Group. According to reporting, attackers gained access to delegated wallet permissions via a compromised laptop and stole funds from Binance hot wallets. Binance co-CEO Richard Teng disputed the idea that it came down to one laptop or one engineer's permissions, saying the investigation hadn't tied responsibility to an individual.

That distinction matters. A compromised device can be part of an attack chain without being the only place controls failed. In a real system, access usually depends on a mix of identity permissions, signing workflows, transaction rules, and monitoring. The key security question isn't just which device was compromised - it's how attackers could authorize a transaction of that size and why safeguards didn't stop them sooner.

Freezing stablecoins changes the attacker's options

Elliptic estimated that around $2.84 billion of the stolen assets - mostly USDC and USDT - was frozen on-chain. That would leave an estimated $3.33 billion that had not been frozen at the time of the report. These were estimates, and the numbers can change as assets move and addresses are identified.

Stablecoins issued by centralized companies can be frozen using administrative controls at specific addresses. That can slow down theft, but it also makes stablecoins behave differently than assets without an issuer that can apply that kind of control. Once attackers realize some funds are blocked, they have strong incentives to exchange freezeable assets for other cryptocurrencies.

The report said attackers borrowed about $44.6 million in ETH against stablecoins on Aave while trying to move funds. It also described cross-chain swaps via THORChain and Relay, including efforts reported to convert some stolen SOL into Bitcoin. Each route adds friction and creates more breadcrumbs for investigators - but it can also make it harder to immobilize the assets with a single freeze authority.

The FBI published addresses tied to the theft and asked node operators, instant exchanges, DeFi services, analytics firms, and other digital-asset businesses to block transactions involving those addresses or funds derived from them. That request put very different types of infrastructure into the same policy conversation, even though their ability - and legal obligation - to comply may not be the same.

The response shows both sides of on-chain transparency. Public ledgers can help follow funds, while issuer controls can stop some transfers quickly. But traceability doesn't guarantee recovery, and freezing doesn't automatically answer who should have the power to impose it.

Why did THORChain's founder oppose a freeze?

THORChain has been used in previous incidents to swap assets across blockchains without relying on a conventional centralized exchange. Its token, RUNE, reportedly rose 13% on expectations of more trading activity. At the same time, several THORChain routes - including Bitcoin, Ethereum, and Tron - were already halted because automated systems paused the protocol for reasons not directly tied to the Binance theft.

THORChain founder JP Thor, who said he was no longer involved with the project, argued that node operators shouldn't take the protocol offline or block transactions in response to the FBI's request. His view was tied to what he sees as the protocol's job: permissionless settlement shouldn't become dependent on an operator's judgment about which transactions are acceptable.

Neutrality has an operational cost

This isn't a simple "good vs. bad" decision. Refusing to intervene can protect censorship resistance. It can also turn the protocol into a path for moving stolen assets, which increases the odds of legal pressure on the people running the infrastructure.

The source report said about 50 people operate THORChain's 100 nodes, citing developer Chad Barraford. Node operators can vote to halt activity when a required threshold is met, and operators who don't want to process certain transactions can leave the network. THORChain can reportedly keep running with as few as four nodes. The practical takeaway is that decentralization isn't only something code provides - it's also something operators enact when transactions become politically or legally difficult.

There are also limits to what a protocol-level freeze can accomplish. If operators block a route, users can try a different bridge, swap service, or chain. A halt can interrupt legitimate activity along with illicit flows, and it may do little to recover assets already moved. Still, refusing to act doesn't remove the consequences. Law enforcement can pursue identifiable operators, and service providers may get forced into hard questions about their role.

A workable way to analyze it is to separate three issues:

  • Technical control: Can operators or governance participants actually pause or restrict the protocol?
  • Legal exposure: What duties apply to those operators in the jurisdictions where they live or work?
  • Protocol values: Is neutrality worth keeping when it lets transactions tied to a major theft continue?

There's no universal answer. But the incident makes the point clear: claims of decentralization need to be tested against real governance, operator concentration, and emergency procedures - not just the way things are described in documentation.

What does the hack mean for markets and crypto security?

The immediate market reaction was widespread but not definitive. The report said crypto-related stocks, including Coinbase, Circle, and Strategy, sold off sharply, while major crypto assets also fell. Bitcoin was reported at $77,745, down 3% for the week, and Ether at $2,260, down 3.5%. Those are snapshots from the reporting period, not proof of the hack's long-term impact.

There were also concerns that attackers had accumulated at least 18,000 Bitcoin and 430,000 Ether through the theft and subsequent swaps. Other address-based tallies mentioned in the report were higher. Large-scale selling could add downward pressure, but money attributed to attackers isn't the same thing as money that can be sold quickly. Some funds may be frozen, monitored, fragmented, or otherwise hard to liquidate.

The DEX argument needs a second look

Standard Chartered analyst Geoffrey Kendrick said the incident could push more users toward decentralized exchanges, pointing to Uniswap as a beneficiary. The logic is straightforward: if a major centralized exchange fails, self-custody and non-custodial trading start to look more appealing.

But switching venues doesn't remove risk; it changes where it shows up. Centralized exchanges concentrate custody and operational controls. DeFi moves risk into smart contracts, governance decisions, liquidity constraints, transaction-ordering issues, and the security of wallets and bridges. A DEX might not hold deposits the same way a centralized exchange does, but users still take responsibility for what happens when they sign transactions and route funds.

The same caution applies to any claims about "protocol safety." No publicly reported exploit doesn't prove a system can't fail. Security depends on code quality, integrations, privileged roles, operational processes, and how the protocol behaves under unusual conditions. Adoption should follow a real understanding of trade-offs, not a one-incident prediction.

Security lessons for exchanges and users

For exchanges, the core lesson is to design for containment, not just prevention. That means limiting what any single credential or device can do, requiring independent approval for high-impact transfers, monitoring unusual signing behavior, and testing whether emergency controls actually work when pressure hits. Binance's replenishment addresses one part of the crisis; the investigation and changes to the controls that failed address another.

For users, spreading assets across custody methods can reduce dependence on one platform - but self-custody shifts responsibility. A hardware wallet doesn't eliminate every risk, including setup and supply-chain concerns. Trezor urged customers who bought devices from a former authorized reseller, CryptoBilis, since March to move their funds after a customer reported a suspected implant in a Safe 5 device. Trezor said it hadn't examined that device yet and hadn't received reports of related losses. It also followed a separate report involving a Ledger device sold through the same reseller.

The practical response isn't to assume every exchange or wallet is unsafe. It's to match custody to your use case, limit exposure to any single point of failure, verify where devices came from, and be clear about what risks you're taking on yourself.

Conclusion: resilience depends on more than reserves

Binance's reopening produced several big flow numbers, but none of them alone proves whether customers trusted the exchange or whether recovery is complete. The more useful test is whether withdrawals stay reliable, the incident is explained clearly, and underlying security controls actually improve.

The THORChain dispute adds a different test: whether decentralized systems can preserve neutral settlement when law enforcement asks operators to intervene - and what operators risk when they refuse. Freezing funds can reduce damage, but it also concentrates power. Refusing to freeze can protect permissionless access, but it can also leave stolen assets moving and operators under scrutiny.

For exchanges, protocol teams, and users, the practical lesson is the same: resilience comes from understanding where control sits, limiting the damage from a failure, and being explicit about trade-offs before a crisis forces the decision.

Frequently Asked Questions

Binance reported about $4.65 billion in withdrawals and $1 billion in deposits after reopening. Zhao said the first 12 hours saw a 2.1% net outflow, while DeFi Llama recorded $5.8 billion in outflows over 24 hours. Different periods and wallet-tracking methods help explain why the figures differ.
Elliptic estimated that around $2.84 billion, mostly USDC and USDT, had been frozen on-chain at the time of its assessment. The remaining amount was estimated at $3.33 billion. These were time-sensitive estimates, and the amount that can be frozen or recovered may change as assets move and investigators identify addresses.
JP Thor argued that halting THORChain or blocking transactions would undermine its commitment to permissionless, neutral settlement. Critics may point to the risk of facilitating stolen-fund movement and legal exposure for operators. The dispute concerns both protocol values and the real-world responsibilities of the people who maintain it.
No. A decentralized exchange can reduce reliance on a central custodian, but users may still face smart-contract vulnerabilities, compromised wallets, bridge failures, poor liquidity or transaction-approval mistakes. Decentralization changes how risks are distributed; it does not make trading risk-free.

Related articles