Historical source report: PANews, published November 6, 2026, at 19:06 (UTC+8), covering Changpeng Zhao’s November 3 appearance at the Oxford Union.
At an Oxford Union discussion billed around his book Freedom of Money, Binance founder Changpeng Zhao faced a more immediate question: why should users trust the exchange after a major security failure?
According to Binance’s November 3 investigation update, as reported by PANews, the incident involved approximately $6.34 billion in digital assets. About $2.85 billion in related assets had been frozen or intercepted by that disclosure. Binance said it had provisioned enough funds to cover losses and was progressively replenishing affected assets.
Those figures describe assets involved and assets frozen or intercepted—not an independently established final net loss.
Who bears the loss?
The November 3 discussion took place five days after the incident, according to PANews. Over nearly an hour and 19 audience questions, Zhao’s central position was that users should not pay for failures inside a centralized exchange.
He acknowledged shortcomings in Binance’s security and operations and said the sophistication of an attack did not excuse the platform’s responsibility. He emphasized communication after the incident and Binance’s commitment to absorb user asset losses.
PANews described the response as drawing on the SAFU user asset security fund, insurance funds and company resources. Zhao put SAFU’s size at roughly the billion-dollar level during the event. The report cautioned that some funding details were his on-stage estimates, not independently audited figures.
The distinction matters: a commitment to make users whole is not the same as independent verification that every affected balance has already been restored. The supplied report records Binance’s provision and replenishment claims but does not establish a completed, independently audited reconciliation.
Security assurances stopped short of a guarantee
Zhao declined to promise that another attack could never happen. He said such an assurance would be unrealistic for a complex information system.
According to his account, the attackers may have spent approximately five to six months preparing. They used methods including fake recruitment and social engineering to approach people inside the business, then used AI to accelerate their understanding of systems. He also said some basic operational security procedures at Binance had not been strictly followed.
His proposed direction was proactive defense rather than a promise of perfect security. PANews also reported his comments on cooperation involving AI companies, cybersecurity specialists, regulators and law enforcement. Whether that cooperation becomes a durable defensive arrangement remained an open question.
Visible transfers do not settle custody questions
Zhao contrasted blockchain visibility with disclosures by traditional financial institutions, noting that outside analysts can sometimes identify anomalies before a platform does.
That visibility can support scrutiny, but it should not be confused with a complete account of customer obligations. Observable asset movements alone do not establish liabilities, legal ownership or the adequacy of customer protections.
A related distinction arose in questions about law-enforcement freezes. Zhao separated the technical ability to transfer assets on-chain from the use of centralized custodial platforms. The ability to make a blockchain transfer does not mean the real-world transaction route sits outside judicial jurisdiction.
Asked whether users were trusting Binance or its founder personally, Zhao said protection of customer assets should be embedded in the platform’s rules and conduct, rather than depend on whether a particular founder remains in office. PANews noted that he was no longer CEO.
Crypto and AI: potential, not proof
Beyond the incident, Zhao argued that crypto could address difficulties with account access, cross-border payments, inflation and round-the-clock transactions. He also acknowledged that this was an argument about technical potential—not proof that every crypto financial service outperforms banks. Fees, compliance, custody and consumer protection remained practical issues.
He described AI as both a security threat and a possible source of demand for digital payments, particularly if autonomous agents generate large volumes of small transactions. On stablecoins, central bank digital currencies and Bitcoin, his position was that different instruments serve different needs.
The discussion offered no firm timetable for mainstream adoption and removed none of the underlying security risks. For centralized exchanges, its clearest issue was institutional: whether a public promise to protect users is backed by funding, effective controls and verifiable execution—not simply confidence in a founder.