Why didn't Binance consider a Bitcoin chain reorganization to recover the stolen funds?
Technically, a reorg could in theory replace already-confirmed blocks, effectively making the hacked transaction "never happen" — but the cost of doing so is extremely high. It would undermine Bitcoin's core trust foundation of "once confirmed, irreversible," essentially damaging the very mechanism that makes Bitcoin trustworthy in order to recover a single loss. Then-CEO Changpeng Zhao publicly ruled this option out after talking with the miner community, in effect separating "protecting a small group of affected users" from "preserving trust in the entire ecosystem" — compensating users through the SAFU fund instead of shaking the blockchain's own immutability.
This decision also indirectly illustrates why a mechanism like SAFU matters: precisely because an exchange can't and shouldn't solve its own security problem by undermining the blockchain's trust foundation, an independent compensation mechanism the exchange itself is responsible for becomes that much more necessary.
What would have happened to the 2019 incident if the SAFU fund hadn't been large enough at the time?
This is a hypothetical with no official definitive answer, but it's possible to infer from the known mechanism design logic: SAFU is fundamentally Binance's commitment to keep building a reserve funded by trading fees, and if an actual loss had exceeded the fund's size at that moment, Binance would in theory still have been on the hook to make users whole — whether by covering the gap from its own corporate capital or taking some other measure, though that would fall outside what the SAFU mechanism itself was designed to handle. Part of why the 2019 incident could be fully covered was that the loss (roughly $41 million) fell within what the fund could absorb at the time — which is also part of why Binance has continued growing the fund's size since, and set a clearer $1 billion target with a dynamic replenishment mechanism in 2026.
This also circles back to an earlier point: the fund's raw size isn't the only thing worth looking at — whether that size has actually been matched against a real loss and successfully covered it is the more meaningful piece of information.
Did the users whose API keys and 2FA were compromised do anything wrong themselves?
Binance's explanation at the time was that hackers used phishing, viruses, and other multi-pronged techniques to obtain the information — meaning affected users likely fell victim to a sophisticated social engineering attack, not simply "user carelessness" as a straightforward attribution. A sufficiently sophisticated phishing attack can catch even security-conscious users off guard, which is also part of why mechanisms like SAFU are triggered based on "a flaw in the platform itself or the Hot Wallet getting hacked" rather than going back and litigating whether an individual user "shouldn't have clicked that link."
That said, this doesn't mean user-side protection doesn't matter. Enabling 2FA that isn't SMS-based (an authenticator app or Hardware Security Key, rather than SMS codes, since SMS is vulnerable to SIM-swap attacks) and staying alert to any link asking for account credentials are protections users can and should take themselves — complementary to, not a substitute for, the exchange-side payout mechanism.
Is Binance's security today meaningfully different from what it was in 2019?
Yes, and it's a direct result of this incident. Following the 2019 event, Binance publicly stated it would strengthen 2FA processes, API access controls, and withdrawal verification mechanisms — this before-the-fact prevention layer is complementary to SAFU's after-the-fact compensation role. The evolution of the SAFU fund itself also shows signs of continued strengthening — from a limited size at its 2018 founding to a clearly defined $1 billion target in 2026, converted into Bitcoin reserves that carry less correlated risk than holding the platform's own native Token. The mechanism itself has kept evolving over time rather than staying static after launch.
Worth noting, though: "the mechanism keeps getting strengthened" doesn't mean "a security incident will never happen again" — that's a guarantee no exchange can actually make. A more realistic way to view this track record is as evidence the exchange has demonstrated learning from past incidents and continuing to improve, not as proof it's now permanently safe.
Most exchanges' user protection mechanisms live entirely on paper — a Whitepaper, a webpage — and have never faced a real stress test. Until something actually goes wrong, no one knows whether the promise holds. Binance's SAFU fund is a rare exception: it genuinely faced a large-scale hack in 2019, and the entire episode left a complete public record, letting us walk through exactly how close "mechanism design" came to "mechanism actually holding up under pressure."
In May 2019, attackers used phishing, viruses, and other techniques to obtain a large number of Binance users' API keys, 2FA codes, and other account information. Binance later explained that the hackers showed unusual patience — executing a coordinated action through multiple seemingly unrelated accounts at the most opportune moment, successfully bypassing the risk controls in place at the time, and draining 7,000 Bitcoin from hot wallets, worth roughly $41 million at the time. The affected funds represented only about 2% of Binance's total assets at the time — meaning Cold Storage as a defense layer held for the vast majority of user assets, and the small slice that did get breached is exactly what put SAFU's reason for existing to the test.
After confirming the breach, Binance immediately suspended deposits and withdrawals and conducted a full security review — this decision was itself part of the mechanism design, stopping any further potential losses rather than letting trading continue while the system was still vulnerable. At the time, discussion emerged in the community about whether Binance would push for a Bitcoin blockchain reorganization to recover the funds. Then-CEO Changpeng Zhao, after conversations with several miners (including Jihan Wu), publicly stated Binance would not pursue this — a chain reorg would undermine trust in the Bitcoin network itself and could fracture the community. This decision drew a meaningful line: SAFU's entire purpose is to make users whole without needing to resort to something that damages the trust foundation of the broader ecosystem.
SAFU (Secure Asset Fund for Users) is an emergency reserve fund Binance established in 2018, designed from the start to be funded by allocating 10% of the platform's trading fee revenue on an ongoing basis, to prepare for a future security incident. The 2019 event marked the first time this mechanism was actually put to use since its creation — Binance announced that all affected users' losses would be fully covered by the SAFU fund, and no user ended up personally absorbing a loss. Worth noting here is the timing gap: from the fund's establishment in 2018 to this 2019 incident, less than a year had passed, meaning the reserve size at the time wasn't yet substantial — but it was enough to cover this roughly $41 million loss. The mechanism faced a real-world test before it had even built up to its intended scale, and it held.
SAFU used to hold a mix of assets including BNB and USDT. In late January 2026, Binance announced a plan to convert its entire $1 billion SAFU reserve from stablecoins fully into Bitcoin within 30 days — completing the final purchase of 4,545 BTC on February 12, bringing total holdings to 15,000 BTC, worth roughly $1.005 billion at a Bitcoin price of about $67,000 at the time. Binance also committed to replenishing the fund back to $1 billion if its value drops below $800 million due to Bitcoin price volatility. The logic behind this conversion is reducing a specific correlated risk — a protection fund holding the platform's own native Token can see that token's value crash in tandem if the platform itself runs into trouble. Bitcoin is one of the most liquid assets with the least correlation to any single platform's fate. This fully-Bitcoin reserve sits at a public on-chain address (1BAuq7Vho2CEkVkUxbfU26LhwQjbCmWQkD) that anyone can verify at any time via a Block Explorer.
2019 wasn't SAFU's only documented use. In December 2020, a vulnerability in the DeFi protocol Cover Protocol led Binance to pay out roughly $10.1 million from SAFU to affected Binance users. Looking at both instances together, SAFU isn't a one-time crisis PR move — it's a mechanism that has continued operating and continued getting used. This is worth more, when evaluating any exchange's protection mechanism, than a large-sounding fund size on its own: a fund of any size, if it's never actually been triggered, tells you nothing about whether the actual payout process is fast, reliable, or works at all.
Following the 2019 incident, Binance publicly stated it would strengthen two-factor authentication processes, API access controls, and withdrawal verification mechanisms — this kind of before-the-fact prevention is complementary to, not redundant with, SAFU's after-the-fact compensation role. What you should do the moment you notice a security incident yourself (immediately checking for account anomalies, enabling extra verification) and this exchange-level payout mechanism are two separate but equally important lines of defense.
For most exchanges' user protection funds, all you can really judge is how reassuring the whitepaper or website copy sounds. The genuinely decisive question is whether the mechanism has been tested by a real event, and what the outcome was. What SAFU demonstrated in 2019 wasn't just "there's money sitting somewhere" — it was fast decision-making when the event hit, full compensation, and continued hardening afterward. This is one of the rare cases where you can point to a specific timeline, a specific dollar amount, and a specific outcome to verify the claim — not a promise that only ever existed on paper. That doesn't mean this should be the only factor in choosing an exchange, but it's a genuinely verifiable reference point, not marketing language.
If the fact that a mechanism has actually been hack-tested and still paid out in full carries real weight in how you evaluate an exchange, you can sign up through Binance using referral code 206413162 — depending on current promotions, this may come with a trading fee discount (check the actual signup page for details).