When a co-CEO of one of the world’s largest crypto exchanges posts a 30-day performance screenshot showing a six-figure drop, people notice. Especially when that drop lands right around a multi-billion-dollar security incident. That’s exactly what happened with He Yi’s account summary, and the online reaction was fast, sharp, and full of accusations.
I’ve been watching these kinds of episodes for years, and the pattern rarely changes. A public figure shares numbers, the community fills in the blanks with the worst possible story, and suddenly the conversation shifts from market performance to questions of trust. In this case the story went further. Some traders claimed He Yi had moved funds out of Binance on the very day the exchange faced its major security breach. Others pointed to the fact that regular users were still dealing with limited withdrawal access while leadership seemed able to shift assets around. The gap between those two experiences felt unfair to many.
What Actually Appeared in He Yi’s Account Summary
On November 16 He Yi shared a clean 30-day overview covering October 17 through November 16. The headline number was hard to ignore: a total change of negative $101.32 million. Roughly $97.32 million of that sat under trading profit and loss. Another $3.99 million appeared under transfers and other adjustments. She added a short note that the pullback was temporary and that spot holders should not panic.
Here’s the part that got lost in a lot of the early commentary. Those numbers measure the change in account value over thirty days. They do not automatically prove that she realized $101 million in permanent trading losses. They also do not prove that any of the money left the exchange. Public dashboards like the one she shared rarely show individual transaction trails or final destinations. That missing detail left plenty of room for speculation.
Within hours the screenshot was circulating widely. One post asked the blunt question of who was actually making money while the broader market struggled. Another went further and claimed she had moved everything out on the day of the incident. A third suggested that earlier posts raising the issue had been pressured into deletion. None of those posts included transaction hashes or blockchain evidence of external transfers. Still, the timing alone was enough to keep the discussion alive.
Why the Timing Felt Especially Sensitive
The security event itself was no small matter. Roughly $6.3 billion in assets held in exchange wallets were affected after internal systems were compromised. Binance paused customer withdrawals while it investigated and began the recovery process. For ordinary users the days that followed involved checking network-by-network status pages and waiting for each chain to reopen. Bitcoin withdrawals and selected networks started returning on November 1. By mid-November the exchange reported that 87 networks were back online and that the remaining ones were targeted for completion shortly after. Corporate funds had been used to replenish the affected assets, and a fresh proof-of-reserves snapshot covered 198 tokens.
In that environment any visible movement of large balances by senior figures was always going to attract attention. Traders who could not move their own funds understandably felt the contrast. One commenter put it simply: customers were stuck while leadership appeared able to shift assets under the explanation of testing an AI strategy. Whether or not that perception was accurate, it was real enough to shape the conversation.
He Yi’s Direct Response
He Yi did not stay silent. In a November 16 reply she explained that she had moved assets into a sub-account specifically to test Binance AI Pro strategies. The testing phase was still being refined and was not running smoothly, so she later moved the funds back into regular spot trading. A separate account associated with the exchange reinforced the point, stating clearly that the money had not left the platform and pointing people back to her explanation.
During that time, I switched to a sub-account to try running strategies with AI pro, but it was still being tweaked and wasn’t running smoothly, so I switched back to spot trading.
That statement sits at the center of the entire episode. It addresses the core accusation without providing a line-by-line transaction history. For some readers it was enough. For others it left too many open questions about exact amounts, exact timestamps, and whether her account faced the same temporary restrictions as everyone else’s.
How Binance AI Pro Actually Works
Understanding the product itself helps separate internal transfers from external withdrawals. Binance launched AI Pro on October 19, ten days before the security incident. The tool lets users describe trading ideas in everyday language, test those ideas with simulated funds, and then approve them for live use with real capital. Once a strategy goes live it runs inside a dedicated virtual sub-account that stays isolated from the main trading account.
Users must manually move funds into that dedicated strategy account. They can also move the funds back out again. Existing sub-accounts cannot simply be connected; the system creates a fresh strategy account for each deployed AI strategy. According to the published rules, only transfers between a user’s own accounts are supported. The AI agent itself has no permission to send assets to external wallets or addresses outside the exchange.
That architectural choice is important. Moving capital from a spot wallet into an AI Pro strategy account is an internal bookkeeping event. It does not appear on any public blockchain as an outgoing transfer. From the outside it can look like money disappeared if someone is only watching the main account balance. Inside the system the funds remain under the same user identity and the same custody umbrella.
A similar set of controls exists on another AI trading product the exchange introduced earlier in the year. In both cases the design principle is the same: the automated agent can trade inside a controlled environment but cannot initiate withdrawals. That distinction matters when people claim that leadership “withdrew” funds during a restricted period.
Internal Transfers Versus External Withdrawals
In crypto the difference between an internal transfer and an external withdrawal is not just technical. It is also psychological. External withdrawals remove assets from the exchange’s control and place them on-chain or on another platform. Internal transfers keep everything inside the same custody environment. During a security incident the latter is far less alarming than the former, even if both actions reduce the visible balance in a main account.
He Yi’s explanation relies on exactly that distinction. She described moving funds into a sub-account for testing and later returning them to spot trading. Nothing in the public statements confirms or denies the precise dollar amounts involved in those moves. Nothing independently verifies the exact calendar days either. What the product documentation does confirm is that the mechanism she described is real and that it does not require or permit external transfers.
I’ve found that this kind of nuance often gets flattened in social media discussions. A balance drops, someone screenshots it, and the narrative becomes “she pulled her money out.” Reintroducing the technical details rarely travels as far or as fast as the original accusation. That is simply how attention works online.
The Broader Context of the Security Recovery
While the debate about one account continued, Binance was still working through the practical consequences of the breach. Restoring withdrawals is not a single switch that flips for every asset at once. Different networks have different technical requirements, different hot-wallet configurations, and different risk profiles. Some chains returned relatively quickly. Others required more careful checks.
By November 14 the exchange stated it had restored service on 87 networks during the preceding week and was aiming to finish the remaining ones by the end of the following week. A day earlier it confirmed that affected assets had been replenished using corporate resources. The November 12 proof-of-reserves update covering 198 tokens was presented as evidence that customer balances were whole again.
These steps matter because they form the backdrop against which individual leadership actions are judged. When users are still waiting for their preferred network to reopen, any visible movement of large sums by executives will be scrutinized more heavily than it would be on an ordinary trading day. That scrutiny is understandable. It is also why clear, detailed communication becomes more valuable during recovery periods.
What the Public Record Does and Does Not Show
Let’s be precise about the evidence that is actually available. We have a 30-day performance summary showing a large negative change. We have He Yi’s statement that she tested AI Pro strategies in a sub-account. We have product documentation confirming that AI Pro uses isolated sub-accounts and does not allow external withdrawals by the AI agent. We have confirmation that customer withdrawals were restricted for a period and then gradually restored.
What we do not have is a full transaction log of her accounts. We do not have independent on-chain proof of external transfers timed to the day of the incident. We do not have confirmation that her account was subject to the exact same temporary restrictions that applied to retail users. Those gaps leave space for continued skepticism among some observers.
In my view the absence of external transfer evidence is the more telling fact so far. Accusations of that seriousness usually require at least some blockchain footprint if the assets truly left the platform. None has surfaced publicly. That does not close the case for everyone, but it does shift the burden back onto those making the strongest claims.
How Traders Reacted and Why the Story Spread
Crypto Twitter and similar platforms reward speed and certainty more than careful qualification. A screenshot of a large negative number is easy to share. A nuanced explanation about sub-accounts and strategy isolation is harder to compress into a short post. As a result the initial framing of the story leaned heavily toward suspicion.
Some users focused on the market context. October and early November were not easy weeks for many portfolios. Seeing a senior figure post a sizable drawdown invited the question of whether that drawdown reflected ordinary market risk or something more coordinated. Others focused on the fairness angle: if leadership could still move capital while retail withdrawals were limited, the system felt uneven.
Neither reaction is irrational on its face. Both, however, rest on assumptions that the public data does not fully support. The performance summary does not isolate trading losses from mark-to-market changes or from internal transfers. The timing coincides with the security event, yet coincidence is not the same as causation. These distinctions tend to arrive later in the conversation, after the first wave of strong opinions has already spread.
Lessons About Transparency During Crises
Episodes like this underline a recurring challenge for large exchanges. Leadership accounts are never purely private when the platform is under stress. Even routine internal activity can be read as a signal. That reality places a higher communication burden on executives who choose to share performance data publicly.
At the same time, users also carry a responsibility to distinguish between what the numbers show and what they might imply. A balance change is a fact. An external withdrawal is a different fact that requires different evidence. Conflating the two makes it harder to hold platforms accountable for the things that actually matter—custody integrity, timely restoration of services, and accurate proof of reserves.
Perhaps the most useful outcome of the whole discussion is a clearer public understanding of how AI Pro accounts are structured. Before this episode many traders had only a vague sense that the product used sub-accounts. Now the isolation rules and the restriction on external transfers are more widely known. That knowledge helps future readers interpret similar balance movements more accurately.
What Happens Next for Trust and Communication
Binance continues to work through the final stages of network restoration. Users still monitor the status of individual assets. The proof-of-reserves process remains an important ongoing signal. Against that backdrop, the He Yi episode is likely to fade as newer market stories take center stage. Yet the questions it raised about leadership visibility during restricted periods will probably reappear the next time an exchange faces operational stress.
For now the available public record supports the internal-transfer explanation more strongly than the external-withdrawal accusation. That does not require anyone to accept every detail without further curiosity. It does suggest that the most dramatic version of the story has not been backed by the kind of evidence that usually accompanies confirmed asset exits.
I’ve watched enough of these cycles to know that trust is rebuilt more by consistent operational performance than by any single clarification. Restoring every network, maintaining transparent reserves data, and keeping customer support responsive will matter more over the coming weeks than any additional statement about one account’s 30-day summary. Still, the clarification itself was necessary. Leaving the screenshot unexplained would have allowed the stronger accusations to harden into accepted narrative.
Key Takeaways From the Episode
- He Yi attributed the visible balance changes to testing AI Pro strategies inside dedicated sub-accounts.
- AI Pro creates isolated strategy accounts and does not allow the automated agent to send funds to external wallets.
- The $101 million figure reflects a 30-day change in account value, not necessarily realized losses or external withdrawals.
- Customer withdrawals were restricted after the security incident and restored network by network over subsequent weeks.
- No public transaction records confirming external transfers timed to the incident day have been presented.
The difference between an internal sub-account transfer and an external withdrawal is not a minor technicality. It is the difference between money that stays inside the exchange’s custody system and money that leaves it. In the days after a major security event that distinction carries extra weight. He Yi’s explanation places her activity firmly on the internal side of the line. Until stronger contrary evidence appears, that remains the most supported reading of the available facts.
Crypto markets move quickly and so does the conversation around them. A single screenshot can generate days of discussion. The useful response is not to dismiss the questions but to answer them with the precision the numbers and the product rules actually allow. In this case that precision points toward strategy testing rather than an emergency exit. The recovery work on the broader platform continues, and that work will ultimately shape user confidence more than any one executive’s 30-day summary ever could.
For traders still sorting through the noise, the practical lesson is straightforward. Always separate the visible balance change from the unproven story about where the money went. Demand the same standard of evidence for accusations that you would demand for any other claim about on-chain activity. And keep an eye on the operational metrics that actually determine whether an exchange is returning to normal—network restorations, reserve attestations, and the gradual return of full withdrawal functionality. Those are the signals that matter most once the initial wave of speculation has passed.
The episode also highlights how new products can create temporary confusion. AI Pro is still relatively young. Many users had not yet examined its account structure in detail. When a high-profile balance movement coincides with a platform-wide incident, that lack of familiarity becomes an open invitation for worst-case interpretations. Clearer product education and more frequent explanations of how internal tools work could reduce the intensity of similar episodes in the future.
None of this removes the legitimate frustration felt by users who faced delayed access to their own funds. That frustration is real and deserves acknowledgment. At the same time, treating every leadership balance change as evidence of preferential treatment risks turning ordinary operational activity into constant controversy. The healthier middle ground is continued scrutiny paired with careful reading of the actual mechanisms involved.
As the final networks come back online and the proof-of-reserves process settles into its regular cadence, the market will move on. The questions raised about transparency during restricted periods will remain useful. They remind both platforms and users that trust is easier to maintain when the difference between internal bookkeeping and external movement is explained early and often. In that sense the discussion around He Yi’s account summary, however heated at the outset, has at least produced a clearer public map of how one important new trading product actually functions under the surface.
Looking ahead, exchanges that offer automated or AI-assisted strategies will likely face similar moments whenever market stress and product novelty collide. The best preparation is simple: document the custody boundaries clearly, repeat them when questions arise, and avoid letting a single performance screenshot become the only public data point available. He Yi’s response followed that path. Whether it fully satisfies every critic is secondary to the fact that it re-centered the conversation on verifiable product rules rather than unverified exit narratives.
For anyone still watching the numbers, the takeaway is steady rather than dramatic. Large account swings can reflect trading results, mark-to-market moves, or internal reallocations. Only the last of those three was claimed in this case, and the product architecture supports the claim. Until independent evidence of external transfers appears, the internal explanation remains the one consistent with the information currently on the table. That is usually the quieter conclusion after the louder accusations have had their moment.