Binance’s artificial intelligence systems blocked over $10.53 billion in fraudulent transactions and flagged 36,000 malicious wallet addresses since the start of 2025 — marking a significant leap in how centralized exchanges defend user funds.
AI Now Runs the Security Floor at Binance
Binance revealed that more than half of its fraud prevention controls are now driven by machine learning models, a shift that the exchange says has materially improved its ability to catch threats in real time. The $10.53 billion figure covers losses that were intercepted before reaching bad actors — including phishing schemes, address poisoning attacks, withdrawal fraud, and account takeover attempts.
The 36,000 blacklisted addresses represent a growing database of known malicious actors. Once flagged, these addresses are blocked across Binance’s transaction layer, preventing users from sending funds to them even accidentally. The exchange’s risk engine processes billions of signals daily, cross-referencing on-chain behavior, login patterns, device fingerprints, and transaction velocity to score each interaction for risk.
This is not a minor operational update. It reflects a strategic bet by the world’s largest exchange by volume: that AI-driven compliance infrastructure is now table stakes, not a differentiator.
Why the Numbers Are Larger Than They Appear
The $10.53 billion headline is striking, but the composition of that figure matters. A significant share likely stems from address poisoning — a fast-growing attack vector where fraudsters generate wallet addresses that visually resemble a target’s legitimate address, hoping users copy-paste the wrong destination. Binance has previously disclosed that this category alone accounts for hundreds of millions in near-misses annually.
The 36,000 blacklisted addresses also hint at the scale of organized fraud networks operating against crypto users. Many of these are not isolated individual scammers but coordinated groups running pig butchering schemes, fake yield platforms, and impersonation operations. The fact that Binance’s systems identified and catalogued this many unique malicious actors in roughly 16 months signals that threat volumes are accelerating, not stabilizing.
For context, Chainalysis estimated that crypto scam revenues globally exceeded $9.9 billion in 2024. Binance’s single-platform interception figure approaching that annual industry total underscores just how concentrated attack surface the largest exchanges represent.
What This Means for Traders
For active traders on Binance, the practical implications are tangible:
Withdrawal screening is more aggressive. If you’re sending to a new address, expect friction — additional confirmation steps, brief delays, or prompts to verify the destination. This is the AI risk engine doing its job. It can feel inconvenient, but it is the same system that prevented $10.53 billion in losses.
Address whitelisting matters more than ever. With address poisoning on the rise, traders who pre-approve withdrawal addresses in Binance’s whitelist feature significantly reduce their exposure. A whitelisted address bypasses the poisoning vector entirely.
False positives are a real cost. More aggressive AI screening means legitimate transactions occasionally get flagged or delayed. Traders moving large sums or using novel wallet addresses for the first time may encounter holds. Having identity verification fully completed and account history in good standing helps reduce friction.
The broader signal for the market: exchanges that cannot demonstrate comparable fraud prevention infrastructure are operating at a competitive disadvantage. Institutional clients and high-net-worth retail users increasingly view security metrics — not just fee rates — as a primary selection criterion. Binance’s disclosure is partly a security report and partly a market positioning move.
For traders evaluating which exchanges to use, security track record and fraud prevention capability are now worth weighing alongside trading fees and liquidity depth. Cex101 tracks exchange security ratings, fee structures, and access information across major platforms to help users make informed comparisons.
