Bybit Says AI Security Tools Blocked $700M in Potential Crypto Losses

Crypto exchange Bybit says its growing use of artificial intelligence in cybersecurity has dramatically accelerated vulnerability detection and helped prevent more than $700 million in potential losses.

Between Jan. 1 and June 15, Bybit’s AI-powered security system blocked more than 30,000 suspicious crypto withdrawal requests affecting nearly 20,000 users. According to the exchange, flagged withdrawals received an initial review within an average of 4.7 minutes.

The company also said its AI tools identified roughly $212 million in funds potentially connected to fraud and helped blacklist more than 10,000 cryptocurrency addresses. These figures represent Bybit’s own estimates and have not been independently verified.

Bybit has intensified its crypto security efforts following the February 2025 hack that resulted in approximately $1.46 billion being stolen. The attack, attributed to North Korea’s Lazarus Group, became the largest crypto theft on record. Bybit is pursuing legal action against North Korea and the hacking organization.

AI has also transformed Bybit’s vulnerability testing. The exchange said AI-assisted auditing detected high-severity security flaws at up to five times the rate of manual reviews. The time between system assessment and testing reportedly dropped from around two weeks to roughly two hours.

Its automated red-team platform scanned 1,489 publicly accessible assets and uncovered more than 100 high-severity vulnerabilities. Bybit also processed over 100,000 cybersecurity alerts with AI assistance during the reporting period.

The findings come as AI cybersecurity tools gain traction across the Bitcoin and cryptocurrency industry. Following recent infrastructure exploits, developers and companies are increasingly using advanced AI models to identify vulnerabilities before malicious hackers can exploit them.

Crypto companies including Coinbase and Block recently joined an open letter calling for AI developers to give security researchers earlier access to powerful models. Meanwhile, the volunteer Bitcoin Red Team has been using AI to analyze hundreds of Bitcoin-related projects for security weaknesses.

Bybit security executive David Zong said the cybersecurity arms race has shifted to an era measured in minutes, while emphasizing that human judgment remains central to critical security decisions.