SKN | Bybit Says AI Helped Prevent $700 Million in Potential Crypto Losses

• Bybit says its AI-assisted security systems prevented more than 30,000 suspicious withdrawal requests between Jan. 1 and June 15, protecting nearly 20,000 users from approximately $700 million in potential losses.
•  The exchange also reported that AI-assisted auditing identified high-severity vulnerabilities at up to five times the rate of manual reviews and reduced the time required to move from system assessment to testing from roughly two weeks to two hours.
• The figures highlight the growing role of AI in crypto security as exchanges, developers and attackers increasingly use automated tools to identify vulnerabilities and respond to threats.

Bybit Turns to AI After $1.46 Billion Hack

Bybit is using artificial intelligence to strengthen its security infrastructure after suffering the largest theft in cryptocurrency history.

In February 2025, the exchange lost approximately $1.46 billion in a hack attributed to North Korea’s Lazarus Group. Bybit is now pursuing legal action against North Korea and the group over the incident.

A year later, the exchange says AI-assisted security systems are helping it identify vulnerabilities and stop suspicious activity before funds can leave user accounts.

Between Jan. 1 and June 15, Bybit said its AI systems stopped more than 30,000 suspicious withdrawal requests, protecting nearly 20,000 users from approximately $700 million in potential losses.

The exchange stressed that these figures represent withdrawals it prevented rather than confirmed thefts that would otherwise have occurred.

AI Accelerates Vulnerability Detection

Bybit said its AI-assisted auditing system identified high-severity vulnerabilities at up to five times the rate of manual reviews.

The technology also dramatically shortened the time required to investigate potential weaknesses.

Previously, the exchange said it could take approximately two weeks to move from assessing a system to testing it. With AI assistance, that process has been reduced to around two hours.

The automated red-team system scanned 1,489 public-facing assets during the period and identified more than 100 high-severity vulnerabilities.

Bybit also said more than 100,000 security alerts were processed with AI assistance.

The exchange reported that the average time required for the first review of a flagged withdrawal was approximately 4.7 minutes.

AI Flags Fraudulent Funds and Addresses

The exchange’s AI systems were also used beyond vulnerability scanning.

Bybit said the technology identified approximately $212 million in funds linked to suspected fraud and blacklisted more than 10,000 addresses.

The company acknowledged that these figures cannot be independently verified.

Still, the scale of the reported activity illustrates how exchanges are increasingly applying AI to risk monitoring alongside traditional cybersecurity processes.

Rather than relying solely on manual reviews, automated systems can continuously analyze transactions, addresses and system behavior for patterns that could indicate malicious activity.

Crypto Industry Faces an AI Security Arms Race

Bybit’s findings come as AI-assisted cybersecurity has become a growing focus across the cryptocurrency industry.

Bitcoin developers and security researchers have increasingly used AI tools to examine open-source code and identify vulnerabilities before attackers can exploit them.

BTCPay Server, which recently suffered an attack that drained merchant Lightning nodes, has also warned that AI is changing the balance between attackers and defenders.

AI models can analyze large codebases and identify potential weaknesses much faster than traditional manual processes.

The concern is that attackers can use the same capabilities, potentially allowing sophisticated groups to discover and exploit vulnerabilities before development teams have enough time to respond.

Bitcoin Red Team Expands AI-Assisted Audits

A volunteer initiative known as the Bitcoin Red Team has also spent the month using AI models to analyze Bitcoin-related codebases.

The group has reported thousands of potential findings across hundreds of projects, combining automated analysis with human review.

The initiative has relied on donated computing resources and sponsored accounts, demonstrating that AI-assisted security research is no longer limited to major companies with substantial cybersecurity budgets.

Bybit’s approach differs in that the exchange developed its own security tooling and has now published quantitative results from its deployment.

The contrast provides an early look at how AI-based security systems can operate at institutional scale.

Crypto Firms Ask AI Labs for Better Defensive Tools

The growing use of AI for cybersecurity has also prompted crypto companies to seek broader access to advanced models.

Dozens of crypto-related companies, including Coinbase and Block, recently signed an open letter asking AI laboratories for early access to their most capable models.

The companies argued that defenders should have access to powerful AI systems comparable to those potentially available to attackers.

The concern is particularly significant for cryptocurrency businesses because blockchain infrastructure is largely open source and digital assets can often be transferred rapidly once a vulnerability is exploited.

That combination can reduce the amount of time available for human teams to detect and contain an attack.

Human Judgment Remains Central

Bybit said the accelerating pace of cyber threats means security teams are entering an era in which attacks and defenses increasingly operate within minutes rather than days.

David Zong, Bybit’s head of group risk control and security, said the exchange considers strengthening security and risk-control capabilities with AI a top priority.

However, the company also emphasized that human judgment remains central to critical security decisions.

That balance could become increasingly important as automated systems take over more routine detection and analysis while security professionals remain responsible for determining whether alerts represent genuine threats and deciding how to respond.

Closing Insights

Bybit’s reported results provide one of the clearest examples yet of how AI could reshape cybersecurity across the cryptocurrency industry. The exchange says its systems blocked more than 30,000 suspicious withdrawals and protected nearly 20,000 users from $700 million in potential losses, while also dramatically accelerating vulnerability detection and testing. The figures remain company-reported and cannot be independently verified, but they underscore a broader shift in which both attackers and defenders are increasingly using AI to operate at machine speed. As crypto firms continue adopting automated security tools, the competitive advantage may increasingly depend on which side can identify vulnerabilities, suspicious transactions and emerging threats first.