Chainalysis: Impersonation and AI caused $17 billion in losses in 2025

By 2025, the landscape of crypto crime has undergone a significant transformation. According to the latest report published by Chainalysis in January 2026, total losses due to fraud have reached $17 billion, but what is even more concerning is the shift in tactics used by malicious actors in the sector.

The analysis firm has documented a disturbing evolution: criminals are dismantling technical defenses and instead focusing on the most vulnerable resource of all, human trust. Impersonation and AI-based scams have now surpassed cyberattacks as the primary method for stealing funds.

The real target? Trust, not technical vulnerabilities

In previous years, the main focus of security in cryptocurrencies was on protecting smart contracts and network infrastructure. Hacks and exploits represented the dominant threat, with documented losses of approximately $2.2 billion in 2024 solely from cyberattacks.

However, 2025 marked a turning point. Chainalysis observed that criminals have gradually shifted their approach, moving from large-scale, untargeted campaigns to more sophisticated and calibrated strategies. Instead of attempting to rob thousands of people with small sums (the “spray-and-pray” tactic), malicious actors now target specific individuals and withdraw significantly larger amounts, with much higher success rates.

How impersonation has become 1,400% more popular among criminals

The numbers provided by Chainalysis paint an alarming picture: impersonation-based frauds have skyrocketed by 1,400% year-over-year. This is not just a marginal increase but a radical change in criminals’ tactical preferences.

The tactic works in a deceptively simple way: scammers create fake profiles pretending to be exchange support staff, government officials, or trusted individuals within the sector. Victims, believing they are interacting with legitimate personnel, are induced to reveal sensitive information, transfer funds, or grant access to their wallets.

A notable case comes from the UK, where an investor lost nearly $2.5 million in 2025 in a scam structured precisely around these dynamics. North Wales Police described this phenomenon as a “worrying trend,” highlighting how scammers skillfully exploit fear and panic through elaborate social engineering schemes.

AI makes scams 4.5 times more profitable

Here emerges the multiplier factor that makes the situation even more critical: when impersonation is combined with artificial intelligence, profit margins become extraordinary. According to the Chainalysis report, AI-enabled scams are 4.5 times more profitable than traditional scams.

Technology is used to create sophisticated deepfakes, generate personalized messages on a massive scale, and automate entire contact and manipulation processes. A scammer can now manage hundreds of conversations simultaneously, tailoring each message to the specific target, all without significant manual intervention. Automated tools produce fake “government alerts,” counterfeit “support agents,” and falsified communications from “trusted insiders,” all with a level of visual and textual authenticity sufficient to bypass most users’ skepticism.

Real cases demonstrating the new face of crypto crime

Historical statistics help frame the severity of the situation. Between 2020 and the end of 2023, nearly 100,000 people in the UK fell victim to investment scams, losing a total of £2.6 billion (approximately $3.5 billion). This amounts to about £13 million stolen weekly during that period, according to the North Wales Cyber Unit report cited in April 2024.

However, these figures only represent officially reported scams. The dark figures—cases not reported due to shame, distrust of authorities, or simple ignorance—are probably at least double. This means the problem is even larger than public statistics suggest.

The phenomenon is not limited to inattentive or naive victims. Even experienced industry professionals have proven vulnerable. Lior Aizik, co-founder and chief operating officer of XBO, revealed that he was personally targeted by impersonation: scammers used his name to create fake profiles to contact industry operators and request money, pretending to represent the exchange.

As Aizik emphasized, these attacks are not based on brilliant technical exploits or critical vulnerabilities. They rely on perceived urgency, relationship building, and psychological manipulation. “If a message appears urgent or confidential,” he warns, “it’s usually a red flag.” Yet many still fall into the trap, even those who theoretically know how to protect themselves.

Why impersonation is harder to combat than cyberattacks

The distinction emerging from Chainalysis data is crucial for understanding the future of security in cryptocurrencies. A cyberattack, no matter how sophisticated, typically creates a vulnerability that can be patched, corrected, or resolved technologically. Exchanges and protocols can strengthen their infrastructure, implement security audits, and close exploited gaps.

Impersonation and AI-based scams, however, operate at a completely different level. They do not target code vulnerabilities but human vulnerabilities. They do not require direct access to systems but only the ability to convince a person—through a credible message—to take an action that puts their funds at risk.

This distinction explains why crypto-related crime is rapidly reconfiguring. It is no longer primarily an industry of hacks and technological exploits. It is increasingly becoming an industry of sophisticated psychological manipulation, where the authenticity of appearance and communication outweighs the importance of technological defenses.

The Chainalysis report ultimately suggests that the cryptocurrency sector faces a more elusive and persistent enemy than traditional cyberattacks. Because even when exchanges do everything correctly from a technical standpoint, impersonation can still prevail. Human trust remains, and will probably continue to be, the most difficult vector to defend against.