Key Takeaways
- • Chainalysis estimates crypto scam losses reached at least $14 billion in 2025, with potential final figures climbing to $17 billion as investigators identify additional illicit addresses tied to fraud networks.
- • Average payment amounts sent to scam addresses surged 253% year-over-year, signaling both increased fraud sophistication and higher-value victim targets.
- • AI-enhanced scams proved 4.5 times more profitable than non-AI schemes, indicating artificial intelligence has become a critical leverage point for fraud operators scaling operations.
- • Approval phishing—a technique that tricks users into authorizing malicious smart contract permissions—remains the dominant scam vector, frequently embedded within broader investment fraud operations involving social engineering and fake advisors.
- • Investigators identified recurring warning signs including scripted explanations from fraudsters, sudden large crypto purchases by inexperienced users, and high-pressure guidance from purported mentors demanding immediate transaction execution.
Summary
Blockchain analytics firm Chainalysis released findings on June 17 indicating that cryptocurrency scam losses in 2025 reached staggering proportions, with documented on-chain fraud generating at least $14 billion and projections extending to $17 billion as additional scam-linked wallets are attributed to active fraud networks. The upward revision reflects ongoing investigative work to map illicit addresses across blockchain networks, suggesting current loss figures remain incomplete.
The financial impact of crypto fraud has intensified substantially on a per-transaction basis. Chainalysis documented a 253% year-over-year increase in the average payment routed to scam addresses, indicating that fraudsters have successfully targeted higher-net-worth victims and convinced them to move substantially larger sums. This metric reveals not merely an increase in scam frequency, but a troubling escalation in victim vulnerability and exploitation sophistication.
Artificial intelligence has emerged as a material force multiplier in fraud operations. Schemes incorporating AI-enhancement proved 4.5 times more profitable than traditional scams, underscoring how machine learning and generative AI technologies—particularly deepfake video and audio—enable fraudsters to create convincing impersonations of legitimate advisors and establish credibility at scale. This development suggests that standard fraud awareness training may now be insufficient for institutional investors and family office participants managing direct crypto exposure.
Approval phishing constitutes the primary technical mechanism enabling asset theft. The tactic exploits users' misunderstanding of blockchain smart contract permissions by convincing victims they are authorizing routine transactions (token swaps, transfers) when in fact they are granting attackers permanent access to wallet holdings. Once approved, scammers can extract funds at will without requiring additional victim interaction. Chainalysis identified that approval phishing typically functions as part of coordinated investment scams incorporating social engineering, fake financial advisors, and systematic migration of victims from regulated exchanges into self-custody wallets—environments where fund recovery becomes virtually impossible.
Red flags documented by investigators include fraudsters providing scripted transaction explanations, sudden large cryptocurrency purchases by individuals with no prior digital asset history, and high-pressure "mentorship" from supposed advisors demanding rapid action. These behavioral markers offer actionable detection points for institutional compliance teams and custodians managing client assets.
The financial impact of crypto fraud has intensified substantially on a per-transaction basis. Chainalysis documented a 253% year-over-year increase in the average payment routed to scam addresses, indicating that fraudsters have successfully targeted higher-net-worth victims and convinced them to move substantially larger sums. This metric reveals not merely an increase in scam frequency, but a troubling escalation in victim vulnerability and exploitation sophistication.
Artificial intelligence has emerged as a material force multiplier in fraud operations. Schemes incorporating AI-enhancement proved 4.5 times more profitable than traditional scams, underscoring how machine learning and generative AI technologies—particularly deepfake video and audio—enable fraudsters to create convincing impersonations of legitimate advisors and establish credibility at scale. This development suggests that standard fraud awareness training may now be insufficient for institutional investors and family office participants managing direct crypto exposure.
Approval phishing constitutes the primary technical mechanism enabling asset theft. The tactic exploits users' misunderstanding of blockchain smart contract permissions by convincing victims they are authorizing routine transactions (token swaps, transfers) when in fact they are granting attackers permanent access to wallet holdings. Once approved, scammers can extract funds at will without requiring additional victim interaction. Chainalysis identified that approval phishing typically functions as part of coordinated investment scams incorporating social engineering, fake financial advisors, and systematic migration of victims from regulated exchanges into self-custody wallets—environments where fund recovery becomes virtually impossible.
Red flags documented by investigators include fraudsters providing scripted transaction explanations, sudden large cryptocurrency purchases by individuals with no prior digital asset history, and high-pressure "mentorship" from supposed advisors demanding rapid action. These behavioral markers offer actionable detection points for institutional compliance teams and custodians managing client assets.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.