Chainalysis data summarized in Tom’s Hardware estimates that cryptocurrency theft via scams surged in 2025, with headline figures in the multibillion‑dollar range. Analysts and law‑enforcement sources say impersonation, AI deepfakes and automated tactics materially amplified scam scale and per‑victim losses.

Analytical reporting summarizing Chainalysis data and published in January 2026 finds that cryptocurrency theft via scams rose sharply in 2025, with estimates of at least tens of billions of dollars in stolen crypto linked to impersonation and AI‑assisted tactics. The analysis highlights a marked shift: industrialized social‑engineering campaigns using AI‑generated voice and video deepfakes, automated messaging and bespoke impersonation lowered the cost and increased the reach of scams, enabling larger per‑victim losses. Romance, investment and impersonation schemes were singled out as major conduits for transfers to mixers, exchanges and convertible on‑ramps. Chainalysis and other blockchain‑analytics firms, together with law‑enforcement sources, warned that traditional detection methods are being outpaced by automated, AI‑driven fraud flows and that stronger cross‑platform detection, regulatory measures for centralized intermediaries and improved public education are urgently needed. The reporting underscores the evolving threat landscape and the need for coordinated technical, legal and international responses to stem rapidly growing AI‑enabled crypto crime.