TL;DR
Americans reported $11.366B in crypto losses in 2025 (FBI IC3), and Chainalysis estimates ~$17B stolen on-chain globally. Crypto investment fraud alone was $7.228B. AI-enabled scams earn 4.5x more per operation. Personal wallet compromises hit 158,000 incidents in 2025. The pattern: infrastructure attacks (keys, frontends) now dominate, not clever math exploits.
| SIGNAL | NUMBER | SOURCE |
|---|---|---|
| Crypto losses (US reported) | $11.366B · 181,565 complaints · +22% YoY | FBI IC3 2025 |
| On-chain scam losses (global est.) | ~$17B (from $14B confirmed) | Chainalysis 2026 |
| Crypto investment fraud | $7.228B · 61,559 complaints | FBI IC3 2025 |
| Avg scam payment | $2,764 (was $782 in 2024, +253%) | Chainalysis |
| AI-enabled scam premium | 4.5× ($3.2M vs $719K per op) | Chainalysis / JPMorgan |
| Personal wallet compromises | 158,000 incidents · 80,000 victims · $713M | Chainalysis 2025 |
| Infrastructure share of hacks | 80% of $2.1B H1 2025 losses | TRM Labs |
| Phishing (H1 2025) | $410M · 132 incidents | CertiK H1 2025 |
| Deepfake fraud recorded | $3.7B · 2025–26 = 89% of total | Surfshark / AI Inc. DB |
| Crypto ATM fraud | $389M · +58% YoY | FBI IC3 2025 |
Read those rows as a system, not a list. Scams are no longer a teenager with a fake website. They are industrialized: phishing-as-a-service kits cost under $500, operate from forced-labor compounds in Southeast Asia, and pivot money through professional laundering networks within minutes. In 85% of bridge hacks, funds moved before the incident was even disclosed. The attacker does not need to be smarter than you — they need to be faster than disclosure.
Two asymmetries define 2025–2026. First, AI is the multiplier: LLMs let one operator run fluent multi-language conversations, deepfake face-swaps handle “video verification,” and bots orchestrate fund movement. TRM finds AI-enabled scam activity up ~500% in a year. Second, under-reporting is massive: some categories are under-counted by up to 85%. Every number in this module is a floor, not a ceiling.
Who bears the loss is skewed. Americans 60+ lost $4.432B in crypto fraud alone in 2025 — more than double the 50s cohort — and FBI Operation Level Up found 78% of victims did not know they were being scammed. This is not a “don’t be dumb” problem. It is a persuasion-engineering problem.
We are moving toward a future in which virtually all scams will incorporate AI into their operations to some degree.
CHAINALYSIS 2026 · CRYPTO SCAMS
WHAT ATTACKERS OPTIMIZE FOR
Believability × scale. AI gives both: fluent in every language, live on every platform.
Speed of laundering. 33-minute full cycles exist. Freezing after 24h is usually too late.
Infrastructure reuse. Same drainer wallets, same approval tricks, same cash-out routes.
WHAT DEFENDERS MUST OPTIMIZE FOR
Verification rituals that work under urgency. “Pause 10 minutes” beats “learn blockchain.”
Revoke and simulate flows that take 60 seconds, not 60 minutes.
Clustering: one report exposes a network. Report early — you protect the next victim.
HOW TO READ EVERY NUMBER IN THIS SCHOOL
FBI = reported US losses (floor). Chainalysis/TRM/CertiK = on-chain attributed losses (closer to reality, still incomplete). When sources disagree, trust the larger, on-chain-attributed figure — and remember true harm is higher. The settlement for this lesson: assume every number you see is an undercount, and every technique you see is already being automated.
DEFENSE CHECKLIST — MODULE 01
[01] Assume AI impersonation is default. Verify through a second channel, every time value moves.
[02] Treat urgency as the attack signal. Scammers need you fast; safety needs you slow.
[03] Learn the 5-minute revoke + monitor routine before you need it (Module 08).
[04] Keep take-profit discipline: an un-settled trade is an open exposure window.
[05] Report drainer addresses fast — reuse means one report protects many wallets.
SCHOOL INDEX
← START
You are at the beginning. Read in order — or jump to Module 08 checklists.
NEXT · MODULE 02 →
Keys Are Everything
Private keys, seed phrases, hot vs cold, hardware wallets, multisig, and session keys — the one concept that explains 80% of crypto losses.
PNTHR SCHOOL · FREE + OPEN · SOURCES: FBI IC3 2025 · CHAINALYSIS 2026 · TRM LABS · CERTIK · IMMUNEFI · TUTELA · EDUCATION ONLY · NOT FINANCIAL ADVICE · NOT LEGAL ADVICE