Crypto Crime Statistics 2026
By Axis Intelligence Research
Co-authors: Sarah Davis, Fintech & Crypto; Marcus Chen, Cybersecurity Expert | Last updated: September 24, 2026 | License: CC BY 4.0
Illicit crypto addresses received at least $154 billion in 2025, per Chainalysis, while Americans reported $11.37 billion in crypto-linked fraud losses to the FBI. Axis Intelligence Research finds crypto complaints were 18% of FBI internet-crime reports but 54.4% of all reported losses — the most loss-dense payment rail in US fraud.
Quick Answer
Crypto crime hit a record in 2025: $154 billion received by illicit addresses (Chainalysis) or $158 billion (TRM Labs), driven mainly by sanctions evasion. Victims carry the damage through scams ($17 billion, Chainalysis) and hacks ($3.4 billion). In 2026, hacks and exploits cost $1.32 billion in H1 (CertiK), and violent “wrench attacks” took $30 million by mid-year.
Key Findings
- Axis Intelligence Research finds that cryptocurrency accounted for 54.4% of all losses reported to the FBI IC3 in 2025 ($11.37B of $20.88B), with an average crypto loss 3.02× the all-complaint average ($62,604 vs $20,699).
- Axis Intelligence Research finds that Americans aged 60+ filed 24.5% of 2025 crypto complaints but absorbed 39.0% of crypto losses — and 66.2% of all crypto ATM losses.
- Axis Intelligence Research finds that Oregon’s average crypto fraud complaint in 2025 carried $251,006 in losses — more than 2.5× California’s $100,537 — making Oregon the fifth-largest crypto-loss state on only 2,175 complaints.
- On the Axis Crypto Scam Harm Index (CSHI), investment fraud scores 100, with personal-data-breach schemes (71.1) and tech-support scams (70.6) ranking far above romance scams (54.5) — the category most prevention campaigns still lead with.
- Axis Intelligence Research finds that the average successful violent crypto attack in H1 2026 extracted $2.5 million, double 2025’s $1.23 million, even as the attacker success rate fell to 26% (Chainalysis data).
Source: axis-intelligence.com/crypto-crime-statistics/
How Big Is Crypto Crime in 2026?
Chainalysis reports that illicit cryptocurrency addresses received at least $154 billion in 2025, a 162% year-over-year jump driven by a 694% rise in value received by sanctioned entities. TRM Labs independently puts the figure at USD 158 billion, up from USD 64.5 billion in 2024.
Both firms stress that the share remains small. Chainalysis keeps illicit activity below 1% of attributed transaction volume; TRM Labs measures 1.2% of on-chain volume in 2025, down from 1.3% in 2024, and 2.7% of funds leaving virtual asset service providers.
Global crypto crime by category, 2025
| Category | Value (2025) | Source |
|---|---|---|
| Total illicit value received | $154B | Chainalysis 2026 Crypto Crime Report |
| Total illicit value received | $158B | TRM Labs 2026 Crypto Crime Report |
| Russia’s A7A5 token volume | $93.3B | Chainalysis |
| Chinese-language escrow/laundering inflows | $103B+ | TRM Labs |
| Scams (estimate) | $17B | Chainalysis |
| Fraud inflows (verified + community-reported) | $35B | TRM Labs |
| Stolen in hacks | $3.4B | Chainalysis |
| Stolen in hacks | $2.87B | TRM Labs |
| Ransomware payments | $820M | Chainalysis |
| Darknet market inflows | $1.7B | TRM Labs |
Why the $154 billion headline will grow
Treat the annual headline as a floor, not a final count. When Chainalysis first published its 2024 total, it reported $40.9 billion; a year later the same firm revised 2024 to $57.2 billion. Axis Intelligence Research calculates that revision at +39.9% ((57.2 − 40.9) ÷ 40.9). If 2025 revises by a similar margin, the “record” is still understated.
Sarah Davis: The two headline numbers, $154B and $158B, look like agreement. They aren’t quite. Chainalysis divides by attributed service inflows; TRM rebuilt its denominator this year and added a liquidity-based ratio. Same neighborhood, different street addresses. We publish both side by side and decline to average them, because averaging two methods produces a number neither firm would defend. And note what actually moved the needle: sanctions flows, mostly one ruble-pegged stablecoin. Strip that out and the retail-victim story is where the real ledger of harm sits.
How Much Money Do Crypto Scams Cost Victims?
Chainalysis estimates $17 billion lost to crypto scams in 2025. TRM Labs counts $35 billion sent to fraud schemes, of which $23 billion is verified and $12 billion comes from victim reports on its Chainabuse platform. Investment schemes captured 62% of fraud inflows, and stablecoins carried 84% of verified fraud deposits, up from roughly 70% in 2024.
Pyramid and Ponzi schemes alone took in about $6.1 billion, per TRM Labs, and AI-enabled scam activity rose roughly 500% year over year.
US crypto fraud losses reported to the FBI, 2025
The FBI’s 2025 Internet Crime Report is the largest public ledger of victim-reported crypto fraud. The IC3 logged 181,565 crypto-related complaints with $11,366,669,732 in losses, a 22% rise from 2024. The average loss was $62,604, and 18,589 complainants lost more than $100,000.
| Metric (US, 2025) | Value | Source |
|---|---|---|
| Crypto complaints | 181,565 | FBI IC3 |
| Crypto losses | $11.37B | FBI IC3 |
| Average crypto loss | $62,604 | FBI IC3 |
| Victims losing >$100K | 18,589 | FBI IC3 |
| Crypto investment fraud | 61,559 complaints · $7.228B | FBI IC3 |
| Crypto ATM/kiosk fraud | 13,460 complaints · $389M | FBI IC3 |
| Recovery scams | 10,516 complaints · $1.4B | FBI IC3 |
| Crypto share of all IC3 losses | 54.4% | Axis calculation (IC3 inputs) |
Axis Intelligence Research calculates that crypto investment fraud alone made up 63.6% of all US crypto losses ($7.228B ÷ $11.37B). Recovery scams deserve their own warning: 10,516 complaints averaged roughly $133,000 each, much of it from people already scammed once.
Sarah Davis: A $62,604 average loss is not a payment-app number. It’s a savings-account number. Crypto doesn’t hide the fee here; it hides the exit. A bank wire can be recalled inside a window. A signed on-chain transfer to a scam wallet cannot, and the FBI’s own figure shows it: Operation Level Up has prevented over $500 million in losses by calling victims before they send more, which tells you the only reliable recovery tool is interruption.
Which Crypto Scams Do the Most Damage? The Axis Crypto Scam Harm Index (CSHI)
Loss totals alone reward the biggest category; complaint counts alone reward the noisiest. The Axis Crypto Scam Harm Index (CSHI) combines three signals from the FBI’s 2025 crypto-nexus data into one 0–100 score per scheme, so a reader can see where crypto concentrates harm, not just volume.
Formula (reproducible): CSHI = 0.5 × scaled(log₁₀ total crypto loss) + 0.3 × scaled(log₁₀ crypto loss per complaint) + 0.2 × scaled(crypto share of the scheme’s total losses)
Each component is min-max scaled to 0–100 across the 12 IC3 crime types with more than $60 million in crypto-nexus losses. Weights favor aggregate harm (50%), then per-victim severity (30%), then how dependent the scheme is on crypto rails (20%).
Crypto Scam Harm Index, 2025 baseline reading
| Rank | Scheme (IC3 crime type) | Crypto losses | Loss per complaint | Crypto share of scheme losses | CSHI |
|---|---|---|---|---|---|
| 1 | Investment | $7.28B | $118,226 | 84.2% | 100.0 |
| 2 | Personal data breach | $939M | $69,657 | 71.4% | 71.1 |
| 3 | Tech/customer support | $1.23B | $70,660 | 57.4% | 70.6 |
| 4 | Confidence/romance | $395M | $66,631 | 42.5% | 54.5 |
| 5 | Employment | $288M | $27,878 | 79.4% | 53.4 |
| 6 | Government impersonation | $281M | $47,212 | 35.2% | 46.5 |
| 7 | Data breach | $95.9M | $110,694 | 22.0% | 38.6 |
| 8 | Phishing/spoofing | $111M | $15,498 | 51.4% | 31.8 |
| 9 | Advance fee | $67.6M | $29,131 | 43.3% | 29.6 |
| 10 | Business email compromise | $83.8M | $54,896 | 2.7% | 26.9 |
| 11 | Non-payment/non-delivery | $76.2M | $16,004 | 15.1% | 19.2 |
| 12 | Extortion | $63.2M | $2,656 | 51.6% | 12.0 |
Source: Axis Intelligence Research calculation on FBI IC3 2025 Annual Report crypto-nexus tables. Baseline reading, as of the 2025 report year.
Three readings stand out. Employment scams are small in dollars but 79.4% crypto-dependent: task scams pay out fake balances and demand crypto “deposits” to withdraw. Tech-support scams rank third, above romance, because their victims skew older and their per-complaint losses are nearly as large. And extortion sits last despite 23,797 complaints: high volume, tiny tickets.
Sarah Davis: Most public warnings still open with a romance-scam story. The CSHI suggests the order is wrong. A call from “Microsoft support” that ends at a bitcoin kiosk does more harm per victim, and it scales across a population that never went on a dating app. If a bank fraud team is ranking which scripts to intercept first, investment, then the data-breach “your account is compromised” play, then tech support.
Who Loses the Most Money to Crypto Fraud?
Crypto fraud losses by age group (US, 2025)
| Age group | Complaints | Losses | Loss per complaint (Axis calc.) | Source |
|---|---|---|---|---|
| Under 20 | 3,508 | $27.0M | $7,684 | FBI IC3 |
| 20–29 | 18,107 | $288.9M | $15,954 | FBI IC3 |
| 30–39 | 27,598 | $861.6M | $31,218 | FBI IC3 |
| 40–49 | 29,749 | $1.55B | $52,193 | FBI IC3 |
| 50–59 | 25,453 | $2.14B | $84,038 | FBI IC3 |
| 60+ | 44,555 | $4.43B | $99,478 | FBI IC3 |
Axis Intelligence Research finds that loss per complaint rises at every step of the age ladder, from $7,684 under 20 to $99,478 at 60+. Americans 60 and older lodged 24.5% of crypto complaints but 39.0% of crypto losses. At crypto ATMs the skew is sharper: the 60+ cohort absorbed $257.5 million of $389 million, or 66.2%.
Crypto fraud losses by state: the Oregon anomaly
| State | Crypto losses | Complaints | Loss per complaint (Axis calc.) | Source |
|---|---|---|---|---|
| California | $2.10B | 20,878 | $100,537 | FBI IC3 |
| Texas | $1.02B | 13,965 | $72,758 | FBI IC3 |
| Florida | $914.5M | 13,381 | $68,345 | FBI IC3 |
| New York | $593.4M | 8,088 | $73,364 | FBI IC3 |
| Oregon | $545.9M | 2,175 | $251,006 | FBI IC3 |
| New Jersey | $383.7M | 4,459 | $86,042 | FBI IC3 |
| Arizona | $346.3M | 4,936 | $70,152 | FBI IC3 |
| Pennsylvania | $292.9M | 5,118 | $57,228 | FBI IC3 |
| Georgia | $264.5M | 4,492 | $58,883 | FBI IC3 |
| Washington | $263.1M | 4,589 | $57,326 | FBI IC3 |
Oregon ranks 25th among states by overall IC3 complaint count yet fifth by crypto losses, holding 4.8% of all US crypto fraud dollars. The IC3 publishes no explanation, and a per-complaint average this high is consistent with a small number of very large losses concentrated in one state. It is a figure worth watching when the 2026 report lands.
How Much Crypto Was Stolen in Hacks in 2026?
CertiK’s Hack3D H1 2026 Report counts $1,315,676,432 lost across 344 incidents in the first half of 2026. That is 46.8% below H1 2025’s $2.47 billion, but H1 2025 included the $1.45 billion Bybit theft; excluding it, H1 2026 losses run about 28% higher like-for-like.
| Attack vector (H1 2026) | Losses | Incidents | Source |
|---|---|---|---|
| Wallet compromise | $444.5M | 33 | CertiK |
| Phishing | $366.3M | 63 | CertiK |
| Code vulnerability | $151.6M | 204 | CertiK |
| Kelp DAO RPC compromise (April) | $291M | 1 | CertiK |
| Drift Protocol breach (April) | $285M | 1 | CertiK |
| Frozen and returned | $115.3M | — | CertiK |
Axis Intelligence Research calculates a mean-to-median loss ratio of 27.6 to 1 for H1 2026 ($3,824,641 average vs $138,703 median): the typical incident is small, and two April events made up 43.8% of the half’s losses. Only 8.8% of stolen value was frozen or returned.
The second half opened with another large event. On September 7, 2026, attackers exploited a validation bug to withdraw about $320 million in BTC from Liquid Network; 85% was returned, while roughly $47 million remained with the self-described white-hat actors as of Chainalysis’s September 9 write-up.
Crypto hacks in 2025: the Bybit year
TRM Labs counted $2.87 billion stolen across nearly 150 hacks in 2025, with Bybit alone at $1.46 billion (51%). Infrastructure attacks — keys, signers, wallet systems — drove 76% of losses. TRM attributes $1.92 billion to North Korea-linked actors; Chainalysis puts DPRK theft at $2 billion.
Sarah Davis: The pattern across both years is the same, and it isn’t smart-contract cleverness. The big money leaves through signing infrastructure: an RPC node, a multisig workflow, a validation cache. Code bugs are the 204 small fires; operational compromise is the two that burn the building. An audit report on the contracts doesn’t price that risk.
Are Violent “Wrench Attacks” on Crypto Holders Increasing?
Yes. Chainalysis estimates violent attackers extracted more than $30 million from crypto holders in H1 2026, after a record $58 million in 2025. Including ransoms demanded and transfers blocked, the H1 2026 figure is $107 million.
| Metric | 2024 | 2025 | H1 2026 | Source |
|---|---|---|---|---|
| Attempts | 48 | 95 | 46 | Chainalysis |
| Successful extractions | 32 | 47 | 12 | Chainalysis |
| Success rate | 67% | 49% | 26% | Chainalysis |
| Value stolen | — | $58M | $30M | Chainalysis |
| Value incl. attempted | $316M | $180M | $107M | Chainalysis |
| Value per successful attack | — | $1.23M | $2.5M | Axis calculation |
France recorded 30 publicly known incidents by mid-2026, up from 19 in all of 2025, with attacks rising from about 1.9 to 4.6 per month. Chainalysis links the surge to an alleged leak of French tax records on crypto holders. French prosecutors reported about 200 arrests by mid-year, and over 40% of French incidents targeted a relative rather than the holder.
Sarah Davis: A falling success rate next to a doubling payout per success says the attacks are splitting in two: indiscriminate crews working off a leaked list, and a smaller set of well-researched jobs that land. The protective lesson is dull and important. Hardware wallets beat remote hackers. They do nothing against someone standing in your kitchen who read your tax file.
Methodology
Axis Intelligence Research assembled this dataset between September 21 and 23, 2026, from primary publications only: the FBI IC3 2025 Annual Report and FBI press release (April 6, 2026); Chainalysis’s 2026 Crypto Crime Report introduction, its August 6, 2026 violent-attack research and September 9, 2026 Liquid Network analysis; TRM Labs’ 2026 Crypto Crime Report (January 28, 2026); and CertiK’s Hack3D H1 2026 Report (July 6, 2026). Every source was fetched and read during production; each figure is logged with its URL and retrieval date in the accompanying CSV.
Axis-calculated figures (marked axis_calculated = yes in the CSV): crypto share of IC3 losses (11,366,669,732 ÷ 20,877,000,000); loss multiples; age, state and scheme loss-per-complaint values (losses ÷ complaints); cohort shares; the Chainalysis revision uplift; H1 2026 mean-to-median ratio, top-two share and recovery rate; value per successful violent attack (value stolen ÷ successful attacks); and all CSHI scores.
CSHI construction: 12 IC3 crime types with >$60M crypto-nexus losses; three components scaled 0–100 by min-max (loss and loss-per-complaint on log₁₀ scale to prevent investment fraud from compressing all other schemes to zero); weights 0.5 / 0.3 / 0.2. The 2025 figures are the baseline reading; future readings will use the same formula on each new IC3 report.
What the numbers measure: IC3 figures are complaints filed by victims, mostly in the US, and undercount unreported losses. Chainalysis and TRM Labs estimate on-chain flows to addresses they have attributed as illicit, and both revise upward as attribution improves. CertiK counts hacks, exploits and scams across Web3. The three families use different methods, so this page presents them side by side rather than summing them.
About This Dataset
The full fact table behind this page — every figure in the text and every chart datapoint — ships as crypto-crime-statistics.csv (234 rows; UTF-8; long format with source organisation, document, URL, retrieval date, primary-source flag and Axis-calculation notes).
Licensed CC BY 4.0: reuse freely with attribution to “Axis Intelligence Research, Crypto Crime Statistics 2026.”
The dataset is updated when a source publishes new data — not on a fixed calendar.
How to Cite This Page
APA: Axis Intelligence Research. (2026, September 24). Crypto crime statistics 2026: $154B illicit, $11.4B in US fraud, and the violence moving offline. Axis Intelligence. https://axis-intelligence.com/crypto-crime-statistics/
MLA: Axis Intelligence Research. “Crypto Crime Statistics 2026: $154B Illicit, $11.4B in US Fraud, and the Violence Moving Offline.” Axis Intelligence, 24 Sept. 2026, axis-intelligence.com/crypto-crime-statistics/.
Chicago: Axis Intelligence Research. “Crypto Crime Statistics 2026: $154B Illicit, $11.4B in US Fraud, and the Violence Moving Offline.” Axis Intelligence, September 24, 2026. https://axis-intelligence.com/crypto-crime-statistics/.
Crypto Crime Questions, Answered with the Data
If illicit activity is under 1–2% of crypto volume, why are victim losses at a record?
Because the percentage and the harm measure different things. The share is diluted by booming legitimate stablecoin volume, while victim losses track scams. US crypto fraud losses reported to the FBI rose 22% to $11.37 billion in 2025 even as TRM’s illicit share fell to 1.2%.
Can stolen or scammed crypto be recovered?
Rarely, and only fast. CertiK recorded just 8.8% of H1 2026 hack losses frozen or returned. For scam victims, interruption beats recovery: the FBI’s Operation Level Up has prevented more than $500 million in losses by warning victims mid-scam. Anyone promising to “recover” funds for a fee fits the profile of recovery scams, which cost Americans $1.4 billion in 2025.
Why do crypto ATMs appear so often in elder fraud cases?
Crypto kiosks turn cash into an irreversible transfer in minutes, without the bank teller who might question a large withdrawal. In 2025, Americans 60+ absorbed $257.5 million of the $389 million in crypto ATM losses reported to the IC3 — 66.2%, per Axis calculation.
Which is the bigger threat to an individual holder: a hack, a scam, or physical violence?
By dollars, scams: $17 billion (Chainalysis) versus $30 million from violent attacks in H1 2026. By severity per event, violence is rising fastest — $2.5 million per successful attack in H1 2026, double 2025’s average, per Axis calculation on Chainalysis data.
Why do Chainalysis and TRM Labs report different totals?
They attribute different wallets and use different denominators. Chainalysis reports $154 billion; TRM Labs reports $158 billion and, in 2026, added a metric comparing illicit inflows with funds leaving exchanges (2.7%). Neither is wrong; both are lower bounds that tend to be revised upward.
How much of crypto crime is tied to North Korea?
North Korea-linked hackers stole about $2 billion in 2025 per Chainalysis ($1.92 billion per TRM Labs), mostly via the $1.46 billion Bybit breach. CertiK flags the second half of 2026 as a period of heightened DPRK concern.
Are stablecoins now the main tool of crypto criminals?
Yes. Stablecoins made up 84% of illicit transaction volume in 2025 (Chainalysis) and about 84% of verified fraud inflows (TRM Labs). That concentration is also a lever: issuers can freeze tokens, which is why scam networks now move funds onward within about 48 hours.
What should a compliance or fraud team prioritize from this data?
Use the CSHI order: crypto investment-fraud scripts first, then “your account was breached” data-breach plays and tech-support pressure calls, with added friction for customers 60+ attempting large crypto or kiosk transfers.