The State of Ecommerce Fraud in 2026: The Numbers Every Trust & Safety Team Should Know

Ecommerce fraud stopped being a flat cost of doing business a while ago. It compounds, and the curve got steeper in 2025. Here are the numbers that define where things stand in 2026, and what they actually mean for the people trying to keep platforms safe.

Losses are compounding, not holding steady

Global ecommerce fraud losses reached roughly $48 billion in 2025. Juniper Research projects merchant losses to online payment fraud will climb from $44.3 billion in 2024 to $107 billion by 2029, a 141% jump in five years.

The damage is bigger than the raw loss suggests. Merchants don’t lose one dollar per dollar of fraud; they lose about $4.61 once you add fees, labor, and the cost of the goods that walked out the door. Across a full year, fraud eats roughly 3% of ecommerce revenue on average, and more in some regions: close to 4.6% in Latin America and 3.1% in Europe.

The defense market is racing to keep up

The fraud detection and prevention market is forecast to grow from around $67 billion in 2026 to $244 billion by 2034. When defensive spending grows that fast, it usually means the threat is growing faster.

AI changed the slope of the curve

This is the part that worries trust & safety leaders most, and the data backs the worry.

AI-facilitated fraud losses are projected to rise from $12.3 billion in 2023 to $40 billion by 2027.
Deepfakes made up about 11% of global fraud attempts in 2026, after surging more than 1,000% in two years, according to Sumsub.
Synthetic identity fraud, where real and fabricated data get stitched into an identity that passes KYC, rose 311% between early 2024 and early 2025.
Experian forecast that AI-powered scams would accelerate sharply through 2026, after consumers lost $12.5 billion to fraud the year before.

The cost of running a convincing scam dropped. So volume went up and quality went up at the same time, which is a nasty combination to defend against.

Friendly fraud is the quiet majority

Not all fraud comes from outside. A customer disputing a purchase they actually made (friendly fraud, or first-party fraud) accounts for somewhere between 40% and 80% of total fraud losses depending on the study, and drives about 61% of chargeback disputes. Most chargebacks aren’t stolen cards. They’re buyers charging back real orders. Chargeback volume is on track to rise from 238 million in 2023 to 337 million in 2026.

Account takeover keeps climbing

Account takeover losses are estimated to have grown from $13 billion in 2024 to $17 billion in 2025, and roughly 61% of ATO attacks target ecommerce. Credential-stuffing volume against login endpoints grew 148% year over year through late 2025 per Sift, and TransUnion measured a 37% year-over-year rise in suspected account takeover in early 2026.

What the numbers actually say

Read past the headlines and three things stand out.

Fraud is shifting away from stolen cards toward stolen and synthetic identities and toward dispute abuse. The old card checks are still necessary, but on their own they now miss most of the problem.

The most expensive category for a lot of merchants isn’t external fraud at all. It’s false declines and friendly fraud: real revenue lost to blunt rules and disputed legitimate orders.

And the teams holding the line in 2026 are the ones that can tell intent apart from identity. Is this a real customer behaving strangely, or a strange actor wearing a real customer’s details? That’s a content and behavior question, not only a payment one.

If you want the specific patterns behind these numbers, we keep a fraud detection library with the signals for scam listings, fake reviews, phishing, impersonation, and payment fraud.

Sources

Sumsub — Fraud Trends 2026 (deepfakes, synthetic identity)
TransUnion — H1 2026 fraud trends report (account takeover)
Experian / Fortune — 2026 fraud forecast
Merchant Risk Council — Global Payments and Fraud Report
Juniper Research and Sift, for payment-fraud and credential-stuffing figures