First-party fraud: three abuse patterns every business should watch

Between November 2025 and February 2026, Stripe observed a sharp rise in abusive free trials across its network. The increase is part of a broader trend toward first-party fraud, where legitimate customers exploit policies by creating multiple accounts, cycling through free trials, or gaming refund procedures.

The problem is widespread. According to the Merchant Risk Council, 62% of merchants report increased disputes tied to first-party fraud in the past year, and each $100 in disputed transactions costs businesses $35 to manage.

To better understand the landscape, Stripe analyzed hundreds of millions of transactions. That analysis points to three fast-growing abuse types, each occurring at a different stage of the customer lifecycle: account abuse at sign-up, free trial abuse during evaluation, and refund fraud after delivery.

AI companies bear the brunt of multi-account abuse

Account abuse, also known as multi-account abuse, happens when one person creates several accounts to repeatedly claim promotions, free trials, or perks tied to new registrations. It can also be used to spread stolen card usage across multiple identities and avoid detection longer.

Research from 451 Research’s Voice of the Connected User Landscape: Connected Customer, Trust and Privacy 2025 found that one in five consumers admit to using multiple email addresses or contact details to access discounts more than once. That figure rises to 29% among Gen Z and 27% among millennials.

The pattern often forms a connected web, with a single payment method linked to dozens or hundreds of email addresses, IP addresses, and names. Any business offering something valuable at sign-up can be affected, but AI companies are especially exposed.

Stripe’s analysis shows that 7.4% of customer sign-ups at AI companies are involved in suspected multi-account abuse. Because AI tools rely on compute resources, abuse is directly costly: a user spinning up five accounts instead of one consumes five times the compute. To tackle this, Stripe is introducing a new Radar feature that evaluates potential abuse at registration and login events, helping businesses distinguish real customers likely to convert from repeat abusers. Early access sign-up is available for those interested.

Free trial abuse expands as traditional blockers fail

Free trials are a fixture of product evaluation, but users can violate trial terms by cycling through multiple attempts to get extended free access. Stripe identifies two forces driving the acceleration:

  • AI companies are prime targets. Free trial abuse is not new, but AI startups are contributing heavily to the current increase. These businesses run on expensive compute, rely on trials for acquisition, and are highly vulnerable. AI startups offering free trials with self-serve sign-ups and direct API access see 10x more attempted abuse than enterprise AI solutions.
  • Older prevention methods are wearing out. Many businesses once blocked virtual cards at sign-up to deter repeat trials. Today, many virtual cards are used legitimately for privacy and security, and blanket-blocking them drives away real customers and hurts conversion. AI companies are caught between needing trials for growth and facing exposure that traditional fraud controls can't address.

Radar has launched a solution that predicts common free trial violations with 90% accuracy. A new analytics page shows all high-risk blocked payments, and for businesses that haven't enabled the control yet, shows which payments would have been blocked if it were on. Businesses can email [email protected] for early access.

Refund abuse costs $100 billion annually

Generous return policies are a competitive necessity for many retailers, but they invite abuse. Customers may falsely claim items were never delivered or arrived defective, then keep the product and the refund. Stripe estimates global losses from refund abuse around $100 billion per year.

A common pattern is "wardrobing," where customers buy expensive clothing, use it, and return it. The National Retail Federation’s 2025 Retail Returns Landscape report found that 27% of shoppers who returned at least one online purchase in a 12-month period admitted to wardrobing. Among Gen Z, that jumps to 49%.

Retailers point to social media as a growing amplifier. Influencers may buy large hauls for content and return the items after filming, leaving merchants to absorb two-way shipping, processing fees, and markdowns on goods that may no longer be sold as new.

Some operations are more systematic. Stripe has seen bad actors generate more than 100 email variations and use multiple payment cards to exploit "no questions asked" refund policies. When retailers impose limits like repeat-return fees, abusers open fresh accounts with new cards to circumvent them.

In these cases, the customer holds valid payment details and makes genuine purchases, which makes detection at the point of sale difficult. The fraud becomes visible only later, at refund time. Stripe is building tools to address this and is inviting businesses to join a preview by contacting [email protected].

What Stripe is building next

Stripe’s aim is to give businesses tools to identify, reduce, and monitor first-party fraud and abuse. With billions of transactions flowing across millions of businesses, it has visibility into repeat abusers, networks of fake accounts, and emerging tactics. The company plans to extend its existing AI infrastructure and Radar capabilities to cover first-party fraud alongside traditional fraud vectors.

More about Radar’s evolving abuse protection will be shared at Stripe Sessions.