Stripe Unveils AI-Powered Foundation Model for Enhanced Payments

At Stripe Sessions 2025, fintech leader Stripe announced a new AI foundation model designed to revolutionize payment processing. Trained on billions of transactions, this model identifies subtle signals in each payment to improve accuracy and efficiency.

Enhanced Fraud Detection

One key benefit of the new model is significantly improved fraud detection. Stripe reports a 64% increase in card-testing attack detection for large businesses, a substantial improvement over previous models. This advancement builds on Stripe's existing efforts, which had already reduced such attacks by 80% over two years.

Emily Glassberg Sands, Stripe's head of information, emphasized the model's ability to leverage vast amounts of data for enhanced performance.

Deeper Partnership with Nvidia

Stripe also revealed a strengthened partnership with Nvidia. Nvidia migrated its entire subscriber base to Stripe Billing in just six weeks, a process Stripe claims typically takes months. This marks the fastest migration to Stripe Billing to date.

Stablecoin Integration and Other Key Announcements

Stripe is partnering with startups like Ramp, Squads, and Airtm to introduce stablecoin-backed, multicurrency cards. This will enable businesses worldwide to operate in a unified currency. This move follows Stripe's recent acquisition of stablecoin platform Bridge.

Stripe also announced several other key updates:

  • Support for 25 new payment methods, totaling over 125.
  • Klarna integration with Stripe's consumer payment product, Link.
  • Stripe Terminal compatibility with third-party hardware.
  • Managed Payments, a new merchant-of-record offering for global expansion.
  • Smart Disputes, using AI for automated dispute handling.
  • Expanded Stripe Tax availability to 102 countries.
  • Global Payouts, simplifying payments to customers and contractors.

Stripe's new AI foundation model, along with these additional features, positions the company to further innovate in the payments space.