Klarna becomes AI distribution layer

Klarna is trying to own the product feed that AI shopping agents read, not just the payment button they click at the end. If merchants plug product catalogs, prices, stock, shipping, and payment options into Klarna once, Klarna can sit upstream of checkout inside chat based shopping flows. That turns Klarna from a lender into routing infrastructure for discovery, comparison, and purchase across agent interfaces.

• The product is a structured catalog API built on Klarna’s existing shopping graph. It exposes 100M plus products and 400M prices across 12 markets, and accepts feeds merchants already use, like Shopify, Google Merchant, Amazon, Facebook Catalog, or CSV and JSON. That lowers merchant setup work and makes Klarna useful before payment starts.
  1 klarna 5 klarna
• This extends Klarna’s older playbook of using BNPL to win merchant distribution, then layering ads, shopping, and merchant services on top. In 2024, ads and merchant services were already helping lift take rate, while Klarna had grown to 790,000 merchants by Q2 2025 and 850,000 by Q3 2025. The protocol gives that merchant base a new traffic source inside AI assistants.
  2 sacra 3 sacra
• The closest analog is one click checkout infrastructure like Bolt, but shifted one step earlier in the funnel. Bolt aimed to save identity and payment details across stores to reduce checkout friction. Klarna is doing that for AI agents at the discovery layer, where the winning company can influence which products are shown before the shopper ever reaches a merchant site.
  4 stripe 6 sacra

If AI assistants become a real shopping surface, the valuable position will move from owning a button on the checkout page to owning the data rails that feed the assistant. Klarna is building toward that position, then connecting it to agent payments through partners like Stripe, which could let it capture more merchant demand and more transaction volume in the same flow.