BNPL Is Becoming Infrastructure

Klarna has 118 million active consumers, 966,000 merchants, and a NYSE listing. It also has a visibility problem. At a growing number of checkouts, the installment option carries the merchant's brand, not Klarna's.

The lending is still happening, but the lender is disappearing.

The obvious explanation is that BNPL providers got better at embedding themselves. That's true, but it misses the structural shift. BNPL is being absorbed into the payment processing stack the same way fraud screening and currency conversion were absorbed before it. The product is collapsing into a feature, and the brand into an API.

How BNPL worked as a product

The original model was straightforward. You download the Afterpay or Klarna app, create an account, browse products inside the app, and check out. The BNPL provider owns the consumer relationship, the brand, and the credit decision.

The standard product splits a purchase into four interest-free installments: 25% up front, three more payments every two weeks. The consumer pays nothing extra. The merchant pays a discount rate, typically 2% to 8% of the transaction, to the BNPL provider.

That discount rate is the business model. And it only works if the BNPL provider controls something the merchant can't easily replicate: a consumer audience that's already approved for credit and ready to spend.

Klarna processed $127.9 billion in gross merchandise volume in 2025 across 118 million active consumers and roughly 966,000 merchants. Affirm handled $36.7 billion in GMV the same year, growing 38%.

But the business those numbers describe is shifting under Klarna's feet.

The brand is disappearing

Stripe offers Klarna and Affirm as payment methods through its API. A merchant using Stripe can add pay-in-four to their checkout with a few lines of code. The consumer sees the installment option. They don't necessarily see Klarna.

Adyen does the same. So does Checkout.com. The payment processor handles the integration, the BNPL provider handles the underwriting, and the merchant gets a native-looking installment option without building a relationship with the BNPL company directly.

Apple tried to own this layer itself. It launched Apple Pay Later in 2023, offering four-installment lending directly through Apple Pay. Fifteen months later, Apple shut it down and partnered with Affirm instead. Building a consumer lending product requires credit infrastructure, loss reserves, and regulatory compliance that even Apple decided wasn't worth owning.

The pattern is consistent. BNPL is moving from the app-driven model, where the provider owns the consumer, to the merchant-partner model, where the provider is a service behind someone else's checkout. And increasingly, the payment processor sits between them.

Who captures the economics

When BNPL was a consumer product, the economics were clear. The provider acquired consumers through its app, drove traffic to merchants, and charged a merchant discount rate for the privilege. The brand was the moat. Consumers opened Afterpay to shop, the way they opened Amazon.

When BNPL becomes infrastructure, the economics shift. The payment processor takes a margin for routing the transaction. The BNPL provider gets a smaller share of the merchant discount. The merchant pays less because there are more providers competing for the same checkout slot.

This is the same compression that happened to payment processing itself. Twenty years ago, a payment gateway was a product you sold directly to merchants. Today, it's a feature inside Stripe, Adyen, or Square. The margins compressed, the volume grew, and the winners were the platforms that owned the merchant relationship.

BNPL is on the same path. The question for Klarna and Affirm is whether they end up as Visa, the network everyone routes through, or as the gateway, the invisible layer that got commoditized.

How risk changes when the brand disappears

In the branded model, the BNPL provider owns the credit risk. Klarna decides whether to approve a consumer, funds the merchant up front, and collects installments from the buyer. If the buyer defaults, that's Klarna's problem.

In the infrastructure model, the credit risk doesn't disappear. Somebody still has to fund the merchant and absorb defaults. But the question of who gets murkier.

A white-label BNPL product offered through a payment processor, backed by a bank partner, with the merchant's brand on the checkout page creates a chain of entities where responsibility for credit losses, dispute resolution, and regulatory compliance is distributed across three or four parties.

The CFPB issued an interpretive rule in 2024 classifying BNPL providers as credit card issuers under the Truth in Lending Act. That rule gave consumers dispute and refund rights similar to credit cards. The CFPB's own research found that more than 13% of BNPL transactions involved a return or dispute, significantly higher than the rate for traditional credit cards.

But the rule was written for a world where the BNPL provider has a direct relationship with the consumer. When the provider is invisible infrastructure behind a merchant's checkout, the consumer may not even know who issued the credit. The regulatory framework assumes a brand the consumer can identify and complain to. White-label BNPL removes that assumption.

The plumbing precedent

Payment processing followed this exact trajectory. It started as a product companies sold directly, became a feature embedded in platforms, and eventually turned into infrastructure with thin margins and massive volume.

The winners stopped selling a product and started selling reliability, coverage, and cost per transaction.

BNPL is about five years into the same transition. Klarna's IPO valued the company as a fintech platform, not a lending brand. Affirm's partnership with Apple positions it as infrastructure, not a consumer app. The white-label providers launching with 4-to-12-week integration timelines are selling commodity lending rails.

The pay-in-four button at checkout used to be a product. It's becoming a feature. And features, once they're everywhere, become plumbing.