FICO built two credit scoring models made specifically to read buy now, pay later loans, then found the credit bureaus did not have enough BNPL data on file to run them properly. Nearly a year and a half after FICO unveiled FICO Score 10 BNPL and FICO Score 10 T BNPL, whether a BNPL credit score moves a US borrower’s file still depends on how much loan data Affirm, Klarna and Afterpay choose to send to Experian, Equifax and TransUnion, and how much they hold back.
Two scores built for a product mainstream models miss
FICO announced FICO Score 10 BNPL and FICO Score 10 T BNPL on 23 June 2025, describing them as the first credit scores from a major scoring provider built around BNPL data. The two models sit alongside FICO’s existing Score 10 Suite rather than replacing it, and FICO offers them to lenders at no extra fee so a bank can test the BNPL-aware score against the version it already uses. Julie May, FICO’s vice president and general manager of B2B Scores, said the models followed requests from large US lenders for a way to read BNPL repayment without penalising borrowers for how these loans are structured.
The problem FICO had to solve is that one shopper can open several short BNPL loans within days of each other, which an ordinary model could misread as a spike in risk. A joint study with Affirm covering more than 500,000 consumers found that grouping a borrower’s concurrent BNPL loans before scoring kept the model’s predictive accuracy while limiting the damage to individual scores: 85 percent of BNPL users saw a score move of fewer than 10 points, and 97 percent of the heaviest users, those with five or more open BNPL accounts, moved fewer than 20 points.
Why a BNPL credit score still is not live
The gap sits upstream of FICO entirely. In a blog post published on 12 March 2026, FICO said the two BNPL scores would be available at the credit bureaus only once BNPL data was being furnished to those bureaus at scale, a condition that, nine months after the original fall 2025 target, had still not been met.
The Consumer Financial Protection Bureau has already measured how large that data gap is. A December 2025 study by the Bureau found that 53.6 million Americans used a BNPL loan in 2023, taking an average of 6.3 loans from a single lender that year, while 63 percent of BNPL borrowers carried more than one loan at once and about a third used more than one provider, obligations that stay largely invisible to a lender pulling a standard credit file.
Affirm reports. Klarna and Afterpay hold back.
Affirm has gone furthest of the big providers. Under its own reporting policy, every Affirm payment plan opened on or after 1 April 2025 is reported to Experian, with coverage extended to TransUnion for plans opened from 1 May 2025. Affirm’s own policy page is specific about what that does not yet mean: the new reporting “won’t be factored into your traditional credit scores in the near term, but may in the future as new credit scoring models are developed.”
Klarna and Afterpay have taken the opposite position. Research published by the Federal Reserve Bank of Richmond in February 2026 recorded that both providers have cautioned against reporting BNPL activity to the bureaus, arguing that models built for longer-term revolving or installment credit could misread frequent short-term BNPL use, or a single missed payment, as a bigger risk signal than it is. Sezzle sits between the two positions, letting individual borrowers opt in to reporting rather than reporting every account by default.
Equifax built the plumbing for BNPL reporting years before either side of this argument existed. The bureau introduced a dedicated business industry code for pay-in-four loans in February 2022 so BNPL tradelines could be filed in the Consumer Data Industry Association’s Metro 2 format, and it still holds the capability to suppress those tradelines from a lender’s existing scoring model while the industry works out how to treat them.
What is riding on the data catching up
The dollar figures explain why FICO, VantageScore and the bureaus are bothering. The Federal Reserve Bank of Richmond put total US BNPL transaction value at roughly 70 billion dollars in 2025, about 1.1 percent of credit card spending, growing close to 20 percent a year since 2021. VantageScore built VantageScore 4.0 as the first tri-bureau model to read trended account data rather than a single snapshot, so it does not need a separate BNPL version the way FICO does. VantageScore’s own leadership argued in 2024 that responsible BNPL repayment could eventually help borrowers reach other credit products, including a mortgage.
None of this has moved a US consumer’s score yet. Nothing in FICO’s, Affirm’s or Equifax’s own published material from this year states that a live BNPL-aware score is running at a US lender today. Each describes a model or a reporting pipeline that exists but is waiting on the other half of the system: the modelling work has been ready, by FICO’s own account, since the fall of 2025, and the holdout sits with the bureaus and the BNPL providers still deciding how much data to send them.
A fuller list of active US lending and credit-decisioning providers working through that data gap is on Fintechly’s lending sector directory.