Klarna's IPO and the BNPL Unit Economics Problem

Buy Now Pay Later is a category that has been alternately declared the future of consumer credit and a systemic risk to household finances, usually within the same six-month window. Klarna’s IPO filing — submitted to the SEC in March 2025 ahead of its September 2025 NYSE listing — provides the most detailed public disclosure of BNPL unit economics ever filed. The numbers tell a more nuanced story than either the enthusiast or the skeptic position.

Understanding the Klarna P&L is useful beyond the Klarna-specific investment question. The BNPL model is now embedded in merchant checkout across e-commerce and physical retail, and the unit economics that determine whether Klarna is a good business are the same economics that determine whether the category survives its regulatory moment.

The BNPL Revenue Model, Decomposed

Klarna earns revenue from two sources: merchant fees and consumer interest/fees.

Merchant fees are the primary revenue driver. Klarna charges merchants a percentage of the transaction value — typically 2-6% depending on product type, merchant vertical, and geography — to offer BNPL as a checkout option. The merchant pays because BNPL drives measurable conversion uplift (Klarna publishes internal studies showing 15-25% basket size increase and 10-15% conversion improvement) and because offering BNPL reduces cart abandonment from customers who prefer deferred payment.

The merchant fee is structurally similar to a card interchange: it’s a percentage of transaction value paid by the merchant to enable a payment method. The difference is that BNPL merchant fees are typically 2-3x higher than card MDR. A merchant paying 1.8% card MDR on credit transactions might pay 4.5% on the same transactions routed through Klarna’s Pay Later product. The merchant accepts this because the conversion improvement more than offsets the fee differential — a claim that holds in most verticals but requires continuous merchant ROI validation.

Consumer interest and fees accrue on Klarna’s installment products — primarily Klarna Financing (longer-term installment loans, typically 6-36 months, interest-bearing) as distinct from Pay in 4 (four equal installments, zero interest, zero fee for consumer). Pay in 4 is effectively a zero-cost credit product for consumers, funded entirely by merchant fees. Klarna Financing is a traditional consumer lending product with APRs ranging 10-30% depending on market and creditworthiness.

The S-1 revenue split: approximately 60% merchant fees, 40% consumer interest and fees. The consumer credit revenue is the higher-margin component — once credit is extended, interest compounds. The merchant fee business is volume-dependent and tied to consumer spending patterns.

The Credit Loss Problem

BNPL credit losses are the existential variable in the unit economics. A BNPL model with 1% gross credit losses on GMV is a healthy business. At 3%, it’s marginal. At 5%, it’s structurally impaired.

The challenge is that BNPL extends credit frictionlessly — that’s the product. There’s no hard credit pull, no extended application, often no income verification for the Pay in 4 product. The underwriting model is behavioral: prior Klarna transaction history, device intelligence, real-time behavioral signals, and external credit bureau data where available. It’s sophisticated ML underwriting, but it operates on very thin decision windows. A customer completing checkout isn’t waiting for an underwriting queue.

Klarna’s disclosed credit loss rate: approximately 0.52% of GMV (provision for credit losses in the twelve months ended June 30, 2025), net of recoveries (Klarna F-1, SEC filing, March 2025). This is the managed, at-scale number following years of model iteration. This is the managed, at-scale number following years of model iteration. It’s also dramatically better than the 2021-2022 period, when rapid geographic expansion combined with the post-pandemic credit environment pushed losses materially higher — contributing to the writedowns and layoffs that dominated Klarna coverage in 2022-2023.

The sensitivity analysis is what matters for the IPO: a 0.5% increase in net credit loss rate — well within historical volatility — reduces Klarna’s operating margin by approximately 8-10 percentage points. At Klarna’s scale ($112B GMV in the twelve months ended June 30, 2025, per Klarna F-1), a 0.5% credit loss increase is approximately $560M in incremental annual losses. This is why the macro environment is so material to BNPL equity valuations: a consumer credit stress cycle doesn’t just reduce volume, it triggers the loss rate sensitivity.

The Gross Margin Structure

Klarna’s gross margin on the merchant fee business is thin. Transaction processing costs (acquiring, fraud, customer service, dispute handling) consume a significant portion of the merchant fee gross revenue. The S-1 doesn’t break out transaction-level margins cleanly, but segment disclosures imply processing cost ratios of 55-65% of merchant fee revenue — leaving 35-45% gross margin on the transaction-level business.

The consumer lending business carries higher gross margins — approximately 60-70% on interest revenue, after funding costs and credit provisions. The blended gross margin across both revenue streams is approximately 40-50%.

These gross margins are lower than pure SaaS (70-80%) but higher than pure payment processing (30-40%). The investor thesis for BNPL has always been that it would evolve toward a higher-margin financial services model — the lending book would grow, interest margin would compound, and the transaction-level business would become a customer acquisition engine for higher-margin credit products. Klarna’s S-1 shows progress toward this but incomplete execution.

The Merchant ROI Calculus by Vertical

BNPL economics vary significantly by merchant vertical. The Klarna merchant fee is justified when the conversion uplift exceeds the fee differential over card acceptance. The math works differently across categories:

Fashion and apparel. Highest BNPL penetration and highest merchant ROI. Average order values of $80-150, high return rates, and a consumer base that’s comfortable with deferred payment. BNPL conversion uplift of 20-30% in fashion is well-documented. The merchant fee of 4-5% is offset by the incremental margin on sales that wouldn’t have occurred. Return rates are a partial offset — Klarna bears the credit exposure but the merchant bears the return handling cost.

Electronics. High average order values ($300-800 for consumer electronics) make BNPL’s zero-interest Pay in 4 a compelling consumer proposition. Merchant BNPL fees are higher in absolute terms but represent similar percentages. Conversion uplift is real but smaller in percentage terms — electronics consumers are often committed to a purchase before reaching checkout, so BNPL is more of a payment method preference than a demand driver.

Travel. Complex and variable. Airlines and OTAs have experimented with BNPL but face a structural issue: travel is a perishable product. If a consumer defaults on a Pay in 4 for a flight that already occurred, the loss is complete. This creates adverse selection — the consumers who most want to defer payment for travel are the ones with the most credit pressure. Klarna’s travel penetration is lower than fashion.

Grocery and everyday spend. Low basket sizes ($30-60) make the math difficult. A 4.5% fee on a $40 grocery basket is $1.80 — real money on thin-margin grocery. The conversion argument is also weaker for necessity purchases. BNPL in grocery has remained niche outside of specific markets (Australia, where BNPL is culturally embedded across more categories).

The Regulatory Headwind

The regulatory environment for BNPL has been moving in one direction since 2022 and represents a structural cost pressure that isn’t fully reflected in current P&L.

Australia. ASIC has been the most aggressive regulator. The Treasury Laws Amendment (Combating Scams and Consumer Harms) legislation, which received Royal Assent in 2024, brings BNPL formally under the National Consumer Credit Protection Act. This mandates responsible lending obligations — affordability assessments before credit extension — that BNPL providers previously avoided by structuring products to fall outside the credit act definitions. The compliance cost of affordability assessment at checkout friction-of-scale is real and ongoing.

UK. The FCA’s BNPL regulation framework, following the Woolard Review, is bringing unregulated BNPL under consumer credit rules. Pre-contract information requirements, complaint handling processes, and FCA registration obligations all add operational cost. UK implementation timeline has slipped several times but is expected effective 2025-2026.

US. The CFPB issued an interpretive rule in May 2024 (effective July 30, 2024) clarifying that BNPL providers are “card issuers” under the Truth in Lending Act, triggering TILA compliance obligations including billing statements, dispute resolution, and chargeback rights (CFPB, May 2024). The Bureau subsequently announced plans to revoke the rule under the post-2025 administration, but the regulatory trajectory — treating BNPL as consumer credit — remains the direction of travel globally.

For Klarna, these regulatory changes have two effects: compliance cost (significant legal and operational infrastructure to build) and credit policy restriction (affordability checks will decline some transactions that would currently be approved, reducing GMV). The net impact on Klarna’s long-run economics is negative, though the magnitude depends heavily on how aggressively regulators enforce and how efficiently Klarna builds compliance infrastructure across 45 markets.

The Competitive Landscape at IPO

Klarna’s primary BNPL competitors — Afterpay (now Block), Zip, and Affirm — have taken divergent paths:

Afterpay/Block. Integrated into Block’s ecosystem alongside Cash App and Square. The BNPL product continues to operate but is increasingly positioned as a Cash App user acquisition tool rather than a standalone business. Merchant fee structure broadly similar to Klarna. Geographic concentration in AU/US/UK.

Affirm. US-focused, Shopify partnership (Shop Pay Installments is Affirm-powered), higher-ticket items. Affirm’s S-1 and subsequent public disclosures show persistently high credit loss volatility — confirming that the credit underwriting problem is sector-wide, not Klarna-specific. Affirm’s consumer lending book is larger relative to its transaction volume, reflecting a more explicit credit product strategy.

Zip. Has retrenched significantly — exited numerous markets and restructured to focus on AU/US. Smaller scale, more credit-focused product mix. Effectively serves as a data point on what happens to BNPL businesses that don’t achieve Klarna’s scale: they become niche credit products without the merchant distribution advantages.

Klarna’s competitive advantages at IPO: merchant distribution (790,000+ merchants in 26 countries as of June 2025, per Klarna F-1), 111 million active consumers, and an AI investment that has materially reduced customer service headcount (Klarna’s AI assistant handles two-thirds of customer service interactions, with Klarna headcount falling from 5,527 at end-2022 to 3,422 by end-2024 — a 40% reduction, per CNBC, May 2025).

The AI Narrative and What It Actually Means

Klarna has positioned itself aggressively as an AI-first company in its IPO narrative. The claims: Klarna’s AI assistant handles two-thirds of customer service interactions, saving $40M annually in support costs. The AI underwriting model has improved credit performance. The AI personalization layer drives higher conversion on merchant surfaces.

The AI cost reduction is real and verifiable from headcount trends — Klarna reduced headcount by approximately 40% from peak to IPO filing, with AI automation being a meaningful contributor. Customer service costs per transaction dropped 40% over two years, from $0.32 per transaction in Q1 2023 to $0.19 in Q1 2025 (CX Dive, 2025). The AI assistant is reported to handle the work equivalent of 853 full-time staff. By early 2026, Klarna began reversing course on AI-only service, adding human agents back after customer satisfaction declined on complex interactions — indicating that the fully-automated model had limits.

The AI underwriting improvement is harder to attribute specifically — credit models improve with scale and data accumulation, and the 2023-2024 credit loss improvement reflects both better ML and a more stable macro environment. The contribution of AI specifically versus environmental normalization is not separately disclosed.

What the AI narrative provides that matters for the IPO: a growth vector that is independent of credit loss cycles and regulatory pressure. If Klarna is a payments data and AI company rather than a credit company, the multiple expansion argument is available. Whether that repositioning holds up to scrutiny depends on whether the software/data revenue becomes material as a standalone line item. Current disclosures don’t yet support that characterization cleanly.

The Verdict on BNPL Unit Economics

Klarna at IPO is a profitable business with genuine competitive advantages, thin margins, high credit loss sensitivity, and a regulatory cost curve that’s still ascending. It’s not the existential credit risk that bears described in 2022, but it’s also not the high-multiple software business that its AI narrative implies.

For merchants evaluating BNPL relationships: the model works where average order values are high enough that conversion uplift exceeds fee differential, and where the consumer base has sufficient credit quality that regulatory affordability requirements don’t drastically change acceptance rates. Fashion, electronics, and higher-ticket retail remain the strongest verticals. Grocery and low-AOV repeat purchase categories are where the math is most difficult and where BNPL’s long-run merchant ROI is least clear.

The more interesting question for the payments industry isn’t whether Klarna as a public company succeeds — it’s whether the BNPL merchant fee structure is sustainable as card networks extend installment products (Visa Installments, Mastercard Installments) that allow issuers to offer comparable BNPL functionality at card-interchange economics. If that competition intensifies, the 4-5% merchant fee that funds the BNPL model faces structural pressure that no amount of AI narrative can offset.