Affirm’s weighted average cost of capital reached 11.2% in 2025, climbing sharply from 8.5% in 2023, while its net interest margin remains stubbornly anchored at 9%. That negative spread is not a temporary macroeconomic glitch. It is a mathematical guarantee that the core business model is bleeding cash, because a lender cannot survive when its funding costs exceed its returns. Despite a brief rally in late 2025, the unit economics of the BNPL sector remain fundamentally broken. The market’s optimism is misplaced, which means the shakeout is not over but merely moving into its messiest phase. The conventional wisdom says the sector's trouble was temporary, a mix of post-pandemic normalization and macro pressure. Analysts at Goldman Sachs and Morgan Stanley have pointed to improving delinquency rates and steadier merchant fees as signs of a turn, and Affirm’s Q4 2025 earnings showed a 12% sequential rise in gross merchandise volume that was immediately hailed as proof the sector had turned a corner. That reading is wrong because the underlying economics still do not work, leaving structural flaws to become more pronounced as capital costs remain elevated.
The Consensus Is Missing the Point
The bull case for the sector rests on three bad assumptions: that merchant fees can keep rising, that credit losses will stay contained, and that scale will eventually solve the margin problem. None of those assumptions holds up to scrutiny when applied to the actual operating data of the major players. Take merchant fees, where the consensus assumes providers can keep charging 3% to 6% per transaction. That expectation ignores the price war already in motion. Klarna cut its U.S. merchant fees to 2.99% in early 2025 to deliberately undercut Affirm and Afterpay, protecting market share at the expense of profitability. Amazon’s entry into the space in 2024 forced incumbents to follow or lose volume entirely by offering fees as low as 1.9%. The result is a catastrophic loss of margin, with estimates from Wall Street analysts clustering together to highlight the damage: JPMorgan’s consumer finance team calculated in 2025 that every 50 basis points of fee compression wipes out roughly $180 million in annual operating profit across the listed cohort, while UBS modeled a massive 260 basis point hit to margin if merchant fees drift below 2.8%. That is not a durable pricing pool. It is a race to the floor driven by massive retail platforms that view financing as a commodity.
Then there is credit risk. Delinquency rates have stabilized only because lenders tightened underwriting standards drastically, reducing approval rates by 20% to 30% across the industry. That is not health. It is retreat. Morgan Stanley estimated that providers rejected more than one in four applicants in 2025, which delivers a direct hit to the mass-market story that once supported the sector’s high valuations. Affirm’s net charge-off rate still sits at 8.7% even after easing from its 2024 peak. This figure remains well above the 5% to 6% range where unsecured consumer lending starts to look profitable for traditional banks, which means the core lending engine is still mispriced.
Finally, the scale argument fails in practice. The idea that these companies will reach profitability through sheer volume assumes economies of scale that simply do not exist in this specific product structure. Unlike credit cards, which benefit from interchange fees and revolving balances that compound over time, the short-term fixed-fee model has no natural moat. Every incremental dollar of gross merchandise volume requires more customer acquisition, more underwriting, and more collections work. Affirm’s operating expenses grew 18% year over year in 2025, outpacing its revenue growth. And yet, the market continues to reward volume. Scale is not solving the problem. It is making the problem larger and more expensive to maintain.
Five Structural Fractures Driving the BNPL Shakeout
The shakeout is being driven by five specific data points that show exactly how weak the model remains under current market conditions.
First, the cost of capital is fundamentally misaligned with the return profile. Providers fund loans through securitizations, warehouse lines, and equity, leaving Affirm with a weighted average cost of capital at 11.2%. By comparison, Capital One enjoys a cost of capital of about 7.8% because it possesses a massive deposit base and a diversified funding stack. Standalone fintech firms lack this foundation, relying instead on capital markets that demand a premium for risk. That structural disadvantage means Affirm must earn returns above 11% just to break even, an impossible hurdle when its net interest margin is stuck at 9%. That leaves the company in a position where every loan originated under these conditions destroys shareholder value, turning origination volume into a liability rather than an asset.
Second, customer acquisition costs are rising while lifetime value is shrinking. Klarna spent $1.2 billion on marketing in 2025, representing a 40% year-over-year increase, yet its active customer base grew only 12%. The math is brutal for the entire sector because acquiring a user costs roughly $60 to $80, but the average customer generates only $30 to $40 in lifetime revenue. Even Affirm, with a stronger merchant network, saw its acquisition cost rise to $72 in Q4 2025 from $55 a year earlier. That is not a growth curve. It is capital leakage on a massive scale. When a company pays more to acquire a customer than that customer will ever return in fees, the business model functions more like a subsidy for retail shopping than a sustainable financial enterprise.
Third, the regulatory vise is tightening across multiple jurisdictions, stripping away the operational freedom that allowed these firms to grow unchecked. The Consumer Financial Protection Bureau’s 2025 rule requiring providers to report loans to credit bureaus added significant compliance costs and, more importantly, increased delinquencies by 15% to 20% as weaker borrowers were screened out by other lenders who could finally see their hidden debt loads. Across the Atlantic, the European Union’s proposed cap on fees at 1.5% would crush Klarna’s European margin profile overnight. The era of regulatory arbitrage that once served as the sector's hidden advantage is ending, forcing these platforms to bear the same compliance burdens as traditional banks without enjoying the same funding advantages.
Fourth, the merchant value proposition is fading rapidly. Early adopters like Shopify and Peloton embraced the product because it lifted conversion rates by 20% to 30%, but that incremental lift shrank as the option became common across the internet. A 2025 study by McKinsey found that the option now lifts conversion by only 5% to 8%. That is a small gain that does not justify a 3% to 6% fee for the retailer, and merchants are pushing back aggressively. Walmart and Target both cut partnerships in 2025, moving toward in-house installment plans with drastically lower costs. The result is that when the largest retailers in the world realize they can build the same functionality internally for a fraction of the price, the external providers lose their most lucrative enterprise channels.
A fifth point strengthens the case against the sector's long-term viability by exposing the true nature of the consumer base. The Federal Reserve Bank of New York reported in 2025 that 37% of users were also carrying revolving credit card balances. That overlap matters because it shows the product is not replacing expensive debt with cheaper debt, as the marketing claims, but rather adding another layer of short-term liability on top of an already strained household balance sheet. That kind of customer is fragile, and fragile customers do not support stable lender economics during a downturn. When macroeconomic pressures mount, these highly leveraged consumers will default on unsecured installment loans long before they default on their primary credit cards or auto loans.
The Embedded Finance Counter-Argument Fails
The strongest defense of the sector points to embedded finance, where proponents argue providers can grow into full-stack financial firms offering savings accounts, debit cards, and even mortgages to diversify revenue. Afterpay’s 2025 launch of a high-yield savings account at 4.5% APY and Klarna’s push into banking licenses in Sweden and the UK are treated as proof that the model can evolve.
That argument ignores a basic fact of banking. These providers lack the core strength of actual banks, which is cheap and stable funding. Klarna’s savings product has gathered $2.3 billion in deposits, but that is tiny relative to its $15 billion loan book. Without a real deposit franchise, the company remains heavily exposed to capital markets, creating a fatal weakness in any rising-rate environment. A lender that depends on wholesale funding cannot pretend to be a bank simply by adding a savings label to its app.
The pivot also requires major upfront spending on compliance, systems, and staff. Affirm’s banking-as-a-service push launched in 2024 has burned through $300 million with no clear path to strong returns. Meanwhile, Chase and Citi are rolling out their own installment products through existing infrastructure to existing customers, requiring zero additional acquisition cost. The embedded finance story delays the reckoning for fintechs without actually solving the underlying unit economics. Traditional banks already own the customer relationship and the deposit base, making it nearly impossible for a standalone installment lender to cross-sell banking products profitably.
Strategic Mandates for the Next Phase
The ongoing contraction is not just about failing startups. It is a structural change that will hit investors, enterprise buyers, and product teams in different ways, leaving each group facing a narrower set of choices than the market wants to admit.
For Investors
Investors need to stop treating this as a growth story and start pricing it as a lending business with weak funding and weak margins. Affirm trades at 12x forward revenue even though it posts a negative 20% free cash flow margin. That multiple makes no sense beside Discover Financial, which trades closer to 6x revenue and produces a 15% return on equity. A lender with no deposit base should never command a software-style valuation, and that spread is a warning rather than an opportunity.
The near-term action is simple. Reduce exposure before Q2 2026 earnings by watching funding costs, gross margin, and approval rates instead of gross merchandise volume headlines. Short interest in Affirm already sits at 18% of float, up from 12% in 2025, and that figure is likely to rise as analysts reset their models. The lead indicators are plain. Watch for wider securitization spreads, weaker merchant fee guidance, and any pause in buyback activity. If those metrics move in the wrong direction, the market will stop calling this a fintech premium story and start calling it subprime lending with a better app.
There is another risk. Investors who assume consolidation will rescue the category are ignoring the shape of the likely deals. Any acquisition of a platform by a larger payments or lending firm will probably happen at a steep discount to book value, meaning the next round of mergers and acquisitions will be about damage control rather than expansion. Pay attention to buyers like PayPal and Block. If those firms hesitate, that is a clear signal that the assets are weaker than the headline user counts suggest.
For Enterprise Buyers
Retailers and merchants should treat this as a checkout tool rather than a strategic partner. Shopify and Peloton once saw real conversion gains, but those gains have faded to low single digits, meaning a fee of 3% to 6% for a 5% to 8% conversion lift is a weak trade. Walmart and Target already moved in-house on installment offers in 2025 because they could lower cost and retain control of the customer relationship, providing the exact model for other large retailers to copy.
The near-term action is to renegotiate every contract now. Merchants should push for fee floors below 2.5%, tighter performance clauses, and the ability to route high-quality customers to cheaper payment rails. If a provider cannot meet those terms, the merchant should replace it with card-linked offers, issuer-led plans, or a direct bank partnership. The data point to watch is incremental basket lift versus fee burden. If the lift falls below 10% and the fee stays above 3%, the deal is losing money in plain sight.
Enterprise finance teams also need to look at customer mix. The product works best in verticals such as healthcare and home improvement, where ticket sizes are higher and repayment horizons are clearer. It performs poorly when used to subsidize low-margin consumer spending, which is why buyers should split usage into narrow use cases and avoid broad checkout placement. Amazon’s low-fee push showed that the platform winner in this category is still the merchant with scale, not the lender with the app.
Related MarketIntel briefing: read BNPL's Unit Economics Still Don't Add Up: And the Shakeout Isn't Over for a connected view on this market signal.