Affirm has never posted a full-year GAAP net profit in its existence as a public company, and yet the consensus on Wall Street spent most of 2024 and 2025 treating the buy-now-pay-later shakeout as yesterday's problem. The prevailing argument assumes that because the sector has consolidated and weeded out the weakest players, the remaining operators are on a glide path to sustainable margins. That argument is fundamentally flawed because it confuses survivorship with structural repair.
The BNPL shakeout is not over because the fundamental unit economics of lending short, absorbing credit losses, and paying for customer acquisition through merchant subsidies have not changed, regardless of which logos survive.
The misconception being challenged here is the idea that market consolidation equals fixed BNPL unit economics. Klarna's 2025 IPO, Affirm's revenue growth, and the quiet exit of smaller operators have been widely read as signs of a maturing industry finally finding its floor. However, the data reveals a much more precarious reality where the survivors are certainly bigger, but they remain unprofitable in any structurally durable way. Scale has arrived, but margin has not followed.
This analysis holds that the model, as currently structured, remains a massive bet on a permanently low credit-loss environment and a cost of capital that simply no longer exists. Both of those assumptions were wrong at inception, and they remain demonstrably wrong in May 2026.
Why BNPL Unit Economics Are Still Bleeding Cash
The dominant narrative runs roughly as follows: the easy-money era inflated valuations beyond reason, weaker players were flushed out, and the remaining operators including Affirm, Klarna, and Block's Afterpay have built real distribution advantages that justify a clear path to profitability. Analysts attempting to value this path have offered fragmented frameworks, with Mizuho and KeyBanc pointing to Affirm's revenue less transaction costs metric to justify price targets clustering in the low-to-mid $60s, even as CFRA and Barclays counter that the model still requires credit losses below 4 percent of GMV and struggles against the rigid ceiling of 3 to 6 percent merchant take rates. Those are all reasonable frameworks for a traditional payments business, but they are missing the point about how this specific credit engine actually functions.
Klarna's IPO, which landed in 2025, gave institutional investors their first clean look at the books in years, revealing a stark contrast between top-line momentum and bottom-line reality. While the headline revenue figures looked encouraging to growth investors, the credit loss line told a different story entirely. Klarna's credit loss provisions as a percentage of gross merchandise volume remained stubbornly elevated compared with traditional consumer lenders, which is particularly concerning even after years of claimed underwriting improvements. The company's path to profitability depended heavily on assumptions about loss rates that have not materialized consistently across credit cycles, which means the underlying risk engine is still miscalibrated.
Affirm's case is perhaps even more instructive for operators trying to understand the gap between scale and health. The company's fiscal year 2024 showed revenue approaching $2.3 billion, which represents a genuine achievement for a business that barely existed a decade ago. And yet, net losses continued to accumulate, while the quality of earnings remained heavily dependent on gain-on-sale accounting from its securitization program. Strip out that specific accounting treatment, and the underlying lending operation looks considerably less attractive to any fundamental investor. Affirm's stock price has historically rewarded revenue growth, but it has not rewarded durable unit economics because those durable unit economics have simply not appeared in the filings.
Block wrote down the value of Afterpay by billions in 2022, and the integration never delivered the massive cross-sell benefits that justified the original $29 billion acquisition price. By 2025, Afterpay's contribution to Block's overall corporate economics remained marginal at best. That is not a shakeout story with a neat resolution, but rather a cautionary tale that is still running its course.
One more market signal matters here, as Sezzle and Zip have both shown that mere survival is not the same thing as financial health. Both names tightened their underwriting standards and trimmed operating spend, yet neither operational change altered the basic arithmetic of thin merchant fees, recurring promotional spend, and elevated loss sensitivity. The consensus mistake across the market is to confuse industry survival with industry health.
Four Structural Breaks the Model Cannot Escape
First, the cost of funds remains a permanent headwind. BNPL operators fund their receivables portfolios through warehouse lines, securitizations, and balance sheet capital. When the Federal Reserve held rates near zero, the spread between the cost of that funding and the fees charged to merchants was highly workable. As rates normalized through 2022 and 2023, that spread compressed sharply, leaving originators with very little room to maneuver. Even with the modest rate reductions seen in late 2024 and into 2025, the funding cost environment is structurally different from the zero-interest-rate anomaly in which these business models were originally designed. Affirm's interest expense as a percentage of revenue climbed sharply between 2021 and 2024, a reality the company is not hiding, though most equity analysts are simply not weighting it correctly in their models.
Second, credit losses at scale have broken the original underwriting thesis. The primary pitch to investors was always that point-of-sale underwriting on short-duration loans would produce loss rates well below those of traditional credit cards. The data has aggressively refused to validate this claim. Research published by TransUnion and Experian through 2023 and 2024 showed that users carried significantly higher total debt burdens than previously disclosed, with many borrowers holding multiple simultaneous obligations that remained entirely invisible to traditional credit bureaus. The Federal Reserve Bank of New York added another hard number to this reality in its 2024 household debt reporting when revolving credit balances crossed $1.14 trillion. That macroeconomic figure matters deeply because point-of-sale financing is not replacing household use, but rather layering directly on top of it. S&P Global Ratings noted in a 2025 surveillance report that representative securitization structures were still being modeled with cumulative net losses near 7.5 percent in stressed cases. That figure is not a rounding error; it is the credit market pricing in a heavy loss regime just to keep the senior tranches standing.
Short-duration loans with merchant fees capped at 3 to 6 percent simply cannot absorb elevated charge-offs and still produce positive returns. The math does not work at any scale.
Third, customer acquisition costs have not fallen as promised. The merchant subsidy model requires constant new merchant partnerships and relentless consumer re-engagement to maintain volume. Klarna has spent aggressively on consumer marketing to build brand affinity, while Affirm has leaned heavily on its Apple partnership and Shopify integration to drive volume. These distribution deals cost real money upfront and immediately compress a fee margin that was already dangerously thin. The long-standing argument that network effects would naturally reduce acquisition costs over time has not proved out at the industry level, forcing operators to keep buying their growth.
Fourth, regulatory pressure is tightening the compliance burden. The Consumer Financial Protection Bureau's 2024 interpretive rule, which classified these loans as credit cards under TILA, imposed strict disclosure and dispute-resolution requirements that increase compliance overhead without adding a single dollar of revenue. This regulatory cost lands hardest on operators without existing, bank-grade compliance infrastructure, which describes almost all of them.
None of these four pressures are temporary, and none of them disappear just because the market is now more concentrated. Consolidation can certainly make the competitive map simpler to read, but it does not rewrite the underlying cost structure.
The Bull Case Deserves a Fair Hearing
The strongest objection to this bearish analysis is that it is fighting the last war. The argument from bulls, including long-term holders at funds like ARK Invest and Tiger Global, is that the survivors have used the consolidation period to build massive data moats. They argue that proprietary underwriting models trained on tens of millions of transactions will eventually produce structurally lower loss rates than the 2021-to-2023 industry averages imply. Klarna's AI-driven credit decisioning and Affirm's data-sharing arrangement with Experian represent genuine, structural improvements to the underwriting stack, meaning these are not merely cosmetic changes designed for investor decks.
This argument deserves to be taken seriously because better data absolutely does improve credit outcomes over time. The critical question for allocators is whether the improvement is large enough to offset the structural funding and regulatory headwinds described above.
The empirical evidence says it is not, at least not yet. The bar for this analysis to be proven wrong is highly specific: Affirm would need to demonstrate two consecutive fiscal years of positive GAAP net income with credit loss rates strictly below 4 percent of GMV, all without relying on gain-on-sale accounting as a material contributor to the bottom line. Klarna would need to show that its IPO-era loss provisions were genuinely contracting in absolute terms, not merely shifting off-balance sheet through complex funding structures. Neither company is close to meeting those specific operational bars as of May 2026. The data, rather than the corporate narrative, will ultimately settle this debate.
Where Capital Gets Burned From Here
The implications of a sector that remains structurally unprofitable are vastly different for each constituency watching this space. The outcomes are not all bad, but they are all highly consequential for how capital is deployed over the next two years.
Institutional Investors
The primary trap for institutional allocators is the IPO narrative. Klarna's public listing created a massive mark-to-market event that essentially forced portfolio managers to take a position. The temptation for growth funds is to buy the dip, reasoning that a steep discount from peak valuation represents a built-in margin of safety. It does not, particularly if the underlying business model has not fixed its fundamental unit economics. Klarna's 2025 IPO price represented a dramatic retreat from the $45.6 billion peak valuation achieved in June 2021. That steep discount priced in historical risk, but it did not price in a resolution to the ongoing margin compression.
Public equity investors should immediately stop treating revenue less transaction costs, or RLTC, as a viable substitute for net income. This metric matters because it shows how much gross profit remains after direct transaction costs are paid, but it does not pay corporate creditors, software vendors, or taxes. Affirm can continue to widen this metric and still miss the mark entirely if charge-offs, servicing costs, and stock-based compensation continue to run hot. The near-term action for portfolio managers is simple: cut exposure to the names that cannot show sequential, quarter-over-quarter improvement in both credit losses and funding spreads over the next two reporting cycles. If a fund owns Klarna, Affirm, or Afterpay-linked exposure, that position should be tied to a strict watch list with hard stop-losses, not treated as a passive growth story.
Credit investors face an entirely separate decision matrix. The senior tranches of securitizations have held up reasonably well because structural subordination matters in fixed income. However, the mezzanine and equity tranches are exactly where the damage will surface first if consumer delinquencies move higher in a softening labor market. The right move in the next quarter is to demand vintage-level loss tables, detailed merchant concentration data, and borrower repeat-rate data before rolling any exposure forward. If those specific disclosures are not made available by the issuer, the position size should be drastically reduced. That is not pessimism; that is basic, fundamental credit underwriting.
The best-run institutions are already moving in that direction. BlackRock and Vanguard do not need a fundamental thesis to own a name through a passive index, but active managers absolutely do. Those active managers should stop confusing sector survivorship with evidence of fixed economics.
Enterprise Buyers and Retailers
For the chief financial officer at a mid-market retailer evaluating a new checkout integration, the calculus has shifted dramatically. Early adopters of Afterpay and Klarna saw real, measurable basket-size lifts, sometimes jumping 20 to 40 percent on comparable transactions. That lift has steadily compressed as market penetration has risen and consumer familiarity has grown into fatigue. Shopify's own data from 2024 showed conversion lift declining relative to earlier cohort comparisons, which serves as a quiet but powerful acknowledgment that the incrementality arguments are weakening across the board.
The merchant fee, still sitting stubbornly at 3 to 6 percent of transaction value, is now significantly harder to justify when the conversion data is moving in the wrong direction. A retailer processing $100 million in facilitated volume at a 5 percent fee gives up $5 million in pure margin before accounting for refunds, chargebacks, or customer support costs. If the true incremental gross margin lift is only 8 to 10 percent on a very small slice of total orders, that fee structure destroys more enterprise value than it creates. Retailers should be renegotiating contracts aggressively right now, demanding performance-based fee structures rather than accepting flat-rate agreements. Any retailer still operating on a flat-rate contract signed before 2023 is almost certainly overpaying for the volume they receive.
The near-term action for merchants is highly concrete. Finance teams must tie these contracts to measurable lift over traditional card-based checkout, rather than to checkout completion alone. They should require monthly reporting on the repeat purchase rate, average order value, refund rate, and the specific impact of installment defaults on customer lifetime value. If a provider cannot mathematically prove that its fee is justified by incremental margin within one renewal cycle, the retailer should aggressively route promotional volume toward cheaper options such as card-linked installments or in-house pay-over-time offers. Macy's, Walmart, and other high-volume merchants already know the math. A higher conversion rate at the top of the funnel is not the same thing as a better contribution margin at the bottom.
Retail finance teams should also begin treating this payment method like paid media, rather than treating it like a strategic necessity. Marketing teams would never sign a blanket, multi-year contract for search ads without strict return-on-ad-spend thresholds in place. The exact same financial discipline belongs here. The category may still help move inventory in select high-ticket bands or for certain younger demographic cohorts. That is a valid, tactical use case, but it is not proof that the provider's economics are repaired.
Product and Engineering Teams
For the builders and engineers inside these companies, the strategic imperative is to move off the pure lending model entirely before the capital markets force the issue. Affirm's move into debit-linked installments and savings products is exactly the right instinct for a company trying to diversify its revenue base. Klarna's pivot toward becoming a broader shopping and financial services platform is a massive bet that its consumer distribution can outlast the deteriorating lending economics. The product teams that will matter most over the next 24 months are those building real-time data-sharing infrastructure with credit bureaus and underwriting integrations capable of getting loss rates below the threshold where the unit economics actually work. That threshold, roughly 3 to 4 percent of GMV in charge-offs, is technically achievable. It just has not been achieved consistently yet across a full credit cycle.
Product leaders should immediately stop celebrating approval growth in isolation. A rising approval rate can actually be a severe warning sign if it is achieved by pushing into weaker credit cohorts or by relaxing merchant-specific risk filters to hit quarterly volume targets. The near-term action for engineering leaders is to ship cohort dashboards that clearly separate first-payment defaults, 30-day delinquency, 90-day loss severity, and repeat borrowing by specific merchant. If engineers and risk managers cannot see those numbers updated daily, then the company is flying blind into a credit storm. Experian and TransUnion integrations are not nice-to-have plumbing for the roadmap. They are the absolute minimum required if the sector wants to look like a disciplined credit product instead of a venture-funded growth subsidy.
The exact same discipline applies to merchant underwriting. A provider such as Affirm or Klarna should not give every retailer the exact same subsidy budget. A luxury merchant with strong average order values and exceptionally low refund rates deserves entirely different terms from a low-margin fast-fashion seller with high return frequency and erratic inventory. The product team should encode those specific risk differences into the pricing engine itself, rather than leaving them to sales teams chasing quarterly bookings bonuses. The immediate task is to reduce subsidy leakage, close weak merchant vintages faster, and force every single new feature to prove it can improve contribution margin within one quarter.
That is how the model becomes defensible over the long term. It happens through tighter controls, cleaner reporting, and fewer excuses, not by repeating marketing slogans.
Two Predictions for the Market
Prediction one: By the end of Q4 2026, at least one major player will be forced to slow origination growth or widen merchant fees because funding spreads and credit losses will fail to improve at the same pace as top-line revenue. The leading indicators for investors to watch are RLTC margin stalling, securitization spreads widening by at least 50 basis points from mid-2025 levels, and merchant concession requests rising sharply in renewal negotiations. If those indicators move in that direction, the broader market will finally stop rewarding top-line growth as if it were proof of model repair.
Prediction two: By the end of 2027, at least one of Affirm, Zip, Sezzle, or a similar standalone operator will be acquired by a traditional bank or global payments network. A buyer like JPMorgan Chase, Visa, or another scaled financial institution will not buy the loan book for its return profile. Instead, it will buy the consumer distribution, the merchant relationships, and the checkout software. The leading indicators for this outcome are already visible across the sector: slower capital markets access, heavier reliance on expensive warehouse funding, board-level review of strategic alternatives, and an increasing number of partnerships with traditional financial institutions. The final exit price will sit well below the target's last public peak valuation. That outcome is not a collapse. That is simply the market finally assigning the business the exact value its BNPL unit economics deserve.
The shakeout was never going to end with a triumphant press release. It ends with a reconciled balance sheet.
Hasn't Klarna's IPO proved that the market believes in BNPL's future?
An IPO proves that there is demand for liquidity from early investors, which is not the same thing as market validation of a business model. Klarna's 2025 listing gave SoftBank and other long-term holders a necessary exit, but the final IPO price sat far below the $45.6 billion peak valuation achieved in 2021. That massive valuation gap matters deeply. It shows that public markets are still heavily discounting the economics, not celebrating them as a solved problem. A stock can trade actively on a public exchange, but that does not mean the underlying credit engine is fixed. The hard question for any buyer is whether Klarna can turn gross merchant volume into durable, GAAP-compliant profit. The empirical evidence still says no.
If merchants keep signing, how can BNPL economics still be broken?
Because merchants optimize entirely for their own conversion rate, not for the payment provider's profitability. A retailer can easily accept a 5 percent fee if the basket lift and customer acquisition benefits outweigh the immediate cost. That dynamic says absolutely nothing about whether the provider actually earns a return after accounting for funding, credit losses, and customer support. Shopify's 2024 data showed conversion lift easing against earlier cohorts, which means the fee is getting significantly harder to defend in procurement reviews. Merchants will keep signing until the incrementality falls below the fee they pay. That inflection point is arriving much faster than the providers want to admit.
Why does regulation matter if BNPL loans are short and transparent?
Short duration does not automatically mean low risk, and transparency is exactly what many of these programs lacked during their hyper-growth phase. The CFPB's 2024 interpretive rule under TILA pushed the sector much closer to credit-card-style disclosure and dispute handling. That regulatory shift adds permanent operational cost, and it does not add a single cent of revenue. It also exposes the consumer overlap that has been easy for bulls to ignore: many of these borrowers already carry heavy credit-card balances and other unsecured debt. When the Federal Reserve reported revolving credit balances above $1.14 trillion, the macroeconomic message was clear. The sector is operating inside a highly stressed consumer balance sheet, not outside it.
Affirm, Klarna, Afterpay, and their peers have successfully proven that consumer demand exists. They have not proven that the operating model can absorb credit losses, higher funding costs, and regulatory overhead all at the same time. That is the core issue. Until those structural costs fall in tandem, BNPL unit economics remain exceptionally fragile, no matter how many survivors remain standing at the end of the year.
Related MarketIntel briefing: read BNPL’s Profit Mirage Is Starting to Crack for a connected view on this market signal.
Source context: readers can compare this market signal with broader data from Reuters.