Klarna: $40 IPO to $15 as Credit Provisions Surged 102% and the CEO Admitted AI Cuts Went Too Far
Revenue grew 25% to $3.5B. The merchant network expanded 42% to 966K partners. The Board Chair bought $50M in stock at $14. The securities class action says credit risk was hidden. Six lenses examined which narrative holds.
+25% YoY, Q4 accelerated to +38%
~$750M, rate rose from 0.44% to 0.72% of GMV
Down 60% from $40 IPO price (Sep 2025)
3.47M shares purchased at ~$14 avg
Klarna Group plc arrived on the NYSE in September 2025 with a compelling pitch: the world's largest BNPL provider, newly profitable, AI-powered to the point where it had halved its workforce, and poised to become a full digital bank. The market embraced the story, pricing shares at $40 and giving the company a $20B market cap on day one.
Within six months, that narrative cracked. Credit loss provisions surged 102% year-over-year. The company swung from profit to a $0.79/share loss. The CEO publicly admitted the AI workforce cuts "went too far." A securities class action was filed alleging Klarna knew the credit deterioration was coming when it filed at $40. The stock collapsed 60% to $15.67.
Then something unusual happened. Board Chair Michael Moritz -- Sequoia Capital partner, one of the most experienced tech investors alive -- bought $50M worth of stock at approximately $14/share in early March 2026. This was not a token gesture. It was the largest insider purchase since the IPO, executed with personal capital at distressed prices.
Want the full 6-lens analysis with signal assessments and model debates?
Opus + Sonnet ensemble. 6 lenses. 9 signals. 11 debates. Full evidence citations.
Signal Assessments: What 6 Lenses Found
102% provision surge, $335M adjusted-GAAP gap, revenue-provision timing asymmetry on US originations
Chairman bought $50M at ~$14 (bullish), but CMO and CCO sold at lock-up expiration via 10b5-1 plans
$3.5B revenue (+25%) with 966K merchants, but every BNPL transaction creates credit risk
Securities class action with strong factual basis; CFPB deregulating but EU tightening
8.2x leverage, $400M GAAP loss, equity market impaired at $15 vs $40 IPO
IPO capital consumed by losses; no disclosed GAAP profitability timeline
966K merchants but PayPal leads US usage (56% vs 38%); merchant switching demonstrated
IPO AI narrative partially retracted; 60% decline has largely repriced the overvaluation
At 1.7x revenue, not cheap for an unprofitable lender with 8.2x leverage and active class action
Key Findings
The 102% Provision Surge Is the Central Data Point
Credit loss provisions surged from ~$370M to ~$750M year-over-year, outpacing the 25% revenue growth and 22% GMV growth. The provision rate as a percentage of GMV increased from 0.44% to 0.72% at its Q3 peak before improving to 0.65% in Q4.
Growth Explanation
Rapid US expansion into a market where 63% of BNPL users carry multiple loans. Under IFRS 9, provisions are front-loaded at origination. High growth quarters will always show elevated provisions relative to revenue.
Deterioration Explanation
The provision rate increase (0.44% to 0.72%) exceeds what pure volume would explain. Late payments rose from 33% to 40% of users YoY (LendingTree). The 2/3 subprime applicant mix creates structural credit vulnerability.
The AI Narrative Was Partially Validated, Partially Retracted
Klarna reduced its workforce from 7,400 to approximately 3,000 employees, generating genuine cost savings and funding 60% pay increases for remaining staff. The OpenAI-powered chatbot replaced the equivalent of 800 customer service agents. These are real operational achievements.
However, CEO Sebastian Siemiatkowski publicly admitted in January 2025 that the cuts "went too far." Customer complaints about chatbot quality forced rehiring of human agents. The AI tools are licensed from OpenAI (commercially available to all competitors), undermining the structural moat narrative. Two lenses (Moat Mapper and Myth Meter) independently classified this as an operational advantage with a 12-18 month lead, not a durable moat.
The Chairman's $50M Purchase Is the Strongest Contrarian Signal
Michael Moritz (Board Chair, Sequoia Capital partner) purchased 3.47M shares for approximately $50M between March 3-11, 2026 at an average price around $14. This followed the lock-up expiration on March 9 and occurred while the stock was near 52-week lows. Board chairs rarely deploy this magnitude of personal capital unless they have high conviction in value recovery.
By contrast, the CMO sold $490K and the CCO sold $380K via pre-established 10b5-1 trading plans. Net insider flow is overwhelmingly positive at approximately $49M of net buying.
Where the Models Disagreed
Growth Cost vs. Deterioration Signal
Provision growth reflects deliberate US expansion into higher-risk demographics. The Q4 sequential improvement (0.65% vs 0.72%) suggests stabilization. This is a growth cost, not a deterioration signal.
The 102% provision surge significantly outpaces 22% GMV growth. The provision rate increase (0.44% to 0.65-0.72%) indicates genuine credit quality deterioration beyond volume effects.
Both views contain truth. The provision rate increase confirms some credit deterioration beyond pure volume effects, but the Q4 sequential improvement suggests active risk management. Classified as QUESTIONABLE.
Overcorrection or Justified Repricing?
At 1.7x revenue with 25% growth, the stock overshoots fundamentals. The Chairman's $50M purchase supports the oversold thesis. Revenue, merchant network, and AI savings all have more value than $6B implies.
The $40 IPO was mispriced. 1.7x revenue is appropriate for an unprofitable lending business with 8.2x leverage, a class action, and subprime exposure. The current price reflects the real business, not the IPO narrative.
MODERATE DISCONNECT. The IPO was clearly stretched, and the 60% decline has overshot to some degree. The business has more value than $6B implies, but material uncertainties justify a significant discount. The gap has narrowed but not closed.
Is the Merchant Network a Moat?
966K merchants at 42% growth rate and 118M consumers create network effects. Merchant integrations create switching costs. The Walmart win validates enterprise sales capability.
Walmart switching from Affirm to Klarna proves BNPL providers are interchangeable. PayPal has higher US usage (56% vs 38%). The "moat" is brand awareness, not structural lock-in.
CONTESTED. Scale advantages exist but demonstrated merchant switching proves relationships are contestable. The moat is real but narrow, dependent on pricing competitiveness rather than structural lock-in.
Where Lenses Agreed
Credit risk is the central vulnerability
Four lenses (Fugazi Filter, Gravy Gauge, Stress Scanner, Regulatory Reader) converged on credit risk as the primary threat. The provision rate increase from 0.44% to 0.65-0.72% of GMV is confirmed at E3 evidence level.
AI cost advantage is real but not a moat
Moat Mapper and Myth Meter agree the workforce reduction created genuine savings, but both classify the advantage as operational (12-18 month lead) rather than structural (durable moat).
Valuation has compressed but risk premium is justified
Myth Meter and Stress Scanner agree the 60% decline partially addresses the IPO overvaluation. At 1.7x revenue, the entry is more reasonable but the stock still prices in margin improvement that depends on credit stability.
Regulatory relief is near-term, not permanent
Regulatory Reader and Gravy Gauge agree that CFPB withdrawal provides breathing room, but the evidence base for future enforcement is accumulating. EU regulation tightens regardless.
What to Watch
Currently 0.65% (Q4), down from 0.72% peak (Q3). Above 0.80% signals credit deterioration beyond management control. Below 0.50% signals stabilization. This is the single most important metric for KLAR.
Lead plaintiff appointed. Motion to dismiss ruling is the next milestone. Dismissal would reduce regulatory exposure materially; proceeding to discovery maintains elevated risk and may reveal internal credit assessments.
Currently 1.9% (FY2025), guided to 6.9% (FY2026). Below 3.5% by Q2 2026 would indicate guidance miss. This is the key credibility test for management after the AI narrative partially unraveled.
335M shares eligible since March 9. So far, selling has been modest relative to the Chairman's $50M purchase. Heavy insider selling (50M+ shares in 60 days) would shift governance from MIXED to MISALIGNED.
Bottom Line
HIGHER SCRUTINY
Klarna has genuine operational strengths -- 25% revenue growth, 966K merchants, real AI cost savings -- but the business model embeds credit risk in every transaction. The 102% provision surge, 8.2x leverage, securities class action, and subprime consumer exposure create a risk cluster that warrants elevated investigation. The Chairman's $50M purchase and the 60% valuation compression prevent an AVOID classification, but credit conditions must stabilize before the risk/reward improves meaningfully.
Path to More Favorable Assessment
- • Credit provision rate declining below 0.50% of GMV
- • Securities class action dismissed on motion
- • GAAP profitability achieved (not just adjusted)
- • Adjusted margin tracking toward 6.9% target
Path to Less Favorable Assessment
- • Credit provision rate exceeding 0.80% of GMV
- • Consumer default rate above 0.70%
- • Heavy insider selling post-lock-up (50M+ shares)
- • FY2026 margin guidance missed materially
This analysis is for educational purposes only -- it is not a recommendation to buy or sell any security.
Public Sources Used (9 documents)
- • Annual Report (20-F) -- FY2025
- • Quarterly Report (6-K) -- Q3 2025
- • Quarterly Report (6-K) -- Q4 2025 Earnings
- • F-1/A Registration Statement (IPO Prospectus)
- • Q4 2025 Earnings Call Transcript
- • Q3 2025 Earnings Call Transcript
- • CFPB BNPL Market Report -- December 2025
- • Securities Class Action Filing Summary -- Hagens Berman
- • Senate Banking Committee Letter to Klarna -- November 2025
Full Analysis with Signal Breakdowns
Explore the complete 6-lens assessment including debate transcripts, evidence citations, and monitoring triggers.
View KLAR Analysis