Affirm Holdings Q3 FY26 earnings call AFRM

Consistent with prior quarters, the call opened with only a short statement from CEO Max Levchin ('Fiscal Q3 was another one for the record books') before moving straight into Q&A; COO Michael Linford and CFO Rob O'Hare fielded most detail. Management repeatedly declined to 'front-run' its 2026 Affirm Investor Forum, scheduled for Tuesday, May 12, 2:00–5:00 p.m. ET (public livecast plus replay). Topics explicitly held back for that event included international market specifics, the Affirm Card and banking roadmap, engagement/transactions-per-active decomposition, Adaptive Checkout metrics, and details behind the 'Big Nothing' promotional event. Analysts should treat the light quantitative disclosure here as intentional, with the forum as the data-rich follow-up.

Management reported stable delinquencies and no deterioration among the Affirm consumer base (explicitly not a comment on the broader U.S. consumer). The QoQ rise in the allowance rate was attributed to seasonality — the sequential downtick from holiday volumes in Q2 to a lower Q3 base — and to elevated prepayments during tax season, which reduce the loan balance and mechanically lift delinquencies off a smaller base while being framed as a positive credit signal. Funding conditions were described as 'exceptionally constructive': three securitizations executed year-to-date (two revolving deals in the quarter plus a static deal priced but not yet closed), heavy oversubscription, and sustained spread tightening contributing to funding costs down roughly 125 bps YoY. Forward-flow demand remained strong, with partners 'clamoring for a bigger allocation.'

The Card base roughly doubled to 4.4 million cardholders, adding about 700,000 in the quarter (down from ~900,000 the prior quarter), at roughly a 17% attach rate and about 20% of actives. Management reiterated the Card is the fastest-growing and most profitable product, now in the billions of dollars of volume, and stressed growth is 'natural' rather than fueled by performance marketing — Affirm does not pay to acquire users or mail cards, positioning the Card as the best mechanism to re-engage known, previously-transacted consumers. Cardholders skew slightly higher credit quality (by design), transact more frequently, spend ~40% more than the average Affirm consumer (rising over time^⏳), and show broader, more even category usage than casual users who transact ~4–6 times per year. An analyst noted a medium-term target of $10B GMV and ~7.5M active Card users; management did not restate or confirm those figures on this call.

Pay in X (Pay in 4) was called the fastest-growing segment, driven partly by one very large program moving to an evergreen 0% Pay in 4 offer, with most Pay in 4/X volume still originating from the Shopify program. 0% loans carry slightly lower RLTC margin but also lower credit costs; management's largest programs all now use 0%, and Affirm is leaning into 0% within the Card. The lower loan-loss on purchase commitment as a percentage was attributed to 0% mix and term length (the discount rate applied to 0% loans), not to changed vendor economics. On Adaptive Checkout, management said it is now effectively sold together with Boost as a single offering ('Affirm Checkout'), is increasingly understood by merchants, and is trending toward an AI-optimized, always-on default — with specifics reserved for the forum.

Management highlighted a notable ramp in agentic-written code (roughly double the prior request volume), calling it 'unequivocally accretive to the bottom line' and citing a recent hackathon where the product team delivered dozens of shippable features. Rob quantified the associated developer-tool spend at low single-digit millions per quarter, continuing into Q4, and immaterial to the P&L. Max flagged unique-to-Affirm checks and balances (no 'AI makes mistakes' fine print acceptable in underwriting), said the company believes it can '10x this productivity further,' and explicitly ruled out AI-related layoffs, noting revenue per employee already sits in 'NVIDIA territory' and headcount has grown little.

Asked whether Q3's GMV strength is sustainable, management said there was nothing unusual or unnatural about the quarter and that Affirm remains 'really, really small' relative to total U.S. payment and e-commerce volume, leaving ample room to take share. Competitive differentiation was framed around 15 years of brand trust (no late fees, ever), deep and sophisticated AI-driven underwriting that competitors have found harder than expected, capital-markets counterparties that trust Affirm's asset, and a compounding network effect: more merchants increase consumer awareness and sign-ups, which drives Card adoption and higher transactions per active (growing above 20%), with 90%+ of transactions from returning users. A bank-charter application is progressing with regulators on an uncertain timeline, kept separate from the consumer product roadmap.