Introduction
Affirm is a point-of-sale lender that lets shoppers split a purchase into a small number of installments instead of using a credit card. Founded in 2012 by Max Levchin — a co-founder of PayPal — along with Nathan Gettings, Jeffrey Kaditz, and Alex Rampell, Affirm was spun out of Levchin’s San Francisco innovation studio, HVF, and listed on the Nasdaq in January 2021.
What distinguishes Affirm from a conventional lender is where its underwriting model sits: directly inside the checkout flow of thousands of merchants, evaluating each transaction and returning an approval decision in seconds. For readers of Brel’s AI in finance coverage, Affirm is a useful case study in machine-learning credit decisions applied at consumer scale rather than behind the scenes at a bank.
What the company does
Affirm’s core product, Adaptive Checkout, offers a shopper biweekly or monthly pay-over-time plans at the point of sale, with annual percentage rates from 0% to 36% depending on the underwriting outcome and no late or hidden fees. The company also issues the Affirm Card, a Visa debit card that can toggle between debit and pay-over-time spending, and the Affirm Money Account, a savings account held at Cross River Bank. Loans made through Affirm’s platform are originated by its lending partners, listed at affirm.com/lenders.
Who it serves
Affirm works with merchants ranging from small online retailers to large platforms — it is integrated directly into Amazon.com and, since June 2023, into Amazon Pay for third-party merchants — as well as with individual consumers who apply for financing at checkout or through the Affirm app. According to figures reported in Affirm’s own financial disclosures and cited on its Wikipedia reference page, the company had close to 26 million users and roughly 478,000 merchants as of its fiscal year ended June 30, 2025. Merchants that adopted Affirm’s Adaptive Checkout in early access reported, per Affirm’s own announcement, meaningfully higher average order values compared with other payment methods offered at the same checkout.
Company background
Affirm was founded in 2012 and, per its own proxy filing, Max Levchin has served as its Chairman and Chief Executive Officer since the company’s founding. Affirm filed with the U.S. Securities and Exchange Commission in November 2020 to prepare for an initial public offering, reportedly delayed its filing that December, and ultimately listed on the Nasdaq Global Select Market under the ticker AFRM on January 13, 2021, raising approximately $1.2 billion. The company is headquartered in San Francisco.
Product and AI capabilities
Adaptive Checkout, launched in September 2021, uses what Affirm describes as its smart decision engine to combine a real-time underwriting decision with a dynamically personalized set of payment-plan options for each transaction, rather than offering a single fixed plan to every shopper. Merchants who adopted the product in early access reported, according to Affirm’s own announcement, an average 26% increase in cart conversion and a 22% lift in approvals compared with offering monthly payments alone. Beginning in its fiscal 2025 year, Affirm layered an AI-powered configuration tool, Adapt AI, on top of Adaptive Checkout, using machine learning to tune which financing offer a given consumer sees at checkout; on its fiscal fourth-quarter 2025 earnings call, the company said an early rollout of Adapt AI produced roughly a 5% increase in gross merchandise value for participating merchants.
Key developments
Affirm was founded in 2012 and spent its first years building direct merchant integrations before filing for an IPO in November 2020 and listing on the Nasdaq in January 2021. It launched Adaptive Checkout in September 2021, its first major push toward a machine-learning-personalized checkout experience, and extended that product to Amazon Pay merchants in June 2023. Most recently, Affirm has begun rolling out Adapt AI, an AI-configured version of Adaptive Checkout that automatically tunes financing offers per consumer rather than relying on manually tuned rules, with early results discussed on its fiscal 2025 earnings calls.
Why it matters
Because Affirm’s underwriting model runs at the moment of purchase, its decisions are felt directly by consumers and merchants at a scale most AI-in-finance products never reach — tens of millions of checkout events rather than a smaller pool of loan applications reviewed after the fact. That makes Affirm a useful, concrete example of how machine-learning credit decisions show up in ordinary online shopping, distinct from the underwriting software that banks license internally or the marketplace model used by lenders that route applications to partner institutions. Affirm’s no-late-fee structure also means its model has to balance approval rates against a revenue design that does not rely on penalty charges when a borrower falls behind, a different incentive structure than a typical revolving credit card.
Sector context
Within Brel’s AI in finance coverage, Affirm sits alongside other AI-driven consumer credit companies such as Upstart, though the two differ in structure: Affirm holds its own merchant relationships and underwrites at checkout, while Upstart licenses its models to bank and credit union partners that originate the loans. Affirm’s underwriting also invites comparison with software vendors such as Zest AI that sell explainable credit models to lenders rather than making consumer-facing lending decisions themselves.
Sources and references
This profile draws on Affirm’s own investor relations announcements, its fiscal 2025 earnings call transcript, and independent reference sources.
- Affirm Investor Relations — “Affirm launches Adaptive Checkout™” (2021)
- Affirm Investor Relations — “Amazon Pay adds Affirm…” (2023)
- The Motley Fool — Affirm (AFRM) Q4 2025 Earnings Call Transcript
- Wikipedia — “Affirm Holdings”