What this covers
This episode of Business Breakdowns examines Affirm through an analogy to American Express. Fraser Christie opens with why Amex has always traded at a discount to Visa and Mastercard: it extends credit on top of its network, while they only run rails. Fraser carries the argument from there. His contrast is with cards, where about half of US cardholders carry a revolving balance averaging $7,000, while Affirm's loans are built to pay down to zero.
The middle of the conversation is mechanics. Loans are fixed-rate with 11 to 12 month terms, so a Fed hike mostly slows approvals and lags yield for a quarter or two, as in 2022 and 2023, instead of producing the credit-loss washout Fraser expects at Capital One and Synchrony. On unit economics, Pay in X earns roughly 1 to 3% of GMV, interest-bearing loans 4 to 6%, and the blend about 4%. Merchant-funded 0% loans, carrying high-single-digit MDRs, are pitched as customer-by-customer discounts. Usage has risen from about two to about seven transactions per user per year, and sales and marketing spend has barely moved since 2021.
The later turns are competitive and cultural. Affirm tackled large, long-term loans such as Peloton financing first and backfilled with pay-in-4, while Klarna went the other way. Walmart's move to OnePay is judged to have had minimal impact. International expansion is pulled by merchants such as Shopify and Costco. Fraser also explains the stock's swings as growth investors and bank investors each misreading it, and describes a flat-headcount engineering culture where 70 to 80 percent of deployed code is AI-generated.
Affirm should be valued as an internet-native payment network that happens to lend, not as a fragile subprime lender, because its transaction-level underwriting, merchant integrations, and founder-led tech culture create network effects and a credit P&L that behaves better than traditional card lenders in rate shocks.
- Affirm's network effects are visible in rising per-user usage and very low dependence on paid acquisition
- Transaction-by-transaction underwriting produces an interest-rate shock that slows growth for one or two quarters rather than causing a credit-loss blowout like revolving card books
- Affirm's customer base skews prime by dollars and older than assumed, so it is not primarily a subprime lender
Levchin's leadership and HVF incubator origins explain Affirm's stable team and survival through volatile stock history.
- The Affirm team has been stable over a long period, with both presidents tenured eight to ten years, which Avidan links to Levchin and the people leading the business as the reason the company survived its volatile stock history.
“a lot of that probably has to do with Max and the people leading the business”
- Affirm got its start within Levchin's HVF incubator, the acronym for hard, valuable, and fun.
“HVF was the name of his incubator, which is how Affirm got started too.”
AI shopping agents like Muse and Instinct create stock uncertainty for Affirm despite agentic commerce upside.
- Concerns that AI shopping agents such as Muse and Instinct could erode payments businesses weighed on Affirm's stock, creating uncertainty despite the upside potential of agentic commerce.
“the stock had dipped on the news of Muse and maybe even Instinct and Groq Bots”
Apple's Pay Later sunset over underwriting difficulty led Apple to partner with Affirm and Klarna.
- Apple's attempt to build Apple Pay Later was sunset because underwriting is very difficult, and Apple partnered with Affirm and Klarna instead.
“Apple tried to do this — I think it was Apple Pay Later — and they sunsetted that because the underwriting side is actually very, very difficult.”