Affirm underwrites each transaction; loss rates run 3.5 times Klarna's
Not just another BNPL app: Affirm vertically integrates credit and payment networks into what is essentially the internet's American Express—with individual transaction underwriting and negative customer acquisition costs.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Credit cards issue a total limit; Affirm prices each transaction individually
Credit cards approve a lump credit line and let customers manage their own spending; Affirm underwrites each transaction separately—rates, economic conditions, and purchase category all feed into a real-time model that adjusts who gets credit, how much, and at what rate. This approach already outperforms traditional consumer lenders like Synchrony despite Affirm's lower Prime customer concentration (about 50% versus Synchrony's 75%). Affirm releases a new transformer-architecture underwriting model every six to seven months, and loss outcomes improve in discrete jumps as more data feeds the system.
— Fraser ChristieAffirm tackles harder loans; Klarna pursues easier ones
Affirm is more than twice as large as Klarna in the U.S. market, but that difference reflects fundamentally different strategies from the start. Affirm focuses on larger amounts, interest-bearing loans with longer terms—like the 24-to-36 month financing for Peloton equipment. Longer terms mean more default risk: as Max puts it, consumers have 12 chances to choose not to repay. Klarna targets smaller, shorter loans like splitting a $100 pair of jeans across three or four payments. The loss profile reflects this: Affirm's monthly installment loans have roughly 3.5% loss rates while traditional short-term BNPL loans run about 1%.
— Fraser ChristieNegative CAC: Affirm turns profitable at point of sale
When a consumer registers on a merchant's website and completes an Affirm loan, Affirm acquires the customer while earning profit on that very loan—so customer acquisition cost on that channel is negative. The only exception: equity grants to Shopify and Amazon effectively count as hidden acquisition costs. This advantage compounds because e-commerce is steadily taking share from retail, and BNPL is steadily taking share from e-commerce. Affirm's growth flywheel is therefore more naturally embedded than most competitors'.
— Fraser ChristieFrom IPO day doubling to 90% crash in just over a year
Affirm's IPO priced at roughly $49 and closed its first day near $100. After one year, the stock had climbed to around $175. Then it plummeted to $9—a drawdown exceeding 90%. But reviewing that period reveals the damage was concentrated in one quarter: December 2022 saw growth collapse and profit margins crater, looking like catastrophe. Yet by the next quarter, margins had noticeably recovered, and the quarter after that, both growth and margins returned to health. The entire 'air pocket' effect persisted for just one to two quarters.
— Fraser ChristieAffirm spreads funding so widely it depends on no single source
About half the loans are packaged and sold off the balance sheet; the other half stay on but still draw third-party funding. Sources range across banks, insurers, and institutions; structures span securitized and unsecuritized, on- and off-balance-sheet, static and revolving, fixed and floating rate. Affirm is now applying for a bank charter to tap consumer deposits. The real near-term risk is mechanical: loan rates lock in at origination and never float. If the Fed moves rates up sharply, the newly originated high-rate loans take one to two quarters to make up 50% or 75% of the outstanding portfolio. This lag is where the 2022-2023 crisis actually struck.
— Fraser ChristieRising rates hurt Affirm's growth, not its loan loss rates
In credit cards, many customers carry revolving balances at floating rates—a rate shock from 15% to 25% triggers concentrated default waves. Affirm's mechanism is different: loan rates lock in at origination and never float. So the impact surfaces as a growth constraint, not a default crisis. When market rates rise and new loans must price at 25%, some customers who previously qualified at 15% no longer pass underwriting. Rather than originating the loan and hoping for repayment, Affirm simply doesn't originate to them—conserving growth rather than risking losses.
— Fraser ChristieAffirm's $100 billion GMV goal is a milestone, not the destination
U.S. credit card outstanding balances total roughly $1.3 trillion; Affirm carries about $20 billion on- and off-balance-sheet. Capital One alone processes 15 times Affirm's GMV. Management has set a target of $100 billion GMV with annual growth exceeding 25%, covering core U.S. operations, Affirm Card, and international business—but excluding Edge (bank white-label lending) and agentic commerce. If those two segments gain traction, the $100 billion target itself recedes further out.
— Fraser ChristieWalmart dropped Affirm, but customers stayed loyal to the brand
Walmart terminated Affirm's direct integration and swapped in Klarna's One Pay private-label BNPL solution. The economic impact on Affirm proved minimal: Walmart was never a high-margin merchant to begin with. More tellingly, the company disclosed that most customers who had used Affirm at Walmart simply switched to using Affirm Card instead—demonstrating brand loyalty that exceeded most investors' expectations.
— Fraser ChristieIn their own words · checked verbatim
the problem to be sold has to be hard, valuable and fun
Fraser Christie7:17
there's no hidden terms. There's no junk or gunk as Max Levshan would describe it. What you see is what you get
Fraser Christie31:11
Max's salary is less than $40,000 a year. He earns no cash bonus, but he owns a couple of billion dollars worth of stock
Fraser Christie48:41
Capital One processes over 15 times a firm's GMV each year. A firm is large in the sense it processes $50 billion of GMV and is integrated with some of the biggest merchants in America, but there's still a very long way to go
Fraser Christie1:01:09
The biggest businesses in the world, whether that be Google, Amazon, Apple, Shopify, have absolutely no interest in extending credit to their customers. They outsource it
Fraser Christie1:11:47
Figures
| Affirm's current GMV | ~$50 billion, annual growth exceeding 30% | 23:56 |
| Affirm's U.S. BNPL market share | ~50%, and still expanding | 29:11 |
| Affirm installment loan loss rate | ~3.5% (versus ~1% for traditional short-term BNPL) | 17:39 |
| IPO to present stock price movement | IPO price ~$49; first day close ~$100; one year later ~$175; fell to $9 (90%+ drawdown); currently ~$60-80 | 44:31 |
| Current market cap and valuation | ~$25 billion market cap; ~20x adjusted P/E | 44:31 |
| RLTC (gross profit) | ~$2 billion annualized | 44:31 |
| Management's medium-term target | GMV growth from $50 billion to $100 billion; annual growth exceeding 25% | 1:00:04 |
| U.S. credit card outstanding balances | ~$1.3 trillion total (versus Affirm's on- and off-balance-sheet loans of ~$20 billion) | 1:00:04 |
| P&L impact from 100 basis point rate increase | ~$90 million (versus ~$2.5 billion gross profit this year) | 55:57 |
| U.S. Top 250 merchants with Affirm integration | 75 | 36:17 |
Glossary
- MDR (merchant discount rate)
- Percentage fee merchants pay Affirm for accepting Affirm as a payment option.
- RLTC (revenue less transaction costs)
- Affirm's internal term for gross profit: revenue minus transaction costs.
- forward flow facility
- Funding mechanism where loans are continuously sold to third-party institutions after origination.
- agentic commerce
- Shopping where AI agents autonomously compare prices, place orders, and execute payments.
How to listen
Investors and founders focused on consumer finance, payment networks, or e-commerce infrastructure—anyone who wants to understand Affirm's business model and moat rather than just the BNPL label.
The opening Portrait ad (0:00-2:04) and details about Max's early immigration story (1:29-6:14).