Scottish Mortgage is Baillie Gifford's flagship investment trust (founded 1909, LSE ticker SMT), known for its maximalist growth style — long-term stakes in Tesla, Amazon and ASML plus bold allocations to private companies like SpaceX and ByteDance. It is the UK retail investor's flagship vehicle for global disruptive growth.
This interview profiles Affirm, a 'buy now, pay later' company that lets shoppers split purchases into installments instead of using credit cards. Its distinctive promise is never charging late fees or hidden costs. That leaves no room for sloppy lending, so it checks each customer's finances in real time before approving every purchase. For ordinary investors, the lesson is that growing a loan business is easy, but growing it without big losses is hard. Affirm is betting that transparent pricing and AI shopping agents will undermine credit-card tricks buried in fine print. Worth reading for a clear look at what makes some lenders sustainable and why customer concentration is a risk.
This report presents Scottish Mortgage's investment analysis of Affirm. Affirm is a buy-now-pay-later (BNPL) platform founded by Max Levchin, positioned as an alternative to credit cards. It promises transparent pricing with no late fees or hidden charges, and offers 0% interest financing on many pu
Max Levchin is the co-founder and CEO of Affirm, and also a co-founder and the first CTO of PayPal. This conversation centers on his motivations for founding Affirm, its business model, and technological moats, with Scottish Mortgage investment manager Tom Slater adding the investment thesis and risks from a shareholder perspective. The most consequential takeaway of the entire episode: Tom Slater believes that in consumer lending, growth is easy — what is hard is growing without losing money, and that is precisely the dividing line between Affirm and most of its competitors.
Max Levchin 断言：Affirm 从不、也永远不会收取滞纳金或任何隐藏费用，因为只有这样才能建立真正的消费者信任，而这要求承保精准到没有容错空间。
Levchin's own credit card trauma was the starting point of his entrepreneurial journey. In 1993, while at the University of Illinois, he obtained a "0% APR" credit card, only to discover 90 days later in the fine print: failing to pay the minimum amount due would trigger retroactive interest of roughly 29%. After PayPal's IPO, he went to buy a car and was denied financing by the dealership for having a poor credit record. He concluded that the U.S. credit industry "invented all kinds of fees just to squeeze out a little more money" — late fees, reminder fees, deferral fees — all categorically rejected by Affirm.
Whether this no-fee model can be profitable hinges on underwriting. Levchin says:
> "The only way you can make a profitable business out of that is you have to be exceedingly good at underwriting." In other words: if you do not make money from fees, the only way is to perfect underwriting.
The appeal to merchants likewise comes from "true zero interest." In his early days, Levchin personally bet merchants: Affirm would fund a weekend 0% promotion at its own expense, and if sales did not visibly improve, it would never bother them again. He won every single bet. Today, when merchants onboard with Affirm, they typically expect at least a 20% increase in sales; compared with a 10% discount promotion, true 0% installment plans stimulate consumption more effectively without disrupting merchants' pricing architecture — frequent discounting only teaches consumers to wait for the next sale, whereas 0% installment plans do not erode pricing.
Tom Slater argues: growing a consumer lending business is easy, but growing it without losing substantial money is extraordinarily hard — Affirm's real-time, transaction-by-transaction underwriting and 15 years of accumulated data form a moat competitors find difficult to replicate.
Slater concluded in the final investment discussion:
> "It's really easy to grow a consumer lending business. It's really difficult to grow a consumer lending business, without losing a lot of money." In other words: it is easy to grow a consumer lending business; the hard part is growing it without losing a lot of money.
The traditional credit card model is "one approval, valid indefinitely," and does not reassess whether a consumer can still afford a given purchase for two or three years after the card is issued. Affirm, by contrast, underwrites every single transaction in real time: at each purchase, it immediately pulls the consumer-authorized financial data, and a machine-learning model decides whether to approve and at what interest rate. Levchin offers an example: if the model detects that an applicant's cash flow is already overextended, Affirm will outright decline the loan, or suggest paying half upfront and financing the other half, while explaining the reason.
This creates a data network effect: the broader the coverage — the U.S., Canada, the U.K., and the new markets it plans to enter — the more accurate the underwriting model; the more accurate the model, the lower the interest rates consumers receive and the lower the cost of merchants' 0% promotions, in turn attracting more merchants and users. Levchin calls this a "powerful data moat." Slater adds another layer of defense: extremely high user net promoter scores — "people who use Affirm genuinely value it" — and the fact that merchants who tie their brand to Affirm must trust the quality of its service, a trust that cannot be quickly copied. The competitive landscape is likewise confirming the first-mover advantage: since Affirm pioneered the no-late-fee model, more than half of U.S. BNPL transaction volume no longer charges late fees.
On risk, Slater acknowledges the business is concentrated among a few large clients: although Affirm has more than 300,000 merchant partners, a substantial portion of business comes from a handful of major platforms, including Amazon. His logic: as long as Affirm consistently delivers merchants "incremental sales that would not have otherwise happened," it is not only a financial partner but also a marketing channel — value that is hard to replace.
Max Levchin judges: AI shopping agents will diligently read all the "fine print," and the era of credit institutions earning hidden profits through convoluted terms is coming to an end — a structural tailwind for Affirm.
Levchin's view on AI shopping is clearly layered: some shopping has an entertainment quality, and consumers will not hand it entirely to robots; but the final step — "paying for an item you want" — people will confidently entrust to an AI agent. He summarizes:
> "Robots are very happy to read all the fine print." In other words: robots are very happy to read all the fine print.
This means consumers will no longer be deceived by "0% with an asterisk." Affirm is already positioning for this through partnerships with Shopify and Stripe: the former has selected Affirm as its exclusive BNPL partner, while the latter — also a Scottish Mortgage holding — is building the payment infrastructure for AI shopping agents. Levchin notes both companies are driven by engineering culture, speaking a common language that enables joint development.
On market opportunity, the data chain is clear: Affirm is projected to facilitate more than $48bn in merchandise and service transactions this fiscal year, yet that accounts for less than 1% of total U.S. retail sales and roughly 2% of e-commerce — against outstanding U.S. credit card balances of approximately $1.2–1.3tn. Levchin uses Australia as a reference point: BNPL there already accounts for more than 20% of retail sales, while the U.S. remains a "laggard." Slater outlines four growth paths: raising brand awareness, getting merchants to use Affirm as a customer acquisition tool, dramatically expanding the range of merchant locations where Affirm is accepted through the Affirm Card, and international expansion into markets such as the U.K. He also stresses that, with proper execution, the vast pool of credit card balances is the long-term target.
| Company | Guest Stance | Key Data |
|---|---|---|
| Affirm | Core holding; bullish on its transparent pricing, real-time risk control, and data moat | Forecast GMV $48bn+ for the current fiscal year; in roughly 30% of transactions, merchants bear all interest; over half of U.S. BNPL transactions now have no late fees |
| PayPal | Background mention; Levchin co-founded it and served as CTO | Went public in 2002; Levchin says his four years at PayPal were the source of all his business education |
| X.com | Background mention; merged with Confinity to form PayPal | — |
| Confinity | Background mention; founded by Levchin and Peter Thiel | — |
| Tesla | Not stated (only mentioned that Scottish Mortgage holds it long-term) | — |
| SpaceX | Not stated (only mentioned that Scottish Mortgage holds it long-term) | — |
| Yelp | Background mention; Levchin served as chairman | — |
| Slide | Background mention; founded by Levchin and then sold to Google | Acquired by Google; by Levchin's own assessment, revenue fell short of expectations |
| Background mention; acquired Slide | — | |
| Andreessen Horowitz | Background mention; early investor in Affirm | — |
| Khosla Ventures | Background mention; early investor in Affirm | — |
| Visa | Partner; co-created the Visa Flex standard with Affirm | Enables Affirm Card to be used across the Visa network |
| Shopify | Partner; shared engineering culture | Chose Affirm as its exclusive BNPL partner |
| Stripe | Partner; also a holding of Scottish Mortgage | Provides infrastructure for AI shopping agents to trigger Affirm installments |
| Amazon | One of the major customers; also a holding of Scottish Mortgage | TS acknowledged business concentration risk without giving a specific percentage |
| Wayfair | Mentioned; a merchant where users report Affirm is available | — |
| Best Buy | Mentioned; a merchant where users hope Affirm will be integrated | — |
| Gopuff | Preview mention; guest on the next interview (neutral) | Promises delivery in as fast as 15 minutes |
1. Tom Slater: Consumer lending is easy to grow, but hard to grow without losing money — management restraint and risk-control authority are the key variables when investing in this industry.
2. Max Levchin: AI agents will read all the fine print, and the era of credit providers earning hidden profits through complex terms is coming to an end; Affirm has already positioned itself at the forefront of this trend.