Affirm: The Internet’s American Express - [Business Breakdowns, EP.249] — Summary & Takeaways · Colossus (Invest Like the Best / Business Breakdowns)

Affirm: The Internet’s American Express - [Business Breakdowns, EP.249]

At a Glance

In this episode, host Fraser Christie (Founder and CIO of Sandstone Global Investment) analyzes Affirm — a "buy now, pay later" (BNPL) company founded by PayPal co-founder Max Levchin. Fraser argues that Affirm is fundamentally more akin to an internet-era American Express (payment network + vertically integrated credit) than a traditional lender. Core theses include: transaction-level risk assessment outperforms credit cards' general credit limit model; merchant relationships generate strong network effects; negative customer acquisition cost (negative CAC); and growth potential from AI and agentic commerce. The episode also discusses the company's resilience during the 2022-2023 rate hiking cycle, competition with Klarna and Afterpay, the expansion of Affirm Card, and key investment risks.

I. Max Levchin: The Founder's Story from PayPal to Affirm

1. Background and Entrepreneurial Motivation

Fraser notes that Max Levchin was born into a family of physicists in Kyiv, Ukraine. Following the Chernobyl nuclear disaster, he left Kyiv with his family at age 10, lived with his grandparents in Crimea for five years, and then moved to Chicago shortly before the dissolution of the Soviet Union. When applying to MIT, he mistakenly wrote the school's name as "MTI" and ultimately attended the University of Illinois.

During his college years, he originally aspired to become a professor of astrophysics but was drawn to the entrepreneurial wave. The seed for Affirm was planted during this period: having grown up in the Soviet Union, he had no concept of credit cards or credit systems. He used credit cards to finance his startup ventures, frequently missed payments, and severely damaged his personal credit record.

2. How Personal Experience Shaped the Company's Mission

Even after Levchin had become highly successful following PayPal's IPO, his poor credit record continued to haunt him. Fraser recounts two key stories:

Car Purchase Rejected: After the PayPal IPO, Levchin flew to Los Angeles to buy a convertible for his girlfriend (now wife), but was denied a loan by the dealership due to his low credit score. He had to wait for a wire transfer to clear before he could take the car.
Friends' Prank: For years, friends would pretend his credit card was declined at restaurants as a way to prank him.

These experiences drove Levchin to found Affirm a decade after leaving PayPal, with a mission that remains unchanged to this day: to provide honest financial products that improve people's lives — no late fees, no revolving compound interest, no deferred interest, and no hidden terms. When he explained this concept to industry insiders, they thought he was "completely insane," as a significant portion of the industry's profits came from precisely these mechanisms.

3. Leadership Style and the HVF Framework

Technology-Driven: As PayPal's CTO, Levchin was an early pioneer in fraud detection and cryptography. Affirm's culture is highly technical, and he attracted math talent who might otherwise have entered academia or hedge funds.
The HVF Framework: Levchin's startup incubator is called HVF, which stands for Hard, Valuable, and Fun. He tends to tackle difficult problems — for example, offering consumers large, long-term credit loans without "safety net" mechanisms like late fees and compound interest.
Discipline: Although the company was generous with equity incentives in its early days, it has achieved a flat headcount for several years, flat G&A expenses for five years, and flat stock-based compensation for six years. Fraser believes this reflects the ability to "make tough decisions."

II. The Complete Flow of an Affirm Transaction

1. Consumer Side

Fraser positions Affirm as a payment network built for the internet and e-commerce, drawing an analogy to American Express. Using the example of a consumer purchasing an item worth approximately $250 on an e-commerce site (e.g., Amazon, Target):

1. The consumer sees Affirm financing options at checkout (e.g., 0% APR for 3 months, 20% APR for 6 months, 25% APR for 12 months) — at this stage, this is merely a theoretical quote, and Affirm has not yet pulled the consumer's credit record.

2. The consumer clicks the Affirm button and provides their phone number, name, date of birth, and the last four digits of their Social Security number.

3. Affirm matches the transaction data (merchant, basket amount, SKU details) and performs real-time underwriting.

4. If approved, the consumer selects the most suitable plan and completes the checkout.

5. The consumer repays according to the agreed schedule — the interest rate is fixed, the repayment plan is fixed, and the balance goes to zero as scheduled. This is a key distinction from credit cards.

2. Fund Flow

Affirm pays the merchant the product price minus the merchant discount rate (MDR). For example, for a $250 item with an MDR of $5, Affirm retains $5 and pays the merchant $245.
For interest-bearing loans, Affirm also profits from the interest paid by the consumer.
Funding side: Approximately half of the loans are sold to third parties (off-balance-sheet), while the other half remain on the balance sheet but utilize third-party funding vehicles.

3. Network Effects & Customer Acquisition Efficiency

The more merchants that integrate Affirm checkout, the more opportunities consumers have to see the offer, and the more customers Affirm acquires — similar to American Express's network effect.
Approximately 95% of transactions come from existing customers, and the average usage frequency per user has increased from 2 times per year at the time of the IPO to 7 times per year.
Sales and marketing cash expenditures have remained nearly flat since 2021, as the majority of customers are acquired organically through offers displayed on merchant websites.

III. Transaction-Level Risk Control: The Core AI-Driven Advantage

1. Fundamental Differences from Credit Cards

Fraser points out the fundamental differences between Affirm and credit cards:

Dimension	Credit Cards	Affirm
Credit Approach	General credit limit (e.g., several thousand dollars)	Per-transaction underwriting
Interest Rate	Variable rate, compound interest	Fixed rate, with a fixed repayment plan
Balance Handling	Can revolve indefinitely	Designed to reach zero
Risk Control Updates	Low frequency of limit adjustments	Each transaction assessed independently; rate, amount, and term dynamically adjusted

2. AI and Model Iteration

Affirm uses a Transformer-based underwriting model, releasing a new version approximately every 6-7 months, with each iteration delivering "materially better underwriting outcomes."
As more data is fed into the model for training, performance continues to improve.
Comparative data: Approximately 75% of Synchrony's customers are prime, while about 50% of Affirm's customers are prime — yet Affirm's underwriting outcomes (e.g., default rates) are actually better.

3. Common Misconceptions About BNPL

Fraser emphasizes that BNPL does not serve only subprime customers:

Affirm's average customer is primarily prime (by loan volume, not by customer count).
The customer base is older (predominantly aged 30+), not just young people.
The product is not "subprime lending" but serves a broader population — including consumers who carry revolving credit card balances (about half of U.S. credit card holders carry revolving balances) but have not received the best products.

IV. Business Model & Unit Economics

1. Three Loan Types & Revenue Structure

Fraser introduces the three loan types and their revenue structure:

Type	% of GMV	Primary Revenue Source	Gross Margin (% of GMV)
Interest-Bearing Loans	~70%	Consumer interest income	4%-6%
0% APR Loans	~15%	Merchant discount rate (MDR, higher)	1%-3%
Buy Now, Pay Later (Pay in X)	~15%	Merchant discount rate (MDR, close to credit card rates)	1%-3%

Interest-Bearing Loans: MDR is very low; primary revenue comes from consumer interest.
0% APR Loans: Interest is subsidized by merchants; MDR is higher (can reach high single-digit percentages).
Pay in X: Short-term, small-ticket loans; MDR is gradually approaching credit card rates.

2. Overall Profitability

Blended gross margin is approximately 4% (as a % of GMV), industry-leading.
Gross margin already deducts all credit losses and funding costs.
Although interest-bearing loans have a higher loss rate (~3.5%), their gross margin remains significantly higher than short-term loans (loss rate ~1%).

3. Affirm Card Economics

Approximately 20% of GMV comes from the Affirm Card, with annual growth exceeding 100%.
Card transactions have a higher proportion of interest-bearing loans, making unit economics even better than the core business.
The card product allows Affirm to reach offline payments and online merchants without direct integration.

V. Why Affirm Is "the Internet Version of American Express"

1. Payment Network vs. Pure Lender

Fraser distinguishes three types of payment networks:

Visa/MasterCard: Pure payment rails, with credit provided by issuing banks (separated model).
American Express: Vertically integrated — payment network + credit.
Affirm: Also vertically integrated — directly integrated with merchants (online) + credit.

Evidence of network effects: Affirm processes over $50 billion in annual GMV, growing at over 30%, and its growth rate is faster than it was several years ago (despite having scaled 2-3x). This aligns with the network effect characteristic of "the larger the scale, the faster the growth."

2. Negative Customer Acquisition Cost (Negative CAC)

When a consumer sees an offer on a merchant's website and registers, Affirm has already profited from that loan — customer acquisition cost is negative.
The only exceptions are large partners like Shopify and Amazon, to which Affirm previously issued equity as consideration for the partnership.

3. Dual Industry Tailwinds

Tailwind	Current Status	Potential
E-commerce as share of retail	Approximately 20% in the U.S.	Continued growth (total retail $7-8 trillion)
BNPL as share of e-commerce	Approximately 6% in the U.S.	Could increase to 15%-25% (referencing Australia, Sweden)
Affirm's U.S. market share	Approximately 50% and growing	Growing faster than Klarna and Afterpay

VI. The Opportunity in Agentic Commerce

1. Why Affirm Is Better Positioned in the AI Agent Era

Fraser believes that agentic commerce will only accelerate the penetration of e-commerce and BNPL. AI agents excel at comparing and selecting optimal credit options, and Affirm’s products are naturally suited for this:

Directly integrated at checkout, offering personalized quotes.
No hidden fees (no late fees, no compound interest), with transparent pricing.
Merchant-subsidized 0% APR loans — credit cards cannot compete with this.
E-commerce native: AI agents can automatically fill in information such as name, phone number, and the last four digits of the Social Security number, operating more smoothly than a human.

2. Partnership Model with ShopPay

Affirm already provides credit for ShopPay through a white-label model (ShopPay Installments is essentially a white-label version of Affirm). Fraser expects that when AI agents build wallets in the future, Affirm will similarly provide credit functionality and share in the economic benefits.

3. Misunderstanding Behind the Stock Price Decline

Fraser points out that market concerns about AI agents have caused Affirm’s stock price to decline, but this is a misjudgment — AI agents will only make Affirm’s transparent pricing and merchant-subsidized loans more attractive, rather than weakening its position.

7. Competitive Landscape: Affirm vs. Klarna vs. Afterpay

1. Core Differences

Fraser points out the core differences between Affirm, Klarna, and Afterpay:

Dimension	Affirm	Klarna	Afterpay
Origin	United States (2012)	Sweden (2006)	Australia
Core Product	Interest-bearing long-term loans (high ticket size)	Pay in X short-term loans	Pay in X short-term loans
U.S. Market Share	~50%, growing	Smaller than Affirm	Smaller than Affirm
Loss Rate (Installment Plans)	~3.5%	~1% (short-term loans)	~1%
Product Breadth	Full category (from Pay in X to 24-36 month loans)	Primarily short-term, expanding into long-term	Primarily short-term
Late Fees/Penalties	None	Still has related revenue	Yes

2. Why Affirm Ultimately Wins

Fraser believes that Affirm chose the harder problem (large-ticket, long-term, interest-bearing loans) and then "worked backward" to solve the easier short-term loan problem. In contrast, Klarna started with easy short-term loans and is now trying to solve the harder problem. This difference in trajectory has led to Affirm being larger, growing faster, and having broader merchant coverage in the U.S.

3. Strategic Significance of International Business

Affirm's international business is currently small, but it is being "pulled" into markets like Canada, the UK, and Australia by U.S. merchants (e.g., Shopify, Costco).
Strategic importance: Global merchants (e.g., Airbnb, eBay) tend to prefer a single global button. If Affirm does not expand internationally, it may lose out to Klarna in competing for these merchants.
In the long term, international expansion is a necessary condition for maintaining Affirm's core U.S. competitiveness.

VIII. Merchant Adoption & Customer Retention

1. Why Merchants Have Not Fully Adopted

Fraser points out the reasons why merchants have not fully adopted Affirm:

Only 75 of the top 250 U.S. merchants have integrated Affirm.
Many merchants view credit card acceptance as a cost center, not a growth lever.
The MDR for 0% APR loans can be as high as high single digits, which some merchants consider too high.
Affirm's sales strategy: Position 0% APR loans as a precision discount tool—merchants can subsidize rates for specific customers rather than offering storewide discounts (e.g., 20% off), thereby protecting profit margins.

2. Walmart Case: Customer Retention Exceeds Expectations

In 2023, Walmart terminated its partnership with Affirm and switched to its own "One Pay" (powered by Klarna). However, the impact was limited:

Walmart was not a high-margin merchant for Affirm (Walmart has strong bargaining power).
Most former Walmart customers continued using the Affirm Card for purchases at Walmart—customer loyalty was higher than expected.

3. Customer Retention Data

After joining, the average usage frequency increased from 2 times per year to 7 times per year.
Customer churn is extremely low.

IX. Resilience During the Rate-Hiking Cycle: Funding and Interest Rate Risk Management

1. Diversified Funding Strategy

Fraser introduces Affirm's diversified, high-redundancy funding base:

Funding Type	Share	Description
Off-Balance Sheet (Forward Flow + Static ABS)	~50%	Loans sold to third parties
On-Balance Sheet (Warehouse + Revolving ABS)	~50%	Use of third-party financing vehicles
Future: Bank Deposits (applying for bank charter)	New	Further diversification

2. The Mechanism of Rate Hike Impact

Unlike credit cards, Affirm’s loans carry fixed interest rates (average term of 11–12 months). The impact of rate hikes is a short-term "air pocket" :

After a rate hike, the interest rate on newly originated loans rises, but the rate on existing loans remains unchanged.
Due to the fast turnover of the loan portfolio: approximately 50% of loans can be repriced to higher rates within one quarter, and approximately 75% within two quarters.
The impact lasts only 1–2 quarters.
Recent data shows: a 100 basis point rise in interest rates impacts the income statement by approximately $90 million (the company’s annual gross profit is approximately $2.5 billion).

3. Impact of Rate Hikes on Growth

Rate hikes affect not only profitability but also suppress growth: when Affirm needs to charge higher rates on new loans, some consumers who would previously have qualified no longer meet underwriting standards, leading to a decline in loan origination volume. However, credit losses do not spike significantly (because existing loan rates remain unchanged, and consumers are not squeezed by floating rates).

Fraser concluded: 2022–2023 represented the "largest and fastest rate-hiking cycle in 40 years," yet Affirm experienced only one quarter of significant pressure before recovering quickly. Rate hikes are not an existential risk.

10. Growth Path: From $50 Billion to $100 Billion GMV

1. Management Targets

Affirm management set medium-term targets at a recent investor forum: doubling GMV from approximately $50 billion to $100 billion, with an annual growth rate exceeding 25%. Management clearly distinguished between growth sources that are "already included in the target" and those that are "incremental."

2. Three Growth Pillars (Included in Target)

Pillar	Current Status	Growth Logic
Core U.S. Business	50% market share, 75/250 large merchants	E-commerce penetration increase (20% → higher) + BNPL penetration increase (6% → 15-25%) + Merchant coverage expansion
Affirm Card	20% of GMV, growth rate >100%	Reaches offline and non-integrated merchants; unit economics comparable to or better than core business
International Business	Early stage	"Pulled" into new markets by U.S. merchants; strategically prevents global merchants from defecting to Klarna

3. Incremental Growth Sources (Not Included in Target)

Edge (White Label): Provides white-label credit services to U.S. regional banks. Consumers can enable a "buy now, pay later" feature on their existing debit cards, with Affirm providing underwriting and servicing.
Agentic Commerce: AI agents automatically select the optimal credit solution.

4. Overall TAM Comparison

Metric	Scale
U.S. Credit Card Outstanding Balances	~$1.3 trillion
Affirm Total Loan Balances (On-Balance Sheet + Off-Balance Sheet)	~$20 billion
Visa/MasterCard/Amex Annual Credit Card Transaction Volume	~$13 trillion
Affirm Annual GMV	~$50 billion
Capital One Annual GMV	More than 15 times Affirm's

XI. Key Risks and Investor Concerns

1. Common Investor Concerns

Fraser points out common investor concerns and his responses:

Concern	Fraser's Response
Credit business is too complex	Requires understanding, but Affirm's mechanism differs from credit cards and carries lower risk
Only serves subprime customers	Misunderstanding — most customers are prime, just not receiving the best products
Intense competition, weak moat	Network effects + transaction-level risk control + technical talent + no hidden fees = structural advantage
Merchant concentration risk	Amazon exceeds 20%, but long-term agreements + customer loyalty (Walmart case) mitigate the risk
Interest rate risk	Only a 1-2 quarter "air pocket," not a survival risk

2. Valuation Divergence

Fraser points out that the market suffers from an "identity crisis" regarding Affirm's valuation:

Growth investors: Buy during tailwinds but lack understanding of credit, selling off at the first sign of risk.
Bank/credit investors: Do not value traits such as technology-native nature or founder leadership.
Similar to the debate over Amazon in its early days — "retailer or tech company?" Affirm is both.

3. Biggest Risk: Market Sentiment During the Credit Cycle

Fraser believes the biggest short-term risk is market panic over credit — whenever news about private credit issues emerges, Affirm's stock price declines. However, history has proven that the company is capable of "self-healing."

XII. Lessons for Founders and Investors

1. Lessons for Founders

Fraser summarizes the HVF Framework: Do hard, valuable, and fun things. Max Levchin chose to solve the difficult problem of "large consumer loans without late fees or compound interest," ultimately building a formidable competitive moat.

2. Lessons for Investors

Fraser summarizes: Look for companies that are technology-native but operate in non-tech industries. Such enterprises combine:

Top-tier engineering talent (technological advantage)
Real-world operational complexity (industry barriers)

Affirm possesses both underwriting talent and engineering talent — a cross-disciplinary advantage that "even Claude Code cannot replicate." The world's largest companies (Google, Amazon, Apple, Shopify) have no intention of providing credit themselves — they outsource this difficult problem, and Affirm is the one that takes it on.

Mentioned Positions

Position	Guest's View or Explicit Action	Key Basis & Data
Affirm	Long-term bullish, described as "the internet version of American Express"	Annual GMV exceeds $50 billion, growth rate over 30%; gross margin ~4% (industry-leading); network effects drive negative CAC; transaction-level risk control superior to credit cards
Klarna	Main competitor, but Affirm is already ahead in the U.S.	U.S. market share ~50% and expanding; Klarna focuses mainly on short-term loans, while Affirm offers a more comprehensive product suite
Afterpay	Australian pioneer, but U.S. market share lags behind Affirm	Primarily short-term loans; penetration rate in Australia ~15-25%
Walmart	Has terminated direct partnership with Affirm, but impact is limited	Not a high-margin merchant; most customers still spend at Walmart via the Affirm Card
Amazon	Largest partner (over 20% of GMV), long-term agreement	Signed in 2020-2021; partnership is stable
Shopify	Important partner; ShopPay Installments is a white-label product for Affirm	Equity partnership; a driver of international expansion
Costco	Partner, facilitating international expansion	Bringing Affirm into Canada, the UK, and Australia
Capital One	Comparison target	Capital One's annual GMV is more than 15 times that of Affirm
Synchrony	Comparison target	~75% of Synchrony's customers are prime borrowers, but its underwriting results are inferior to Affirm's (the latter has only 50% prime customers)
Peloton	Historically important client (largest source of business at IPO)	24-36 month loans; Peloton's business collapse once hit Affirm's stock price

Memorable Insights & Thought Processes

1. "Affirm is not a lender, but an internet-era American Express" (Fraser Christie)

Vertically integrated payment network + credit, generating network effects; negative CAC; 95% of transactions come from existing customers; growth accelerates as scale increases.

2. "Transaction-level risk control vs. general credit limits: a structural advantage" (Fraser Christie)

Credit cards: one-time credit approval, consumers decide how to use it. Affirm: each transaction is independently assessed, with dynamically adjusted interest rates, amounts, and terms. Even with a lower credit score customer base, underwriting outcomes are still better.

3. "The HVF framework: doing hard, valuable, and fun things" (Max Levchin)

Choosing to solve the difficult problem of "large consumer loans with no late fees and no compound interest" ultimately built a moat that competitors find hard to replicate.

4. "Rate hikes are not an existential risk, but a 1-2 quarter air pocket" (Fraser Christie)

The loan portfolio turns over quickly (50% resets in a quarter); a 100 basis point rate increase only impacts approximately $90 million in profit (against ~$2.5 billion in annual gross profit).

5. "The Walmart case proves customer loyalty is higher than expected" (Fraser Christie)

Even when a merchant terminates a direct partnership, most customers continue spending at that merchant via the Affirm Card — the network effect transcends a single merchant relationship.

6. "AI agents are not a threat, but an accelerator" (Fraser Christie)

Transparent pricing, merchant-subsidized 0% APR loans, and a native e-commerce experience — AI agents will only make Affirm's credit solutions more attractive.

7. "Look for companies that are technology-native but operate in non-tech industries" (Fraser Christie)

Such companies combine engineering talent with industry moats, creating a competitive advantage that "Claude Code cannot replicate." The world's largest companies have no intention of providing credit themselves — this is precisely Affirm's opportunity.

Affirm: The Internet’s American Express - [Business Breakdowns, EP.249]

In plain words