What Affirm actually is
Affirm is a lender that sells itself at checkout. When a shopper reaches the payment step at a merchant that has integrated it, Affirm offers a loan for that specific purchase instead of a card charge, underwrites the shopper on the spot, pays the merchant, and then collects from the consumer in instalments. It listed on Nasdaq under AFRM in January 2021, was founded in 2012 by Max Levchin, Nathan Gettings, Jeffrey Kaditz and Alex Rampell, and facilitated US$36.7 billion of gross merchandise volume in the fiscal year ended 30 June 2025 across approximately 377,000 active merchants.
Three product shapes account for that volume. Short-term interest-free plans of up to four instalments were 14% of FY2025 GMV. Interest-free monthly instalment plans, where the merchant subsidises the consumer's interest, were 13%. Interest-bearing monthly instalment loans were 72% — the largest share by a wide margin, which is the fact that separates Affirm from most of the buy now, pay later category.
Two design choices distinguish Affirm's consumer product. Interest is simple and fixed rather than compounding, so the total shown at checkout is the total owed. And Affirm charges no late fees at all — its 10-K states it has charged US$0 in late fees since founding and that consumers do not pay more than what they agreed to at checkout even if a payment is missed. Whatever else is true about BNPL, that structure is more defensible under regulatory scrutiny than a revolving balance with penalty pricing.
How an Affirm transaction actually works
The chain behind a single approval is longer than the checkout suggests, and each link carries a different risk.
In the United States, substantially all loans facilitated through Affirm's platform are legally originated by Celtic Bank, an FDIC-insured Utah state-chartered industrial bank, and Lead Bank, an FDIC-insured Missouri state-chartered bank. Affirm holds direct lending licences in some jurisdictions but has no bank charter of its own. That structure exists for a specific reason: originating through a state-chartered insured bank lets the loan rely on the bank's federal interest-rate authority under Section 27 of the Federal Deposit Insurance Act, rather than on the lending law of the consumer's state. The whole pricing structure of the product rests on that doctrine.
Affirm then purchases the loans from the originating bank and services them, which is where the risk lands. It funds the portfolio through four channels — warehouse credit facilities, term and revolving securitisations, pass-through loan sales, and forward flow arrangements with banks, investment banks, pension funds, asset managers and insurers. There are no deposits. FY2025 funding costs were US$425.5 million and the provision for credit losses was US$616.7 million. Affirm carries origination, credit and funding risk on its own balance sheet rather than pushing it all to a partner bank.
For the merchant the mechanics are simpler: Affirm approves or declines the consumer, the merchant is paid for the sale, and repayment performance is Affirm's problem. Affirm also issues the Affirm Card, a debit card that pays in full from a linked bank account or converts a purchase into an instalment loan before or after the fact, and issues virtual cards for use at merchants that have not integrated it. It is not a payment processor: no card acquiring, no terminals, no merchant settlement, no cross-border collection.
How Affirm prices
Affirm's pricing is two-sided, and only one side is public. Merchants pay a fee when Affirm converts a sale, negotiated per merchant, published nowhere. Consumers pay simple fixed interest on interest-bearing plans, nothing on interest-free plans, and no late fees under any circumstances.
The structural point a merchant needs to internalise is that interest-free promotional financing is not free — it is the most expensive product Affirm sells to a merchant, because the merchant is buying down the consumer's interest through a higher merchant fee. "Interest-free for the shopper" and "cheap for the retailer" are opposite statements. A merchant choosing between promotional financing and an interest-bearing plan is choosing how much of the conversion lift it wants to pay for directly.
Nothing about merchant contracts is published — no rate card, no minimum, no term, no termination provisions. What is published, unusually, is the consumer side: a fixed total disclosed at checkout, no compounding, no penalty fees. Most lenders do the reverse.
Where Affirm is genuinely strong
Per-transaction underwriting. Affirm approves or declines each purchase individually rather than extending a revolving line and hoping. That is a materially different risk posture from a credit card, it lets approval standards move with conditions in days rather than quarters, and it is the core competence the whole business rests on. The FY2025 provision for credit losses of US$616.7 million is the price of running it.
Higher-ticket, considered purchases. Furniture, electronics, fitness equipment, powersports, travel — categories where a monthly payment changes whether the purchase happens at all. Affirm's longer interest-bearing loans fit those baskets in a way a four-instalment plan does not, and the GMV mix reflects it.
Consumer terms that survive scrutiny. No late fees, no compounding, no revolving balance, a total fixed at checkout. For a merchant whose brand is exposed to how its financing partner treats customers, this is not a soft benefit — it is the difference between a partner that generates complaints and one that does not.
Distribution and integration. Affirm publishes merchant APIs and SDKs and is available natively through platforms including Shopify and Stripe, so enabling it is usually configuration rather than a build. The Affirm Card extends the same financing into physical stores.
It finally works as a business. FY2025 net income of US$52.2 million on revenue of US$3,224.4 million was Affirm's first profitable fiscal year, after net losses of US$517.8 million in FY2024 and US$985.3 million in FY2023. Repeat behaviour is strong: 94% of FY2025 transactions came from returning consumers, at 5.8 transactions per active consumer.
Where Affirm falls short
Merchant pricing is invisible, and it varies by product. A merchant cannot compare Affirm against Klarna, Afterpay or PayPal without running a sales process with each, and the fee depends on which plan mix is presented at checkout. Two merchants quoted "an Affirm rate" may be buying quite different things.
Funding is a real constraint, not a footnote. Affirm does not take deposits. Origination capacity depends on warehouse lines, securitisation markets and forward flow buyers remaining open at a workable cost — FY2025 funding costs were US$425.5 million. A deposit-funded competitor such as Klarna has a cheaper and less cyclical cost of funds, and in a stressed market that gap widens exactly when a merchant most wants approvals to hold up.
Approval rates move with the credit cycle. Affirm's economics depend on repayment, so when credit performance deteriorates, underwriting tightens and terms narrow — which is precisely the moment a merchant is counting on financing to defend conversion. Any conversion-lift model built on a benign credit environment is fragile.
Narrow geography and no payments infrastructure. The United States, Canada and the United Kingdom, in local currency. No cross-border collection, no FX, no recurring billing, no card acceptance, no terminals, no orchestration between financing providers. Affirm presents its own product only.
Concentration risk runs both ways. Large retail partnerships move. Walmart, one of Affirm's earliest and largest relationships announced in 2019, was reported to have moved to Klarna in March 2025 — a reminder that in BNPL the merchant relationship is contestable and re-tenders are frequent.
Regulation: a rulebook that reversed itself
Buy now, pay later has been the subject of a federal regulatory attempt that was made and then unmade within roughly a year, and any merchant integrating it should understand that sequence.
The CFPB opened an inquiry into BNPL in December 2021, covering debt accumulation, regulatory arbitrage and data harvesting. On 22 May 2024 it issued an interpretive rule treating BNPL digital user accounts as credit cards under Regulation Z, which would have imported dispute and refund rights and periodic statement requirements into the product. On 2 June 2025 the CFPB confirmed retraction of that interpretive rule. Both dates are disclosed in Affirm's own FY2025 Form 10-K. The practical position today is that US BNPL is governed by general consumer credit law, state lending and money transmission licensing, and the Fair Credit Reporting Act, rather than by a BNPL-specific federal rule — and that the federal position has already reversed once.
The more consequential exposure is preemption. Affirm's model depends on originating banks' federal interest-rate authority under Section 27 of the Federal Deposit Insurance Act and the FDIC's Federal Interest Rate Authority Rule of 22 July 2020. Affirm's 10-K discloses that the attorneys general of California, New York and Illinois sued the OCC, and that the attorneys general of California, Illinois, Massachusetts, Minnesota, New Jersey, New York and North Carolina sued the FDIC, over the rules underpinning that authority. It also flags private litigation and governmental enforcement actions seeking to recharacterise which party is the "true lender" in bank-partner arrangements. If preemption narrows, the pricing structure of the product is directly affected — not just Affirm's, but that of every bank-partner lender.
Elsewhere: Affirm U.K. Limited is authorised and regulated by the Financial Conduct Authority, and Affirm holds lending, servicing and brokering licences across US, Canadian and UK jurisdictions plus money transmitter licences in certain US states through Affirm Payments, LLC. In February 2023 it laid off 19% of its workforce and closed its cryptocurrency unit; a June 2024 breach at partner bank Evolve Bank & Trust affected some Affirm cardholders.
How to evaluate Affirm
For a merchant, the questions that determine whether this works are commercial and behavioural rather than technical.
- Price each plan type separately. Get the merchant fee for interest-bearing plans, interest-free monthly plans and short-term instalments individually, and model your realistic mix rather than the mix in the proposal.
- Ask what happens to approvals in a downturn. Understand who decides which consumers see which plans, how quickly credit policy can change, and whether you receive notice when it does.
- Measure incrementality, not attribution. The question is not how many orders used Affirm but how many would not have happened otherwise. Run a holdout before concluding that promotional financing pays for itself.
- Read the returns and disputes flow. Establish exactly how refunds, partial returns and chargebacks work when the consumer's obligation sits with a lender rather than a card issuer, and who the customer contacts when something goes wrong.
- Check credit-reporting disclosures directly. Affirm is subject to the Fair Credit Reporting Act, Regulation V and CARES Act credit-reporting requirements, but which loan types are furnished to which consumer reporting agencies is not stated in the FY2025 10-K. If your customer service team will be asked, get the current answer from Affirm's own consumer disclosures rather than from press coverage.
- Contract terms. Nothing is published. Establish term length, exclusivity if any, notice periods, and what happens to in-flight loans and consumer relationships if you switch providers.
Affirm suits merchants with considered, higher-ticket baskets who are prepared to pay for conversion and want a financing partner whose consumer terms will not embarrass them. It does not suit merchants who need cost certainty before a sales process, merchants outside its three markets, or anyone shopping for a payment processor — because Affirm is not one.
Capability assessment
Every company profiled on this site is assessed against the same eighteen dimensions, so the profiles can be read against one another. Each rating carries one sentence of evidence. There is no score out of ten, because a score is not defensible and a sentence is.
Core strength means a primary, differentiating capability. Supported means genuinely offered and documented, but not a differentiator. Limited means partial, geographically restricted, gated behind an enterprise tier, or delivered through a third party. Not offered means what it says. Unclear means the company markets the capability but does not document it well enough to judge — which is itself a finding.
Who Affirm suits
- Merchants selling considered, higher-ticket purchases — furniture, electronics, fitness equipment, travel, powersports — where a monthly instalment materially raises conversion and average order value.
- Merchants willing to buy conversion by subsidising 0% APR promotional financing, since Affirm charges a higher merchant fee on those plans and passes the benefit to the shopper.
- Merchants who want a financing partner whose consumer terms are defensible under scrutiny — no late fees, no compounding interest, no revolving balance, and a total cost fixed at checkout.
- Merchants on Shopify, Stripe and other platforms where Affirm is available as a native integration rather than a bespoke build.
- Merchants that want a counterparty with audited public financials: FY2025 revenue of US$3.224 billion, net income of US$52.2 million — its first profitable fiscal year — and GMV of US$36.7 billion across approximately 377,000 active merchants.
- Retailers wanting reach beyond their own checkout, since the Affirm Card lets a consumer bring Affirm financing into any store.
Who Affirm is a poor fit for
- Merchants who need to compare cost before entering a sales process. Affirm publishes no merchant pricing at all, and the fee varies materially by product — 0% APR promotional financing costs the merchant more than interest-bearing plans, because the merchant is buying down the consumer's interest.
- Merchants uncomfortable with regulatory whiplash in BNPL. The CFPB issued an interpretive rule on 22 May 2024 extending credit-card-style requirements to BNPL providers, then confirmed retraction of that rule on 2 June 2025, per Affirm's own 10-K. The federal position has reversed once within about a year, so a merchant integrating BNPL is integrating into a rulebook that is not settled.
- Merchants exposed to 'true lender' litigation risk by association. Affirm's 10-K flags that there have been private litigation and governmental enforcement actions seeking to recharacterise lending transactions, and notes state attorneys general lawsuits against the OCC (California, New York, Illinois) and the FDIC (California, Illinois, Massachusetts, Minnesota, New Jersey, New York, North Carolina) over federal interest-rate preemption — the doctrine Affirm's bank-partner model depends on. If preemption narrows, the 0-36% APR structure is directly affected.
- Merchants whose customers are credit-stressed. Affirm's provision for credit losses was US$616.7 million in FY2025, and its economics depend on consumer repayment; approval rates and available financing terms tighten when credit performance deteriorates, which is precisely when a merchant most wants the conversion lift.
- Merchants who want a payment processor. Affirm is a lender, not an acquirer — no card acceptance, no terminals, no settlement, no cross-border. It always sits alongside a processor, never instead of one.
- Merchants outside the US, Canada and the UK. There is no broader international footprint.
- Anyone assuming Affirm's own funding is unconditional. Affirm does not fund from deposits; it depends on warehouse credit facilities, securitisations, pass-through loan sales and forward flow arrangements with banks, investment banks, pension funds, asset managers and insurers. FY2025 funding costs were US$425.5 million. A closed securitisation market constrains origination in a way a deposit-funded lender would not face.
- Consumers or merchants who assume BNPL activity stays off a credit file. Affirm is subject to the FCRA and Regulation V and reports certain loans to consumer reporting agencies; which loans go to which bureau has changed over time and should be checked against Affirm's current consumer disclosures rather than assumed.
Competitors and alternatives
| Company | Why a business would choose it instead |
|---|---|
| Klarna | Named as a competitor in Affirm's own 10-K; a deposit-funded bank with a cheaper cost of funds, broader international reach and a much larger consumer base. |
| Block (Afterpay) | Named in Affirm's 10-K; bundles BNPL with Square's merchant acquiring and Cash App's consumer base, so a merchant can get acceptance and financing from one vendor. |
| PayPal (Pay in 4 and Pay Monthly) | Named in Affirm's 10-K; already present in most checkouts, so enabling BNPL is often a toggle rather than an integration. |
| Synchrony | Named in Affirm's 10-K; the incumbent in private-label and promotional retail financing, with deposit funding and decades of merchant programme experience. |
| Bread Financial | Named in Affirm's 10-K; competes directly for retailer instalment and private-label programmes. |
| Chase, Citi, Capital One, Bank of America and American Express | Named in Affirm's 10-K; card issuers offering plan-it and instalment features on existing cards, which reaches the same consumer without a new merchant integration. |
| Zip and Sezzle | Smaller pure-play BNPL providers competing on merchant fee and on approval rates for thinner-file consumers. |
| The merchant's own financing programme | Affirm's 10-K notes merchants increasingly offer proprietary financing in parallel, which cuts the third-party fee entirely for large retailers. |
Affirm — frequently asked questions
How does Affirm make money?
Mostly from interest, not merchant fees. In the fiscal year ended 30 June 2025, total revenue was US$3,224.4 million: interest income of US$1,608.2 million (about half), merchant network revenue of US$882.7 million, gain on sales of loans of US$381.6 million, card network revenue of US$231.3 million and servicing income of US$120.6 million. Interest-bearing monthly instalment loans made up 72% of GMV, interest-free monthly plans 13% and short-term instalment plans 14%. On interest-free promotional plans the merchant pays a higher fee because it is subsidising the consumer's interest.
Does Affirm charge late fees?
No. Affirm's FY2025 Form 10-K states it has charged US$0 in late fees for missed payments since founding, that it does not profit from consumers' mistakes, and that consumers do not pay more than what they agreed to at checkout even if a payment is missed or late. Interest on interest-bearing loans is simple and fixed rather than compounding, so the total shown at checkout is the total owed. Missing payments can still affect a consumer's ability to be approved for future Affirm loans.
Who actually lends the money when a consumer uses Affirm?
In the United States, substantially all loans facilitated through Affirm's platform are originated by Celtic Bank, an FDIC-insured Utah state-chartered industrial bank, and Lead Bank, an FDIC-insured Missouri state-chartered bank; Affirm also holds direct lending licences in certain jurisdictions. Affirm then purchases and services the loans, funding them through warehouse credit facilities, term and revolving securitisations, pass-through loan sales and forward flow arrangements. Affirm has no bank charter of its own — its FY2025 10-K describes pursuing a federal charter only as a contingency if bank partnerships were lost. The Affirm Money Account is provided through Cross River Bank.
What did the CFPB do about buy now, pay later?
The CFPB opened an inquiry in December 2021 into BNPL debt accumulation, regulatory arbitrage and data harvesting. On 22 May 2024 it issued an interpretive rule extending Regulation Z credit card requirements — including dispute and refund rights and periodic statements — to BNPL providers using digital user accounts. On 2 June 2025 the CFPB confirmed retraction of that interpretive rule, as disclosed in Affirm's FY2025 Form 10-K. US BNPL is therefore currently governed by general consumer credit law, state lending and money transmission licensing and the Fair Credit Reporting Act rather than by a BNPL-specific federal rule.
Does Affirm report to credit bureaus?
Affirm is subject to the Fair Credit Reporting Act as amended by the FACT Act, to Regulation V and to CARES Act credit-reporting requirements. However, its FY2025 Form 10-K does not name which consumer reporting agencies receive data or which loan types are furnished, and reporting practice across the BNPL industry has changed repeatedly. Anyone who needs a definitive answer — a consumer, or a merchant briefing its support team — should check Affirm's current consumer disclosures directly rather than rely on secondary reporting.
Is Affirm a payment processor?
No. Affirm is a point-of-sale lender. It does not acquire card transactions, sell terminals, settle merchant funds, provide recurring billing or offer cross-border collection and FX, and it operates only in the United States, Canada and the United Kingdom. It sits alongside a merchant's existing payment processor as a financing option at checkout, typically enabled through a platform such as Shopify or Stripe or through Affirm's own APIs and SDKs.
How does Affirm differ from Klarna?
Funding and product mix. Klarna is a licensed Swedish bank that funded 90% of its lending from US$13 billion of consumer deposits at 31 December 2025 and operates in 26 markets. Affirm has no charter: it originates through Celtic Bank and Lead Bank and funds through warehouse lines, securitisations and loan sales, paying US$425.5 million in funding costs in FY2025, and operates in three markets. Affirm charges no late fees at all and leans toward longer interest-bearing loans, 72% of its FY2025 GMV; Klarna leans toward short interest-free instalments. Affirm reported net income of US$52.2 million in its fiscal year to 30 June 2025; Klarna reported a net loss of US$273 million for calendar 2025.
Sources
This profile was built from the following primary and secondary sources. Where sources disagreed, the disagreement is stated in the text rather than resolved silently.
- https://www.sec.gov/Archives/edgar/data/1820953/000182095325000080/afrm-20250630.htm
- https://www.sec.gov/Archives/edgar/data/1820953/000182095325000080/R5.htm
- https://www.sec.gov/Archives/edgar/data/1820953/000182095325000080/FilingSummary.xml
- https://data.sec.gov/submissions/CIK0001820953.json
- https://en.wikipedia.org/wiki/Affirm_Holdings
- https://investors.affirm.com/news-events/newsroom
- https://www.affirm.com
- Credit-reporting specifics. Affirm's FY2025 10-K confirms FCRA, Regulation V and CARES Act credit-reporting obligations but does not name Experian, TransUnion or Equifax, does not say which loan types are furnished, and gives no dates. Widely discussed 2025 developments — Affirm beginning to furnish Pay in 4 loans to specific bureaus, and FICO introducing BNPL-inclusive scores — could NOT be verified from a primary source in this research. Do not publish a bureau name, a date, or a FICO product name without a primary source.
- Affirm's bank charter status. The English Wikipedia article states a bank charter application was filed in 2025, but Affirm's FY2025 10-K (filed 28 August 2025) describes pursuing a federal charter only as a contingency if bank partnerships were lost, and discloses no pending application. Treat the Wikipedia claim as unverified and probably wrong.
- Item 3 Legal Proceedings of the FY2025 10-K was not read in this research, so no statement can be made about pending litigation against Affirm beyond the risk-factor language quoted.
- Active consumer count and 30-day delinquency rate for FY2025 were not extracted and should not be published without checking Affirm's supplemental metrics.
- Fiscal Q3 2026 results were reported on 7 May 2026 but the figures were not retrieved, so all financial data here is as of the fiscal year ended 30 June 2025 and should be labelled accordingly.
- Merchant partner claims sourced from Wikipedia — Amazon, Walmart (reported moved to Klarna in March 2025), Target, Apple, Best Buy, Shopify, Costco — were not confirmed against Affirm's filings or press releases and should be checked before publication.
- The Returnly acquisition price of US$300 million and the IPO raise of approximately US$1.2 billion come from Wikipedia rather than a filing.
- Merchant pricing is not published in any form, so no merchant discount rate, minimum or contract term can be stated.
- Whether Affirm operates in Australia could not be confirmed; the 10-K names the US, Canada and the UK.