Affirm just posted its most profitable quarter in company history, and investors sold the stock down 4.26% anyway. That gap tells you something. The market isn't doubting what Affirm already built. It's squinting at what comes next.

What comes next is a card.

The Affirm Card is the company's main growth story now, and on the August 27 earnings call, founder and CEO Max Levchin was direct about the ambition and the difficulty simultaneously. He's stepping back from day-to-day operations, handing President duties to Michael Linford and global markets to Pat Suh, so he can work on products that, in his words, "will only show up in fiscal '28, '29, and so on." The present business, he implied, can run itself. The future business needs his full attention.

What the card actually is

The Affirm Card runs on something called the Visa Flexible Credential, a technical setup that lets a single card toggle between different funding sources at the point of sale. Practically, that means a shopper can use their Affirm account to split a big purchase into installments while using the same card for everyday spending. Rewards come through merchant-funded zero-percent financing programs, which Affirm says delivers value equivalent to 8% to 15% cash back, better than most traditional rewards cards.

The early numbers are compelling. Cardholders spend twice as much on Affirm's network as average users. Thirty percent of Affirm Card transactions are now happening in physical stores, which is a meaningful shift for a company that grew up entirely online. Card holders represent 19% of active users, and that cohort is deepening, not just widening, the company's relationship with its customers.

Affirm processes roughly 100 million transaction requests per quarter, which gives it a real-time view of credit behavior it can adjust quickly. The company is also still at only 10% of e-commerce merchants and just 80 of the top 250 US e-commerce sites, which management frames as a long runway ahead. Add a "Services" vertical that nearly doubled in volume after two major platform integrations, and the business looks, on paper, like a flywheel gaining speed.

Where it gets hard

The rewards math depends entirely on zero-percent consumer financing, and Levchin described underwriting those loans as a "really, really hard science." A small misjudgment in who gets approved for zero-percent terms doesn't just hurt margins. He said it can produce "a lot of unprofitable transactions," fast.

That risk gets harder to manage as Affirm moves into physical retail. Online purchases are clean data environments. In-store is messier: connectivity issues, different consumer behavior, less predictable transaction signals. Levchin said the "bar for error is much lower" in physical stores, which is a candid warning from a CEO about his own expansion plan.

Enterprise sales, meaning large merchants adopting Affirm's platform at scale, are also moving slowly, which matters because the card's economics depend partly on merchant-funded rewards and broad merchant acceptance.

Affirm's fiscal 2027 guidance projects the same revenue margin (revenue after transaction costs) as 2026, at 4.2%. Holding steady isn't a disaster, but it signals that the physical card push won't show up in the near-term financials. The payoff, if it comes, is what Levchin is building toward in his more product-focused role, including a bank partner platform called Affirm Edge, with pilot programs planned for the second half of this year.

The deeper pattern here is one the buy-now-pay-later industry keeps running into. Online installment payments were a clean product for a clean environment. The real world, with its physical stores, mixed connectivity, and harder-to-model behavior, is a different problem. Affirm has better underwriting infrastructure than most of its competitors. Whether that's enough to make zero-percent in-store financing work at scale, without a flood of losses, is what the next two years will answer.

Investors, apparently, aren't sure yet either.