Happy Friday, Fintech Takers!

Prediction markets are not the subject of today’s newsletter (thank goodness), but I do have to comment on one thing real quick.

On January 13th, 2025 — about a week before his father was inaugurated — Donald Trump Jr. became a paid advisor to Kalshi, a role that reportedly earned him $300,000 in equity. On August 26th, 2025, Trump Jr.'s venture capital firm (1789 Capital) invested $200M in Polymarket and he joined the board as an advisor. On August 31st, it was reported that his firm committed an additional $300M to Polymarket. On November 3rd, 2026, voters in the U.S. gave control of Congress (and numerous state legislatures and governorships) to Democrats, handing Republicans a historic defeat. Among the top issues cited by voters in this election was concern over political corruption. In the wake of the election, multiple Democrats in both the House and the Senate have promised to scrutinize the prediction markets and their ties to the White House.

And yesterday, Kalshi announced that it has ended its advisory relationship with Trump Jr.

Allow me to offer you two different justifications for this decision:

1. The most recent investment by 1789 Capital in Polymarket created a conflict that Kalshi and Trump Jr. both felt was disqualifying for him continuing in his role at Kalshi. This is the explanation that the company has provided.
2. Kalshi is trying to minimize the negative effects of the scrutiny that Congressional Democrats are going to bring to the prediction market industry, particularly as it relates to political corruption.

Reasonable people can come to different conclusions regarding this news (this is mine), but I will point out that explanation #1 strains credulity because Trump Jr. had already been an investor in and advisor to Polymarket for more than a year. Additionally — in support of conclusion #2 — Bloomberg has reported yesterday that Kalshi is in talks to raise money from investors at a $65B valuation, mere weeks after closing a funding round at a $40B valuation, and most of their business rests on the exploitation of a single, massive regulatory loophole that could be closed at any time.

Anyway, thanks for indulging me. Now let’s get onto more interesting and important news.

- Alex

Can Revolut Dominate the U.S. Without Taking Risk?Copy anchor linkCopied

There are a few quotes that I want to talk about.

The first is from an interview that Revolut CEO Nik Storonsky did with Bloomberg, in which he was asked about his company’s plans for expanding into the U.S. and competing directly with credit card giants like American Express and JPMorgan Chase. Storonsky didn’t equivocate, stating that Revolut’s credit card product in the U.S. would be “at par” with Amex and JPMC, and that he is coming for JPMC CEO Jamie Dimon’s “dinner” and wants U.S. consumers to see Revolut as “a new JPMorgan.”

“Everywhere we want to go, we want to dominate,” Storonsky said. “Otherwise, what’s the point?”

So, that’s the plan. Go right at the biggest banks in the U.S. and don’t shy away from competing directly with them in the consumer product category (premium credit cards) where they are strongest.

Bold!

Then, this week, Storonsky went on Stripe’s Cheeky Pint podcast, where he affirmed that Revolut’s go-to-market approach in the U.S. (where Revolut just got preliminary conditional approval for a national bank charter) would be “credit-card led,” and that the company would go head to head with Amex Platinum and Chase Sapphire by offering “the same benefits or better.”

Storonsky was also asked in the Cheeky Pint interview if Revolut would have a traditional, credit-led bank business model. Here was his answer:

We want to have the least amount of risk. Even if we lend, it won’t be a large portion of the balance sheet. At the moment, target state is less than 10%, and then the rest we would securitize.

When asked why, he elaborated:

First, the simpler the model, the better from a regulatory point of view. Plus, again, if we want to be a tech business, we would rather be in the subscription business. As soon as you go into the lending business, your multiples become similar to banking multiples. If you look at tech business multiples, you can usually get like 40-50 PE. In the banking business, you’ll be lucky if you get 10 PE.

And elaborated some more:

If you look at the credit business in banking, general credit business — every time you do a loan or credit card, you need to put provisions against it. In reality, the amount of these provisions is highly dependent on the models that you use, on economic conditions. Because in reality you don’t know ... In three, or four, or five years, will you receive the loan back, right? You kind of model it. On average you hope to receive money back, but really things can change, right? This uncertainty is being priced in by investors and generally by lower multiples. Because the bank can produce a lot of money now from the credit portfolio. But look, in 2-3 years’ time unemployment goes from 3% to 6%.

And concluded with this:

It’s similar to selling options, right? You kind of collect a premium for 10 years, and then year number 11, you lose it all, right? That’s a problem.

Those answers are … kinda weird. And worth unpacking a bit more.

Why the Obsession Over Price-to-Earnings Multiples?Copy anchor linkCopied

Let’s start, as Storonsky does¹, with the argument that subscription-based business models earn tech-level valuations (40-50x price-to-earnings multiples) while lending-based business models earn bank-level valuations (10x P/E multiple, if you’re lucky).

To be blunt, that’s a very bad argument to make for your business model of choice.

The business model you choose should be the one that you think creates the strongest possible business and the most value for your customers. And then you’re supposed to trust that investors will, over time, reward you for building a great business that your customers love.

To his credit, Stripe’s John Collison (who was conducting the Cheeky Pint interview) made this exact point in his response to Storonsky:

I’m always skeptical of multiples-based arguments because, ultimately, the business is good or it isn’t. And, you can educate investors if the business is, in fact, good. When we started Stripe, people thought payments businesses were horrendous and deserving of low multiples, and everything like that. Over time, they came to appreciate payments businesses more. But, if we had been fixated on the multiple that people ascribed to the business, we might have been scared off this path. 

Yes! That is exactly the right way to think about it!

When pressed about it, Storonsky went on to explain why he thinks lending off of your own balance sheet is a bad business (this is where his comments on loss provisioning and the comparison to selling options come in). We’ll get to that argument, but before we do, it’s worth pausing and trying to understand why his first instinct was to defend Revolut’s business model using a valuation-based argument.

I think the answer might be sitting in his compensation plan.

Storonsky owns roughly 29% of Revolut, and he has what has been described as an Elon Musk-style incentive package. Instead of a big salary, he is awarded additional shares in stages, as the company’s valuation crosses a series of pre-agreed thresholds. It’s already working. By his own account, his stake climbed from about 25% to about 29% as Revolut’s valuation rose from $45B to $75B. And if the company reaches $200B (it’s at $115B today), he gets another 10%, which would put him at roughly 40% of the company and a stake worth about $80B. According to the FT, he’s also in talks with investors on a second package that would kick in at $500B.

Notice what’s missing from that list: Profit, return on equity, anything having to do with customers. The scoreboard is valuation. And valuation is just earnings times a multiple.

Storonsky has even done the math out loud. He’s said that getting to $200B would take $5B-$6B in annual net profit, which works out to a multiple somewhere between 33x and 40x. At the 10x he says banks are lucky to get, he’d need $20B in profit, more than ten times what Revolut earned last year. When your personal net worth is a function of your company’s valuation multiple, it’s not surprising that it’s one of the first things out of your mouth when someone asks you about your business model.²

This also helps explain the brash messaging around Revolut’s planned expansion in the U.S.³ To get to a $500B valuation, you need to show investors that you are willing to pick a fight with the biggest bullies in the schoolyard.

But can Revolut win that fight?

Too CautiousCopy anchor linkCopied

Setting aside Storonsky’s focus on P/E multiples, his more fundamental argument against lending as a good business evinces a surprising lack of understanding of how lending actually works.

Just look at the quotes:

“Every time you do a loan or credit card, you need to put provisions against it.”

In many cases, yes. But not always! In fact, for installment loans it’s becoming a bit of a trend among fintechs-turned-banks to reject the standard amortized cost accounting model and the CECL loss provisioning framework that goes with it. SoFi has done this for years, and Happen Bank (formerly LendingClub) recently made the switch to fair value option accounting as well. These are the banks that Revolut will be most directly competing with in the U.S., at least in terms of investor comps, and they have already found a way to address this concern.

“In reality you don’t know ... In three, or four, or five years, will you receive the loan back, right? You kind of model it. On average you hope to receive money back, but really things can change, right? … Because the bank can produce a lot of money now from the credit portfolio. But look, in 2-3 years’ time unemployment goes from 3% to 6%.”

“You kind of model it” is such an interesting phrase. It conjures an image of a couple of guys in an office throwing darts at a dartboard while drinking scotch in the middle of the day. Just a very casual and unserious approach to predicting and pricing risk; an approach that can and almost certainly will be thrown into chaos when unemployment ticks up by a few percentage points.

That is, to put it mildly, quite different from the way that chief risk officers and data scientists in financial services think about loan underwriting.

There is a level of uncertainty inherent to loan underwriting, certainly. No one knows what might happen in the future and lenders have definitely blown themselves up by making big bets on the wrong predictions. However, Storonsky’s framing almost makes it sound as if there’s no speed slow enough for banks and fintech companies to safely lend money, and that’s just not true. The best lenders have very sophisticated risk underwriting models, trained on historically rich datasets. They have experience navigating changing credit cycles and they know how to tighten their risk tolerance and reduce their exposure (both upfront during underwriting, and later in loan servicing) in response to the performance of their portfolios and changes in the macro environment.

Put simply, it’s possible to run a consumer lending portfolio without blowing up the entire company. In fact, I’d argue that catastrophic credit risk blowups are actually the exception rather than the rule in our industry.

So why does Storonsky seem to believe otherwise? Why does he compare lending money to selling options?

I don’t know, but it may have something to do with the fact that he started his career as a derivatives trader at Lehman Brothers, from 2006 to 2008. As he said in a recent interview, his years on trading desks taught him, “you just shouldn’t take a risk in complex things.”

That’s an understandable position to hold. I would probably feel the same if my formative experiences in financial services had been at fucking Lehman Brothers in the lead up to the Great Financial Crisis. However, it’s a tough philosophy to square with Storonsky’s stated goal to go head to head with JPMC and Amex and to dominate in the U.S. market.

Allow me to explain why.

JPMorgan ChaseCopy anchor linkCopied

If you want to offer a credit card that is “at par” with Chase, the first thing you need to understand is how Chase makes money on credit cards.

It’s not interchange.

I mean, Chase collects a lot of interchange revenue, obviously. Its customers spent $1.94T on their debit and credit cards last year. But after paying for rewards, co-brand partner payments, and the cost of acquiring new accounts, JPMC’s consumer bank kept just $2.2B in net non-interest card revenue. That’s about 11 cents for every $100 its customers spent. And that number is going the wrong way. It was $3.1B the year before, and it fell because, in the bank’s own words, growth in card spending “was more than offset by higher rewards costs and partner payments.”

Chase has essentially the same business model that movie theaters have: Break even-ish on the ticket sales (after paying the studios) and make your profit on the popcorn.

For the bank, the popcorn is interest and fees on revolving credit card balances.

Chase ended 2025 with $248B in credit card loans. On its card loans, it earned roughly $23.8B in net interest income last year. Its total net revenue from credit cards over that same time period was roughly $23.4B.

Read those numbers again. The interest alone was bigger than the total. That’s because everything else Chase makes on those cards (interchange, annual fees, etc.), after subtracting the costs that come with them (rewards, partner payments, acquisition costs, etc.), added up to less than zero. In other words, the points and airport lounges and travel credits aren’t the product. They’re the customer acquisition and retention costs for the product.

The product is the loan.

And it is very much a risk-forward business. Chase charged off roughly 3.3% of its card loans last year, which worked out to somewhere around $7.7B in losses. That’s more than four times what Revolut earned, in total, across its entire global business last year. Chase also holds $15.6B in reserves against future card losses, which is about two and a half times Revolut’s annual revenue.

Chase isn’t doing this because it’s reckless. It’s doing it because it is exceptionally good at it. It knows, with a great deal of precision, which customers will revolve, which ones will pay, and how both of those numbers will move when unemployment goes from 3% to 6%. The bank’s experience and analytic expertise are what allow it to hand essentially all of its credit card interchange and annual fee revenue back to customers and partners and still run one of the most profitable consumer franchises on the planet.

So when Storonsky says that Revolut’s premium U.S. credit card will be at par with Chase’s, while also saying that lending will be less than 10% of the balance sheet and that the goal is to have “the least amount of risk,” I have a simple question: Who’s paying for the points?

American Express   Copy anchor linkCopied

“OK,” you might say. “So, Revolut has a really long way to go to successfully compete head to head with JPMorgan Chase. But what about Amex? They make a ton of money from their credit cards and they have a payments-focused business model, not a lending-focused one. Why can’t Revolut become the new American Express?”

It’s a fair question. And on the surface, the numbers look encouraging.

Last year, American Express generated $72.2B in revenue. More than half of that ($37.4B) came from the fees it charges merchants. Another $10B came from annual fees. Net interest income was just 24% of the total.

That’s remarkably similar to Revolut, which gets 76% of its revenue from fees and 22% from interest. Swap “subscriptions” for “card fees” and the two income statements become hard to tell apart.

But income statements aren’t the whole story.

Amex is a mullet. Payments business in the front. Lending business in the back.

Look at the balance sheet. Amex ended 2025 with $214B in card balances. About $152B of that was card member loans, meaning balances that customers are able to carry and pay interest on. The other $62B was the old-school charge card business, where the balance is due in full every month. But even that part isn’t risk-free. It’s short-term unsecured credit to customers who spend a lot of money, and Amex wrote off more than 1% of it last year.

Then look at where the profit comes from. Amex earned $17.4B in net interest income last year and set aside $5.3B for credit losses. The difference, about $12B, is roughly 90% of the company's entire pretax income of $13.8B. All of that merchant fee revenue? About half of it goes right back out the door as rewards ($18.4B last year).

Sound familiar?

Additionally, lending is one of the fastest-growing parts of the business. Net interest income grew 12% last year, twice as fast as merchant fees. It was 22% of Amex’s revenue in 2023. It was nearly 25% last quarter. The spend-centric company is becoming a little less spend-centric every year.

And, of course, Amex can be as spend-centric as it is because it’s both an issuing bank and a card network. The company collects roughly 2.2% of every dollar its cardholders spend because it is the issuer, the acquirer, and the network. Revolut will issue its U.S. cards on Visa’s and Mastercard’s rails and collect only a portion of the resulting interchange fee revenue, like every other issuer in a four-party network.⁴

So, no. I don’t think Revolut can be the new Amex either. Not with a lending book targeted at 10% of the balance sheet. Amex’s payments fee revenue is real, but it sits on top of a closed-loop network and $214B in credit exposure.

AffirmCopy anchor linkCopied

Let’s walk through one more example, because it may be the most instructive one for Revolut.

Max Levchin has been saying for years that Affirm is trying to build the modern equivalent to American Express. And unlike Revolut, it has the network to show for it.

As of last quarter, Affirm had 28.7 million active consumers and 621,000 active merchants, including Amazon, Walmart, and Shopify. It processed $14.4B in volume in the quarter (up 34%) and expects more than $65B this fiscal year. The Affirm Card, which lets consumers take the network with them to merchants that haven’t integrated it, has 6 million active cardholders and did $3.3B in volume last quarter, more than double a year ago.

It’s a real two-sided network, and is, in Levchin’s words, a “fully closed loop”: Affirm is the issuer of the credit, the acquirer, and the risk manager. And merchants pay for it. Affirm’s merchant fees ran at about 2.3% of total volume last fiscal year, which is right in line with what Amex collects.

But here’s the critical part. Affirm didn’t build that network by offering merchants one more way for their customers to pay. Merchants don’t want any more help on that front. What they want (and what they are willing to pay for) is more sales, and Affirm delivers more sales by doing what Storonsky seemingly doesn’t want to do: Take on the credit risk. Nearly every Affirm transaction is a loan, underwritten individually, at the point of sale. That’s the product merchants are paying for.

And Affirm’s financials reflect it. Roughly two-thirds of its revenue last fiscal year came from credit (interest income, gains on loan sales, and servicing), not from network fees, and it was holding $9B in loans on its own balance sheet at the end of June.

To Win in the U.S., You Have to Take Credit RiskCopy anchor linkCopied

There are lots of ways for Revolut to win in the U.S.

However, if you want to win at the scale that Storonsky wants to win at (“we want to dominate”) and earn the valuation multiple that Storonsky’s compensation incentivizes him to hit, you have to build a credible competitor to what the biggest and most successful banks — your JPMorgan Chases and American Expresses — have built.

And to do that, you need to get comfortable taking credit risk.

You can try to build the most compelling value proposition for consumers. That’s the Chase playbook. But the points and the lounges aren’t free, and interchange and annual fees will barely cover the costs. You’ll need the lending revenue to be profitable.

You can try to build the most compelling value proposition for merchants. That’s the Affirm playbook. But merchants don’t need another payment button. What they’ll pay for is someone who is willing to take on credit risk in order to help them sell more stuff.

You can try to build a closed-loop network that earns a higher take rate than Visa’s and Mastercard’s rails will ever give you. Amex and Affirm have both done it. And yet, even with that advantage, two-thirds of Affirm’s revenue and nearly 90% of Amex’s pretax profit still come from lending.

You can even try to do what Storonsky says Revolut will do, which is to originate the loans and then get them off your balance sheet. However, there are two problems with that plan.

The first is that it’s a lot harder to do with credit cards than it is with installment loans. An installment loan is a discrete thing. Fixed amount, fixed term. You can sell it and be done with it. A credit card is an open line of credit. The customer keeps borrowing and the issuer keeps the account. You can securitize it, but it’s much trickier to do than it is for an installment loan. For this and other reasons, most large credit card issuers keep most of their balances on their books.

The second problem is that, even where selling loans does work, it doesn’t excuse you from being good at underwriting. Affirm sells a large share of what it originates, and it has $30B in funding capacity from loan buyers and capital markets partners to prove it. But those buyers show up because they’ve watched how Affirm’s loans have performed for more than a decade. To sell loans, you first need to demonstrate that you are good at assessing and pricing risk.

And that’s really the whole point. The winners in U.S. consumer finance aren’t the companies that found a way to avoid credit risk. They’re the companies that got better at it than everyone else. Amex, the supposedly spend-centric one, wrote off 2.2% of its card loans last year, compared to Chase at 3.3%. Affirm is widely considered to be one of the most sophisticated consumer credit risk shops on the planet, and Levchin has described this advantage by saying, “we love math here more than just about anything else. I think most of our competitors just don't. And, that's our strength. Our advantage is we live better through mathematics.”

There’s no way around it. Whichever strategy you pick, winning in the U.S. at the scale that Revolut wants to win at (and that Storonsky needs it to win at) requires getting very good at safely taking credit risk.

Which leaves us with the two statements from Nik Storonsky that I opened this essay with:

1. We want to dominate.
2. We want to have the least amount of risk.

In the U.S., at least, he’s going to have to pick one.⁵

WHERE I'LL BECopy anchor linkCopied

Next week! Get excited!!

✈️ AFC Policy Summit | November 17 | Washington D.C.Copy anchor linkCopied

I am speaking. I’ve been assured that I am, indeed, speaking at the event. The topic keeps changing, but that’s fine. I am adaptable!

✈️ Fintech NerdCon | November 18-20 | San DiegoCopy anchor linkCopied

This is shaping up to be quite the event. Plus, you know, lots of sunshine!

WHERE I'VE BEENCopy anchor linkCopied

*The Science (and Fiction) of Friction (by me, with Spinwheel) 📚Copy anchor linkCopied

Okay, technically this is where I was back in September, but the conversation was too good not to resurface. At our virtual event, Spinwheel CEO Tomás Campos graded financial services a “solid B+” for cutting unnecessary friction and a “C+, or so” for friction associated with fraud, risk and regulatory controls. Missed the event? Our new report has a five-question framework for deciding which friction stays. Read it here.

*This rec is brought to you by one of our fantastic brand partners.

Thanks for the read! Let me know what you thought by replying back to this email.

— Alex