Happy Monday, Fintech Takers!

Sorry for the late arrival time on today’s newsletter. It was a busy day, coming off of a rather astonishing weekend.

Here’s a partial of what happened between Friday and Sunday:

• SpaceX jumped 19% on Friday after the biggest initial public offering ever. The stock closed at around $161, valuing the company at $2.1 trillion. I truly have no idea what numbers mean anymore!

• The World Cup kicked off. I watched the U.S. – Paraguay match, but that’s about it so far. I have no idea how soccer works or what to expect!

• After warning everyone that its latest frontier model — Mythos — is basically a cybersecurity weapon of mass destruction, Anthropic disabled customer access to Fable (the commercialized version of Mythos) after Amazon researchers got around its safeguards and the U.S. government ordered Anthropic to suspend all use by foreign nationals. I have no idea how to contextualize anything Anthropic says or does these days!

• The New York Knicks defeated the San Antonio Spurs in the NBA Finals in 5 games, despite trailing for most of all of the games in the series and only outscoring the Spurs, cumulatively, by 12 points. I thought I understood how professional basketball works, but I have no idea how New York pulled that off!   

In short, I know nothing!

But thanks, as always, for reading along as I try to figure things out. Today, we have plenty of interesting fintech questions to dig into.

- Alex

P.S. — If you’re a stablecoin skeptic (a position I completely understand!) you’ll want to tune into this session that I am having with Western Union and Rain in a few weeks. I promise you’ll learn something.

Was this email forwarded to you?

Subscribe Here

Cyrano de Bergerac by Zacharie Heince.

3 FINTECH NEWS STORIES

#1: Pagaya Sues Klarna

What happened?

A fintech company that I am personally obsessed with just sued one of its former partners:

Pagaya Technologies sued its former partner, Klarna, last month for allegedly misappropriating its trade secrets related to point-of-sale loan underwriting.

The New York-based AI underwriting fintech and Sweden-based buy now, pay later firm have worked together since 2022, with Pagaya assessing subprime customers for Klarna at checkout.

But Pagaya alleged that Klarna “distilled Pagaya’s core underwriting technology for POS loans and then willfully misappropriated its trade secrets” – and allegedly breached agreements.

Klarna’s objective, Pagaya asserted, was to “absorb Pagaya’s trade secrets, use them to build its own competing capabilities, and cut Pagaya out of the very business Pagaya made possible for Klarna.”

All the while, Klarna was allegedly “misrepresenting to Pagaya an intention to expand the partnership, thus causing Pagaya to invest in a venture that, unbeknownst to Pagaya, had no real future.”

So what?

Consumer lending + Complex fintech infrastructure + A spicy lawsuit between former partners = a topic tailor-made for me. Seriously, inject this one directly into my veins.

In case you’re not familiar with Pagaya, it operates an embedded second-look lending network. Second-look lending isn’t a new concept, but Pagaya introduced a new wrinkle, as I wrote about last year:

Pagaya’s innovation is to integrate directly within the lender’s digital origination workflow. When an applicant is rejected by the lender’s decision engine, they (and all of their data) can be instantly passed to the Pagaya decision engine to see if it can approve them. If it can, the lender books the loan, sells the paper to Pagaya, and retains the servicing relationship with the customer.

The front-end experience is seamless. Whether the customer is approved by the lender or by Pagaya behind the scenes, it feels to them like the lender said yes. They download the lender’s mobile app, make payments, and contact them if they have any problems. They never see or feel Pagaya’s presence at all.

To fund these loans, Pagaya relies on a variety of different mechanisms, including selling groups of loans as asset-backed securities (ABS) to investors, forward flow agreements with large investors (the investor commits to buying a stream of loans originated by Pagaya’s lending partners on an ongoing basis), and capital funds controlled directly by Pagaya.

Pagaya provided this second-look lending capability to Klarna, specifically for its “Fair Financing” product in the U.S., which is a point-of-sale installment loan, typically with a term between 3 and 36 months and interest rates ranging from 0% to 35.99%. The product — which has been one of the fastest growing products for Klarna over the last few years, especially in the U.S. — has historically been distributed through Klarna’s merchant partners. However, the company has stated that its ambition is to move more of that lending volume from merchants over to its own direct acquisition channels, particularly for repeat customers, which is a common tactic in BNPL.

What Pagaya is alleging is that Klarna “distilled” Pagaya’s proprietary approach to subprime installment loan underwriting, during the course of the companies’ four-year partnership, replicated that underwriting capability internally, and then wrongly terminated its commercial relationship with Pagaya in order to cut it out of the business that they had built together.

This is a genuinely weird claim to make. Pagaya is not arguing that Klarna stole its subprime underwriting model, because — and this is the thing almost no one covering this case has gotten right — Klarna never had access to Pagaya's underwriting model. Klarna wasn't embedding Pagaya's model inside its decision engine. It was passing rejected applications over to Pagaya (without the end customer's knowledge) for Pagaya to underwrite using its proprietary model, inside Pagaya's decision engine.

So what Pagaya is actually arguing is that Klarna sat next to this arrangement for four years, carefully observed the outputs the model produced — the approve/decline decisions, the pricing, the subsequent repayment performance of every loan — and used those observations to backwards-engineer a competing model of its own.

I find this argument odd for two reasons.

First, as I've written before, I doubt Pagaya's model is as special as Pagaya thinks it is. My read has always been that Pagaya's business is fundamentally an arbitrage — it monetizes the gap between the sophistication and risk appetite of the lenders on one side of its network and the investors on the other. It does not possess some proprietary sorcery for pricing subprime credit that the rest of the industry hasn't cracked.

Second, even if you buy the premise that Pagaya has built a superior underwriting model to what Klarna had built (or was capable of building on its own), you still have to prove that Klarna misappropriated it. That seems like a tough thing to prove! Where exactly is the line between using your partner’s trade secret illegally and getting smarter by working next to them for four years? Because that second thing is legal, and it's what every partner in every commercial relationship does. If I learn to run a better lending program because I spent four years watching a great one operate next to me, did I steal something — or did I just learn?

That question does not have a clean answer. Pagaya is arguing its way around it "on information and belief," which is lawyer-speak for “we're inferring this from Klarna's suspicious-looking new capabilities and its bragging on earnings calls, but we don't have the evidence yet.” Maybe discovery turns up an internal message that amounts to Klarna executives saying, "let's just train the model on the Pagaya results." Or maybe it turns up four years of Klarna data scientists gradually getting better at their jobs. Those are very different cases, and right now Pagaya is asserting the first while only able to prove it had the opportunity for it.

To be fair, there are some other allegations in this lawsuit from Pagaya that have nothing to do with this model distillation argument. These other claims allege, among other things, that Klarna stopped paying its invoices from December 2025 through March 2026 and that it refused to cooperate on securitizations it was contractually obligated to support. These seem much more like run-of-the-mill contract breach claims, and we will have to see how they end up getting resolved.

For what it’s worth, Klarna says that all of Pagaya’s allegations are false and that it will defend itself vigorously in court.

#2: A Bunch of Debanking Stuff Happens

What happened?

The Department of Justice is investigating the big banks for debanking:

The investigation into “debanking” allegations is being handled by Jeanine Pirro, the US Attorney in Washington, said the person, who asked not to be identified because the probe is confidential. The Wall Street Journal reported earlier Wednesday that Pirro had sent subpoenas to lenders including JPMorgan Chase & Co. and Bank of America Corp.

Reputation risk is officially out as a supervisory tool:

The Federal Deposit Insurance Corporation (FDIC), the Office of the Comptroller of the Currency (OCC) and the Board of Governors of the Federal Reserve System jointly announced on June 2, 2026 that they had removed all references to “reputational risk” in an array of interagency documents, including guidance on lending, risk management, and cyber security. This move arrives at the same time that the OCC and FDIC’s joint final rule removing reputation risk from their supervisory programs, which was published in the Federal Register on April 10, 2026, goes into effect. The agencies’ statements emphasize that these revisions are meant to complement the new supervisory policy and that the removal of reputation risk from these documents is intended to help focus supervisory decisions on “material financial risks.”

Citizens Bank is losing deposits because of its business relationship with companies that operate Immigrations and Customs Enforcement (ICE) detention centers:

Entities including the Greater Boston Interfaith Organization, Newport Democratic Citizens Committee and Greater Cleveland Congregations, as well as two Democratic Massachusetts state representatives and one New Jersey city, have all either defunded their Citizens accounts or pledged to do so if the bank doesn’t reconsider financing CoreCivic and the GEO Group, which each operate numerous ICE detention centers.

GBIO announced Wednesday that it was withdrawing $2 million from its Citizens accounts in addition to the $1 million it withdrew early last month.

Also on Wednesday, Jersey City, New Jersey became the first municipality to publicly divest its Citizens account in response to the bank’s ICE ties. GEO Group operates Delaney Hall, an ICE detention facility in nearby Newark that’s gained national attention for its conditions and alleged human rights abuses.

And the CFPB has issued guidance encouraging lenders to consider immigration status when determining a prospective borrower’s ability to repay a loan:

The CFPB has now issued guidance on the consideration of immigration status in connection with the Regulation Z ability to repay requirements for credit cards and mortgage loans.

So what?

I want to analyze these specific stories, but first a little context.

There are lots of different types of risks in banking and, over many decades, bankers and regulators have developed a shared framework for how to think about those risks and a shared language for how to talk about those risks. When that shared framework and shared language are working well, regulators are able to subtly steer the banks they supervise away from areas that are excessively risky, without ever explicitly telling them what to do, completely scaring them away from taking any risks at all, or broadcasting signs of trouble that might negatively impact the market’s perception of the banks. This subtle steering happens, primarily, through supervisory exams. Bank examiners spend weeks embedded inside a bank each year, and the informal feedback they provide — a comment that lingers a little too long on a particular loan portfolio, a question about whether management really understands the risks in a certain business line — is often all it takes to change a bank's behavior. When examiners want to be a little less subtle, they can issue written findings (MRAs & MRIAs) that go directly to a bank’s board, downgrade a bank’s CAMELS rating, or enter into a memorandum of understanding (MOU) with a bank. These are more pointed actions that often have specific impacts on banks’ operations and balance sheets, but, crucially, they are confidential. If regulators want to dispense with subtlety altogether, they can publish guidance (interagency statements, "Dear CEO" letters, supervisory circulars, etc.) or pursue formal public enforcement actions (consent orders, cease and desist orders, civil money penalties, etc.) against specific banks.

The theory of debanking that I am most sympathetic to is that bank regulatory agencies have, in recent years, let political considerations overly influence their perceptions of the risks that banks were taking and allowed that influence to improperly seep into their confidential supervisory processes and published guidance, resulting in bank decisions — which accounts to open, which loans to make, which accounts to close, etc. — that weren’t entirely based on the individual banks’ business judgement.       

The key to this theory — what we might call “debanking with a lowercase d” — is that it doesn’t require a smoking gun. There doesn’t need to have been some centralized directive or explicit instruction to close specific consumer or commercial accounts (this is what we might refer to as “debanking with an uppercase D” or, as others call it, “Operation Chokepoint 2.0”), which is why this theory of debanking persists despite the lack of bombshells coming from the debanking investigations by the OCC and the House Committee on Financial Services (these investigations mostly uncovered anecdotal stories of account closures and internal bank policies on industry-specific risk management).

With that framework in mind, let's get back to our news stories.

I find the DOJ’s investigation into debanking to be strange and troubling. Per the Wall Street Journal, Pirro's office is reportedly examining whether the banks violated FIRREA, the post-S&L-crisis fraud statute that DOJ has long used to go after financial institutions. FIRREA is a fraud statute — to make it stick, you need underlying fraud that affects a financial institution. But a bank closing an account it considers too risky, which is what most "debanking" actually is, doesn't have an obvious fraud victim. Within the bounds of anti-discrimination laws like ECOA (which are fairly specific), declining or exiting a customer on risk grounds isn't a crime — and it's even harder to paint as fraud. Reaching for FIRREA looks less like a prosecutor following the evidence and more like one casting around for any statute that fits their boss’s desired outcome (this seems like the most likely explanation given President Trump’s personal grievances on this topic).

A better approach for addressing the subtle, behind-the-scenes form of debanking would be to remove as much subjectivity from the bank supervisory process as possible. If you’re worried about politics influencing the opinions of bank examiners, the obvious solution is to eliminate the ability for examiners to criticize banks on matters of opinion. This is, essentially, what the current administration is doing by eliminating reputational risk from their supervisory programs and interagency guidance.

I think this move makes sense. Reputation risk is a very tough, subjective thing to assess.

However, it’s important to note that it is still a valid risk, even if examiners won’t be allowed to opine on it anymore. Citizens Bank is getting a reminder of this right now. Jersey City alone is pulling roughly $265 million out of the bank, about $150 million of it already moved. And so far Citizens isn't blinking. It has defended the relationships and hasn't committed to ending them. And honestly? That’s good! It’s the system working the way it's supposed to. Reputation risk is being weighed and repriced in the open — by the depositors, municipalities, and institutions who actually bear it — instead of being whispered into a bank's ear during a confidential exam. 

I haven’t seen the CFPB’s new guidance on ability to pay and immigration status be referred to as “debanking,” nor have I seen anyone in the crypto community up in arms about it. However, if you look at the guidance in detail, it appears less like an honest effort to help lenders avoid risk and more like an attempt by the CFPB to discourage lenders from lending money to immigrants, legal or otherwise. The reason I say this is that it’s rigorously non-prescriptive, as the Consumer Finance Monitor article notes:

The guidance warns creditors of the risk of not complying with ability to repay requirements by not considering immigration status, particularly the status of an applicant not lawfully within the country, but fails to provide detail on situations in which a creditor would need to decline an application because of immigration status. The guidance, thus, increases the compliance burden of creditors. In particular, it appears that creditors may need to become experts in immigration law or engage counsel with such expertise.    

The practical effect of this guidance will be that lenders will be less likely to lend to anyone they think might be an illegal immigrant, which will, of course, cause legal immigrants (and, likely, some U.S. citizens) to be denied credit.

Interestingly, under the last administration, the CFPB and the DOJ issued a joint statement regarding the potential civil rights implications of a creditor’s consideration of an individual’s immigration status under ECOA, which was, at the time, criticized as an attempt to scare lenders into approving immigrants for credit, regardless of their legal status.

I’m not sure what term to use to describe that — it’s basically the inverse of debanking … “compelled banking” maybe? — but I do find it interesting that neither the last administration nor the current one seems content to simply let banks make their own decisions.

#3: Zelle Announces a Stablecoin

What happened?

EWS announced its forthcoming stablecoin:

Early Warning Services (EWS), the parent of the Zelle payments app, announced the launch of its ZLUSD stablecoin, which is intended to be used for cross border remittances. While the stablecoin is live, the Zelle app will only start supporting stablecoin payments by the end of the year, with India as the first friends and family corridor.

The announcement was a little thin on details, such as which blockchain will host the stablecoin and technically who is the issuer, because it doesn’t have to be EWS.

So what?

From what I understand, EWS is going to extend Zelle internationally first — U.S. consumers sending to friends and family in India, with the reverse corridor presumably to follow — and the stablecoin will come online later as the rail underneath.

My quick take: This is smart. EWS isn't positioning ZLUSD as a user-facing product; it's the infrastructure that might power one down the road, assuming EWS can win users in a brutally competitive remittance market.

It’s a sharp contrast with The Clearing House’s strategy to build a tokenized deposit network to keep deposits inside the bank perimeter, which I wrote about last week. Given that TCH and EWS are both owned by an overlapping collection of the same big U.S. banks, I think it suggests that those banks believe that the biggest crypto-related threat to their deposit franchises is in corporate payments and treasury (where TCH is aiming) versus consumer payments (which is where EWS is focused).

Sponsored by Ocrolus

Cash flow tells you what a merchant earned. It doesn't tell you how many other lenders they've approached this month, or how their margins hold up against similar businesses.

Ocrolus Intelligence changes that.

Built on a network processing hundreds of thousands of small business applications a month, it surfaces loan inquiry signals and peer benchmarking tied to NAICS classification directly inside the underwriting workflow.

The median SMB appears 1-7x in a 30-day window; knowing when a merchant is well outside that range changes the decision before it's made, not after.

For SMB lenders trying to move faster without sacrificing rigor, the gap was never the cash flow data. It was the context around it.

See how Ocrolus Intelligence works

2 READING RECOMMENDATIONS

#1: It's All Debt To Me (by Kate Elengold) 📚

I found this exploration of how and why the law treats different forms of consumer debt so differently to be extremely thought provoking. I’d highly recommend giving it a read!

#2: REPORT: The Enterprise AI Operating Model Playbook. (by Simon Taylor, Fintech Brainfood) 📚

Simon is far more immersed in AI than I am, so I appreciate his dispatches from the future. This report is especially dense with insights.

*Bonus: Agents in AML: A Production Playbook from the Investigator's Seat (by Oscilar)📚

Examiner scrutiny doesn't grade on plausibility. Seth Sattler, Head of Compliance/AML at Oscilar, lays out what it takes to build an AML agent that survives SOPs, model risk validation, and the questions examiners ask months after deployment. Want to make AML agents work inside your compliance program? Get the playbook from the practitioner's side.

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

1 QUESTION FROM THE FINTECH TAKES NETWORK

There are a TON of interesting questions being asked in the Fintech Takes Network. I’ll share one question, sourced from the Network, each week. However, if you’d like to join the conversation, please apply to join the Fintech Takes Network. 

I get why Visa, Mastercard, and Stripe are considering teaming up to launch a stablecoin, but I don’t understand why Coinbase is considering joining them. Circle is stuck in vassalage to Coinbase for quite a while and that relationship is enormously profitable to Coinbase.

Why is Coinbase looking beyond USDC?

If you have any thoughts on this question, reply to this email or DM me in the Fintech Takes Network!

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

— Alex