In partnership with

Happy Monday, Fintech Takers!

I hope you had a good weekend and didn’t let the rising tide of AI doomerism get you down.

I’ll be honest with you: I find the whole “there’s a 10% chance that AI kills us all” thing pretty difficult to parse. I don’t trust the folks working at the frontier AI labs on this, because they seem pretty bad at making predictions about their own technologies and they have clear economic incentives to make their products seem as advanced as possible and to secure regulatory capture. I don’t trust the tech VCs who oppose any regulatory slowdown of AI because they are completely coin operated and have a clear economic incentive to protect the competitiveness of open weight models. And I don’t trust politicians, on either side of the aisle, because I don’t think they understand how AI works, but, despite that, are strongly motivated to turn AI into a wedge issue in order to win elections.

So … you know … it’s not great. Hopefully, the current freakout leads, long-term, to more clarity and objectivity, especially as it relates to public policy.

In the meantime, there was SO MUCH fintech news last week, which means there is plenty of meat on the bone for use to choose from in today’s newsletter.

Let’s get to it!

- Alex

P.S. — Most of us guess at how much friction is too much.

Spinwheel has actually studied this quantitatively and has data and insights into what level of friction in lending is helpful and what’s actually counterproductive.

Their CEO, Tomás Campos, joins me Sept 30 to break down the data. Join us live or RSVP for the recording.

Was this email forwarded to you?

Subscribe Here

Sponsored by Plaid

Welcome to a new installment of Uncovering the Credit Blind Spot, where each month, I'll pull on a different thread of the cash flow data lenders need to make smarter credit calls.

Many borrowers carry debt that never shows up on a bureau report.

A recent study conducted by Plaid and Datos Insights found that 29.1% of consumers surveyed used BNPL in the past year, most of it unreported to bureaus.

35% of lenders are seeing delinquencies rise, but only 26% are using cash flow data.

The debt is there, but lenders can't see it build in real time.

That blind spot affects who you approve up front, and how fast you act when a borrower slips.

Hear how LendingClub and Plaid are closing that visibility window with cash flow data (in conversation with Fintech Takes!).

Watch here

Arrangement with white jug, orange and book by Vilhelm Lundstrøm.

3 FINTECH NEWS STORIES

#1: We’re a Bank (but we think of ourselves as a payments company)

What happened?

Chime is buying one of its partner banks:

Chime … announced that it has entered into a definitive agreement to acquire Stride Bank, N.A. (“Stride”) for $590 million in cash. Stride is a nationally chartered bank that has been Chime's bank partner for more than seven years. Upon closing, Stride will become Chime Bank, N.A. and operate as a wholly owned subsidiary of Chime.

The transaction marks an important milestone in Chime’s evolution from industry challenger to category leader. Chime’s technology-driven, payments-led model has reshaped the industry and now helps more than 10 million Active Members make financial progress. The combination of Chime’s digital core, trusted brand, and primary account relationships with Stride’s national charter and bank infrastructure will create an end-to-end platform built for the AI era.

So what?

Over the years, Chime’s public relationship to the term “bank” has shifted quite a bit, depending on the context.

When speaking to customers, it has used the term “bank,” even to the point of getting in trouble for deceiving consumers in California. When speaking to private market investors in 2020 — when ZIRP was in full effect and investors were assigning insane multiples to anything that smelled, vaguely, of compounding growth — Chime specifically claimed that it was “more like a consumer software company than a bank.” And, earlier this year, when speaking to public market investors, the company said that it was “more of a when, not if” that it would become a bank.

That progression tells you more about the changing risks and value of being perceived, publicly, as a bank than it does about Chime’s actual corporate strategy over the last 14 years. That strategy is better articulated in the company’s S-1, which it filed with the SEC before going public last year:

Traditional banks rely on a net interest margin-based business model with nearly 70% of their revenue coming from customer deposits and lending. This approach works well for the most affluent customers with higher credit scores, who have high deposit balances and large borrowing needs, but is ineffective for everyday Americans, most of whom live paycheck-to-paycheck and often have more modest account balances and limited credit histories. We believe that the majority of Americans can be better served by a payments-based banking model rather than the net interest margin-based model used by traditional banks. Everyday Americans earning up to $100,000 annually are estimated to account for less than 35% of consumer deposits yet over 75% of debit card transaction volume. We are focused on serving everyday Americans with a banking model aligned with this financial reality.

To put that in simpler terms, Chime believes that the way that banks make most of their money (net interest margin or NIM) creates an incentive and, over time, a culture for how banks build and price products and service their customers. That culture tends to produce punitive fees and lazy, inertia-driven product roadmaps, which do not serve the interests of their customers, especially those with less than $100,000 in annual income.

So, we can think of Chime’s insistence on thinking of itself as a payments company as, in some senses, a way to keep it accountable to its members. Transactional business models are, arguably, more honest because they force companies to earn the customer every time they swipe. There's no trapped balance, no captive spread, no incentive to profit from inertia or punish mistakes.

However, the problem — and people have been critiquing Chime on this front for a long time — is that transactional payments businesses are not very compelling businesses; margins are thin and dependent on a regulatory advantage (the Durbin exemption) that is tightly constrained and frequently contested.

In response to this basic reality, Chime has been steadily trying to build out deeper, more sophisticated capabilities, on both sides of its balance sheet, to push it beyond simply being a Durbin-exempt debit interchange business.

And, from what I can tell, it’s done fairly well on this front.

Chime ended 2025 at 9.5M active members, and a large majority of them treat it as their primary account (defined as a member with recurring direct deposit or 15+ transactions a month). Paycheck direct deposit is the stickiest liability in consumer banking, and Chime has it at scale. Purchase volume ran at $34.4B in Q4 2025 and average revenue per active member has climbed past $257.

And while the company’s lending products, which focus on meeting members’ short-term liquidity needs, are newer, the growth and performance are impressive. MyPay (an earned wage access product) hit $4.5B of origination in Q2 2026 at a 0.9% loss rate, driving transaction profit to $73M — triple the prior year — on a 64% transaction margin, at a $400M+ revenue run rate. Instant Loans (a short-term unsecured lending product) grew 70% QoQ to $300M of originations and is exiting Q3 at a $100M+ annualized revenue run rate, with repeat-borrower loss rates running up to 50% lower.

And now Chime is buying Stride for $590M and, assuming it gets regulatory approval (which I’d bet that it will), the company will realize “more than $100 million in net synergies, driven by sponsor bank fee savings, expansion of lending products, and a significantly lower cost of funds.”

I’m going to go out on a limb and guess that a large part of that $100M in short-term synergies will specifically come from the expansion of the lending business and the reduction in funding costs (sponsor bank fees have probably already been compressed through negotiation given Chime’s scale). More importantly, much of Chime’s long-term stock price growth will hinge on its ability to build out the lending side of its business (while keeping the payments business healthy and growing … Chime has said that it’s planning to stay under $10B in assets for the foreseeable future).

In other words, Chime is — economically — becoming more and more like a bank, and this acquisition will accelerate that shift. I’m optimistic, but it remains to be seen how successful the company will be in preventing this new economic incentive from reshaping its culture and the way it serves its members.

#2: Nubank is Playing Checkers and Chess

What happened?

Nubank is launching in the U.S.:

Brazilian digital bank Nubank (NU.N), will start offering financial products in the U.S. on Thursday through bank partners, it announced in a securities filing.

The initial products will be high-yield savings accounts, credit cards and remittances through its U.S. bank partner, CEO David Velez and Nubank co-founder and CEO in the U.S. Cristina Junqueira said during an event ⁠in Miami.

Nu is partnering with FDIC-insured Lead Bank to provide deposit accounts with a 3.50% yield, as well as free international money transfers and a no-fee credit card with 1.5% cashback.

So what?

First, I want to add in a couple of additions to the reporting I shared above:

• The interest on the savings account goes up to 4.5% on balances up to $10k if you hold the credit card and use it three times a month. Similarly, the cash back rate on the card will go up to 2% if customers meet certain yet-unnamed conditions. Nubank is clearly designing around bundles, not product silos.

• The U.S. products are being launched with Lead Bank today, but the expectation is that they will migrate over to Nubank’s U.S. bank, which was conditionally approved by the OCC in January and expected to be fully armed and operational by next year.

• Transfers from the U.S. to Brazil, Mexico and Colombia (all countries where Nubank has large and established customer bases) will be free from day one.

Second, I continue to be incredibly impressed by both Nubank’s tactical discipline and strategic thinking. Looking at the details of its U.S. launch, it’s becoming increasingly clear to me that the company is playing both checkers and chess … and playing both at a high level.

Let’s start with checkers.

In every market that it enters, Nubank is very disciplined about not spending more money than it has to. In Brazil, the competitive opportunity was service and transparency against fee-laden incumbent banks, so Nubank won without offering rewards at all. In Mexico, the opportunity was yield in a thin-deposit market, so it led with above-market savings rates (13–15% APY). In the U.S., there is no fee-and-service opportunity left; the neobanks already seized it. And it’s a very competitive rate environment. So Nubank is aiming for parity on savings and card rewards, and is, instead, attempting to differentiate itself by focusing on specific underbanked segments that may already have a positive impression of its brand (e.g., Hispanic, thin-file, cross-border, etc.)

Now chess.

In every market Nubank goes after, it wins by actually banking people. It gets the necessary regulatory approvals to operate directly in country and it builds a real balance sheet, rather than just a rail sitting on top of someone else's. It’s a slower path, but the advantages created by this depth compound. Take international remittances for example. Because both the sender and the recipient are Nubank customers, an international transfer between markets that it operates in stops being a cross-border transaction and becomes an on-us ledger move (instant, free, fully visible, fully serviceable). No correspondent bank, no FX partner, no "your money is pending." This is why Nubank’s decision to focus, narrowly, on a few specific and underserved customer segments in the U.S. isn’t just tactically sound, but also strategically smart; they are the same segments that are likely to find free international remittances to Mexico, Colombia, and Brazil particularly appealing.

I think this unlocks a very exciting long-term opportunity.

Think about the cross-border money-movement market as it exists today: It's old, slow, and expensive (Western Union), fast and cheap but self-service (Wise), or really fast and cheap but you're-on-your-own (crypto). What nobody has built is the American Express of remittances — a service-and-relationship layer in a category where service should obviously matter (you're sending money to family) and yet has never existed. Nubank can build it, because it owns both ends of the corridor and has built a famously excellent customer service operation. Wise owns the money movement rail but not the endpoints. It pays into someone else's bank account, and its relationship with the recipient ends the instant the money lands. Nubank owns the endpoints. So Nubank can own the experience.

Again, this is a deliberately slow strategy. It sacrifices broad coverage (what Wise has built) for integration depth, unit economics, and customer trust and affinity. However, those latter attributes are the ingredients of a company that can be enormously successful for 100+ years, and that seems like the type of company that Nubank is trying to build.

#3: Good for Banks. Bad for Cores. Silent on BaaS.

What happened?

Some new proposed regulatory guidance just dropped:

The Federal Deposit Insurance Corporation, the Federal Reserve Board, the National Credit Union Administration, and the Office of the Comptroller of the Currency (collectively, the agencies) requested comment on proposed guidance to assist financial institutions with managing risks associated with third-party relationships.

The proposed guidance reflects the agencies’ supervisory experience and lessons learned from examining financial institutions’ third-party risk management practices. It is intended to assist banks and credit unions to better align and tailor their third-party risk management practices to the risks of individual third-party relationships. The proposed guidance focuses on a principles-based approach and, as with all supervisory guidance, is non-binding.

When finalized, the federal bank regulatory agencies plan to rescind existing third-party risk management guidance and replace it with the finalized guidance to promote consistency and prudent innovation in the banking industry.

Additionally, three of the agencies (not the NCUA) issued a joint statement on the relationship between community banks and their core providers and the Fed proposed a companion guide to the TPRM guidance for banks under $30B in assets:

The federal bank regulatory agencies are issuing a statement on community banks’ engagement with core service providers. The statement discusses certain factors the agencies will consider in making supervisory and enforcement decisions related to these core providers.

Also today, the Federal Reserve Board separately requested comment on a proposed third-party risk management guide specifically for Federal Reserve-supervised community banks, which is intended to serve as a companion document to the proposed guidance.

So what?

For banks, this is a good news/weird news situation.

First, the good news. This guidance (and the accompanying interagency statement and Fed guide) are clearly intended to help banks, particularly community banks, leverage third-party relationships to increase their competitiveness in the market.

They do this in three main ways:

1. Encouraging banks to innovate through third parties. This is a small point, but it’s worth making right off the top. The agencies are explicitly encouraging banks to view third parties as an avenue for innovation. This is in sharp contrast to the position of the previous administration, which tended to view these relationships primarily through a risk lens. Add in the fact that the NCUA is finally joining this shared framework (credit unions have been operating under a 2007 supervisory letter — written for a pre-cloud, pre-fintech world — for nearly two decades), and it’s fairly obvious that this guidance is intended to help put smaller financial institutions on better footing with their more technologically sophisticated competitors.

2. Tailoring over box-checking. The guidance moves from prescriptive checklists toward a principles-based approach where the intensity of your oversight is supposed to scale with the actual risk of the relationship. The agencies state pretty clearly that the existing TPRM guidance (released in 2023) incentivized check-the-box behavior and was received by banks so conservatively that it chilled partnerships with newer, innovative third parties.

3. More scrutiny on the cores. The interagency statement catalogs every core-provider grievance you've ever heard — opaque pricing, "back-billing" windows, undefined deconversion fees, clauses that block outside vendors from integrating — and then floats a real theory of liability: That certain core providers may qualify as institution-affiliated parties under the Federal Deposit Insurance Act, as "persons . . . who participate in the conduct of the affairs of an insured depository institution." Because cores are "integral to carrying out the business of banking," the agencies argue, they "may be held liable for the practices or violations of a CBO as an institution-affiliated party." In plain English: The regulators are reserving the right to reach through the bank and pursue the vendor directly, with the same enforcement tools they use on bank directors and officers. Spicy stuff!

Now let’s get to the weird part: None of this is a solution for complex bank-fintech partnerships.

The Fed's community-bank guide is explicitly "not intended for community banks with more complex business models or third-party relationship profiles, such as complex bank-fintech partnerships." And, believe it or not, a guide written for small banks never grapples with BaaS at all, despite small banks being the primary providers of BaaS historically. Governor Barr, the lone dissent in the Fed's 6-1 vote, said as much on the record: "many banks with complex business models are especially in need of guidance that better addresses their particular third-party risk management issues, which is not addressed in these proposals." And Governor Cook, who voted for the package, acknowledged this gap by asking commenters whether the agencies should specify "the allocation of responsibilities for consumer protection, record management, and anti-money laundering in bank-fintech partnerships."

Honestly, this guidance reads as though the agencies diagnosed the recent BaaS meltdown as “the cores were bad + the regulators were too prescriptive,” and then built the entire response around fixing exactly those two things.

That's an odd reading of history. Synapse wasn't a core system failure (though its necessity in the first place was due, in part, to the limitations of the cores), and it wasn't caused by an overly rigid checklist. It was a middleware layer and a partner bank where nobody owned the ledger reconciliation, nobody owned the end customer, and, ultimately, real people couldn't get to their own money.

Now, it’s possible that these challenges are purposefully being descoped from this TPRM guidance, so that other solutions, like industry-led standards (which I talked about on the podcast last week) can be leveraged instead. The proposed guidance does talk a little bit about this, as Jason Mikula noted in his newsletter on this topic yesterday:

One very notable addition, particularly following recent reporting the FDIC is working with banking and fintech trade groups to create a “standard setting organization,” is a section on the “Use of Co-ventures, Consortia, Standard-Setting Organizations, Consultants, and Other Third Parties that Provide Risk Management Services.”

The proposed guidance acknowledges banks can leverage “new arrangements” to manage third-party risk …

The proposed guidance argues that such approaches could “create new efficiencies, provide banking organizations additional leverage in conducting due diligence on, negotiating with, or monitoring third parties, and facilitate access to new technologies and strategic expertise.”

The proposed guidance also speaks to the potential for insurance and contractual indemnification to mitigate risk in third party relationships, but notes the need to assess a third party’s ability to meet any obligations and for banks to consider if they have the capacity to cover costs associated with pursuing, filing, disputing, or litigating a claim.

If this is, indeed, the approach that the agencies are taking on complex bank-fintech partnerships, it puts even more pressure on the FDIC’s standard-setting initiative to get it right.

Sponsored by Plaid

Lenders have spent years refining their credit models.

But Plaid's latest report (with Datos Insights) finds that better models haven't bought lenders more confidence.

75.5% report improved predictive performance, yet only 52.5% feel more confident lending on bureau data alone, and 24.5% feel less confident than a year ago.

That instinct turns out to be right.

Delinquencies are rising for 35% of lenders, but only 26% are using cash flow data (catching BNPL obligations that bureau files miss).

The model isn't the problem; the data feeding it is.

Read the report

2 READING RECOMMENDATIONS

#1:  Regulators Move To Rescind Post-Synapse Risk Guidance, Lay Groundwork For Fintech Standards (by Jason Mikula, Fintech Business Weekly) 📚

I linked to Jason’s story above, but I wanted to share it here as well. Really useful history on BaaS, TPRM, and how we got to where we are.

#2: How Technology, Competition and Rates Changed Core Deposits (by Kiah Haslett, Fintech Takes Banking) 📚

People toss out the term “core deposits” a lot without ever thinking critically about what it means or how it has changed as banking overall has changed.

Kiah remedies that in this piece.

*Bonus: It's Time to Start Thinking About Cash Flow Infrastructure (by me, with Nova Credit) 📚

Every technology has a moment where it transcends a collection of use cases to become infrastructure. CRMs crossed that threshold when Salesforce turned deal tracking into a system of record. Cash flow data is crossing it now: proven in underwriting, it spread to everywhere lenders use bureau data, and that sprawl demands an intelligence layer beneath the whole credit risk lifecycle. I wrote about the four questions lenders should be asking about architecture, cost, and compliance. Read it here.

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

1 QUESTION FROM FINITY

There are a TON of interesting questions being asked in Finity (our digital community for fintech and banking nerds). I’ll share one question, sourced from the community, each week. However, if you’d like to join the conversation, please apply to join!

Why did Chime buy an existing bank rather than apply for a de novo charter?

I have thoughts about this question, but I’d love to hear yours! Reply to this email or DM me in Finity!

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

— Alex