In partnership with

Happy Wednesday, Fintech Listeners!

I’m not against advertising. And I’m not against subscription pricing.

I would like to make that very clear, right off the top.

My salary is paid for by our wonderful sponsors, who advertise in the newsletter and podcast and work with me on sponsored content. I think this model works extremely well. And ditto for subscriptions! We don’t put anything at Fintech Takes behind a paywall, but it’s the right decision for many content creators and I myself subscribe to many wonderful newsletters.  

All that said, I’m never going to buy a BMW as long as I live.

The company infamously tried to charge a monthly subscription for access to its vehicles’ heated seats feature, which is insane. It backed down, after enormous public backlash. But now it’s back on the same old bullshit. In partnership with Sony, BMW pushed an ad to its vehicles (2020 or newer) for the new Spider-Man movie. An ad … on the dashboard displays of cars purchased over the last six years.

There’s nothing wrong with subscriptions or ads. However, there is, obviously, something wrong with rapaciously layering both business models on top of your existing business model (selling your cars for a large upfront sums of money) and not telling your customers about it beforehand.

It’s also stupid and self-defeating. BMW sells great cars and has a great brand. That brand is worth more than whatever short-term benefits these pricing gambits are generating.

Just make good cars and sell them at a fair price! This isn’t hard!

Anyway, today’s podcast contains a few good rants, so I figured this was a good energy to start today’s newsletter with.

— Alex

Was this email forwarded to you?

Subscribe Here

Sponsored by Stitch

A $40 billion syndicated loan is also a $40 billion reconciliation problem.

More than 30 institutions now hold pieces of SoftBank’s OpenAI bridge loan.

Each lender has to book, service, and report its position while reconciling its numbers against the agent bank’s record.

If the facility is extended, refinanced, or amended, the complexity multiplies.

Operating systems like Stitch keeps the lending lifecycle in one connected system, from origination and credit decisioning through disbursement, servicing, collections, and reporting.

It supports consumer lending, SME finance, microfinance, and BNPL programs, with workflows built to meet local regulatory requirements.

The risk is supposed to be distributed; the reconciliation burden shouldn’t be.

See the stress test

3 BIG IDEAS FROM THE PODCAST

In our latest episode of Fintech Recap, Jason Mikula joined me to summarize a story we almost cut from the episode’s outline for being too tangled. We also chat about BaaS (without stranding ourselves on BaaS Island!), the second-order effects of bank charters being so available, and the Fed’s impossible master accounts problem.

Tune in for the full conversation here

And read below for my three big ideas...

#1: The Corporate Card Loophole

Corporate cards work on a basic assumption. A bank does its diligence on a company as opposed to every employee and contractor who gets handed a card under that company's name.

Jason's reporting on ALT5 Sigma's Canadian subsidiary Mswipe (a middleware-style payment card issuing platform that also operates under the name Stradacarte) shows what happens when we assume.

(Editor’s Note — My dad said the “makes an ass out of you and me” aphorism to me and my brother CONSTANTLY when we were kids and made the mistake of sharing a bad assumption with him.)

There is a huge demand in the market for so-called “No-KYC Cards” and corporate card programs offer an easy-to-exploit loophole for creating them because once a company has been onboarded to a program, cards can be created for individual employees and contractors, usually without any additional verification of those individuals’ information or if they have a legitimate relationship with the company.

In this case, the infrastructure provider exploiting this loophole isn’t some anonymous shell. Mswipe is a subsidiary of ALT5 Sigma, the same crypto treasury company that took $1.5 billion from World Liberty Financial and watched its token collapse from ~25 cents to a nickel.

Jason’s reporting raises a bigger question than ALT5 Sigma and Mswipe, though. How many other corporate card programs are riding on the same faulty assumption?

In addition to his excellent reporting on ALT5 Sigma and Mswipe, Jason also wrote a broader investigatory piece on no-KYC cards and this corporate card issuing loophole, which I would encourage y’all to read!

#2: Make it BaaS (but Hold the Middleman)

Increase, the BaaS middleware platform, just provided another data point in my ongoing question about the viability of the banking-as-a-service market.

As I documented on Monday, Increase launched in 2020 selling developer-first BaaS middleware through partner banks like Blue Ridge and First Internet, with Ramp, Stripe, and Gusto as customers. Buckley tried buying his own bank once already in 2022/2023 and got rebuffed by the FDIC. In 2025, with the charter window wide open, he picked up a controlling stake in Twin City Bancorp, a tiny community bank in Longview, Washington, and got the FDIC's non-objection the second time around.

After acquiring Twin City, Buckley told Techcrunch that the bank wasn’t going to get into sponsor banking because sponsor banking requires specialized capabilities and expertise that Twin City didn't have.

A year later, Twin City Bank is now Increase Bank. Apparently, it is now specialized.

There’s a metaphor I like to use for what Increase has built: A BaaS Warp Core, which I define as the fusing of a bank charter with developer-first tech infrastructure. It’s the financial services equivalent of combining matter and antimatter, two things that are powerful when brought together, but also extremely volatile (which is exactly why regulators have historically been wary about it).

Jason pushed the root cause even further back beyond the banks themselves, to the core banking providers, FIS, Fiserv, and Jack Henry, which didn’t modernize fast enough to meet community banks’ needs for enabling BaaS.

Middleware emerged as a workaround for two problems stacked on top of each other: A fifteen-year bank charter drought following the 2008 crisis (which limited overall supply), and core providers too slow and expensive to innovate (which limited the ability of community banks to meet market demand for BaaS directly).

Today, this middleware workaround is starting to feel like borrowed time now that both underlying constraints are loosening.

Enough BaaS banks now have modern ledgers and core systems (either purchased from newer core providers or built internally) that FIS/Fiserv/Jack Henry are not the blockers that they were. And bank charters are now very available, which isn’t great for the BaaS Warp Cores (Increase, Column, Lead, etc.) but is (I would argue) worse for the BaaS middleware platforms that don’t have a bank charter.

#3: Threading the Credibility Needle

Regulatory credibility isn't something you build in the moment a crisis hits. It's something you build up through every decision you make in the years prior, and it’s what you draw down when you respond to a crisis and ask the market to trust you.

The reason Jason and I were talking about this is because the current crop of federal regulators seem, at least from my vantage point, to be pursuing policies that risk undermining that credibility.

Take the OCC as an example. The agency was founded during the Civil War. It is incredibly old. Its credibility with banks and with the market is based, in large part, on the culture of prudence and independence that it has nurtured over the last 160+ years.

The current leadership at the OCC obviously knows why the agency’s culture and market perception are so important. They understand why institutional credibility matters.

And you can see them trying to hang onto it, while also executing on the (often chaotic) deregulatory agenda of the Trump Administration.

They approved Erebor — a de novo bank with incredibly strong government connections — for a national bank charter, but set its tier 1 leverage ratio at 12%, which is extremely high even for a de novo bank. It’s almost as if they’re saying, “Yes, we’re taking risks to increase competition in the market, but we’re doing it carefully!”

World Liberty Financial (partner and lender of last resort to ALT5 Sigma!) will be the ultimate test case of whether the OCC’s leadership can thread the needle and find a way to hang onto its credibility.

As I’m sure you recall, World Liberty Financial applied for a national trust bank charter from the OCC, despite being owned by the President of the United States and his family and a member of the UAE royal family. Now, on the one hand, World Liberty Financial’s charter application was officially submitted in January and still hasn’t been approved. This gives the impression that the OCC is seriously deliberating on it and understands the unprecedented circumstances surrounding the application (to say nothing of the more concrete risks inherent to the business). On the other hand, it has been reported that the OCC is expected to approve the application soon, with unnamed former agency officials describing the approval as “all but guaranteed.”

Would approving World Liberty Financial’s charter application, but making it wait slightly longer than other recent applicants be enough to avoid seriously damaging the OCC’s credibility?

I guess we’ll see!

WHAT I'M LISTENING TO

#1: The next big breakthrough will be AIs learning on the job (Dwarkesh Podcast) 🎧

I had to use AI to dumb down this podcast about AI enough so that I could understand it … BUT! … once I understood it, the relevance to financial services became apparent. Some of the most idiosyncratic and sparse real-world training data in the world is found inside of banks and the AI labs need to find a way to export that training data and bring it back inside their walls without tripping any alarms.

#2: Why Robinhood Are Going All in on Tokenized Stocks and Prediction Markets (Tokenized) 🎧

An interesting episode that covered a lot of ground, even if the Robinhood portion made me want to rip off my own eyebrows.

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

— Alex