Quick favor: Our records indicate that you aren’t opening this email. But records can be wrong. Please click here if you’d like to remain subscribed to Fintech Takes. 

Happy Monday, Fintech Takers!

I hope you had a safe and enjoyable Fourth of July weekend! 

My family and I took advantage of cheaper flights over the weekend and went on a quick family vacation. We’re still in the midst of it, so please pardon me if I’m a bit slower than normal in replying to messages. (BTW — LEGOLAND California puts on one hell of a fireworks show.)

I’m also in the process of working through the Boston Celtics’ trade of Jaylen Brown, one of my all-time favorite players. I’m sure I’ll get over it, at some point, but for now I’ll just say that Brown’s greatest legacy in Boston is going to be his contributions off the court, which is a pretty remarkable statement about an NBA champion and a Finals MVP.

- Alex

P.S. — I got some wonderful suggestions for Tolkien-sourced fintech startup names from readers after last Friday’s newsletter. My favorite was from Raj Bajwa: “Moria,” for a lender that extends credit collateralized against mineral rights. That’s such a good idea that when I saw it I got legitimately angry that I hadn’t thought of it!   

P.P.S — For years, we've built systems to identify automated behavior and keep out the bots. But AI agents are changing all that. This creates a challenging set of questions around trust, fraud, KYC, AML, and liability — and nobody has all the answers yet. On July 29th, I'm joined by Ross Freiman-Mendel (Head of Product Growth, Persona), Yuliya Kazakevich (Head of Risk and Compliance, Lithic), and Drew Edmond (Partner, Glenbrook Partners) to think through what this shift means for financial services. Come think through it with us.

Was this email forwarded to you?

Where Nature's God Hath Wrought (1925) by William Wendt.

3 FINTECH NEWS STORIES

A new stablecoin was announced:

More than 140 companies, including Visa, Stripe, Mastercard, BlackRock and Coinbase have joined Open Standard to launch Open USD (OUSD), a new stablecoin that shares most of the earnings from its reserves.

Open Standard said businesses will be able to mint and redeem Open USD without fees or volume limits, while most of the income made by OUSD's reserves will be distributed to participating businesses after a small management fee. According to Open USD's website, the stablecoin is expected to launch later this year.

The FAQ section on the site says that companies that join Open Standard will use Open USD as a core payment asset within their products and services, receive technical and integration support and earn revenue based on the stablecoin's adoption. Open USD will be managed by an independent organization with governance shared among partner companies, instead of a single issuer in control.

Stablecoins are a utility. They always have been. They're payments infrastructure, and payments infrastructure has never been a good standalone business. To the extent stablecoin issuance has looked like a good standalone business, that's mostly an artifact of interest rates being relatively high for the last six years. 

The two companies that dominate stablecoins today dominate in very different ways. Tether is the choice of users who want to keep their money in U.S. dollars but, for various reasons, can’t. Circle is the choice of users who want to transact in full compliance with applicable regulations. Both have been vertically integrating on top of their stablecoins in recent years, precisely because they understand that issuance alone isn't where durable profits live.

What stablecoins do have is a rare combination of attributes. Plenty of payment rails are instant, or global, or low-cost, or easily programmable. Very few are all of those at once. The card networks might be the only other example.

Which brings us to the easiest read on this news: The companies that make money on payments, but not on stablecoins, are teaming up to democratize access to a common stablecoin that none of them control and that won't generate a tremendous amount of profit for any of them. 

The point isn't the stablecoin. The point is that a neutral, jointly-owned stablecoin lets each of them keep winning in the areas where they already win, while preventing the dominant stablecoin issuers from disrupting them. Put simply, this is everyone who isn't Circle or Tether getting together and running the big banks' consortium anti-disruption playbook.

Some things I'll be watching:

• Governance and execution will be the biggest challenges. As Christian Catalini — who lived through this exact problem at Libra — has pointed out, governance and execution are where these things sink or swim. And the details here are scarce. Which companies are on the board? How were they chosen? How long do they stay in control? Do smaller members get a vote? Do all of the members even fully understand what they signed up for? Are they even aware that they signed up?
• The central bankers have stayed quiet, so far. There is something slightly alarming about a group of large U.S. companies getting together and deciding, amongst themselves, to build the future of the global payments system. When Meta tried this, central bankers flipped out. That reaction — right or wrong — is what killed Libra. It's a mark of how far we've come, and how much political influence crypto has accumulated, that central bankers seem fine with this version. I'm assuming Stripe at least gave regulators a heads up, but unlike 2019, they probably didn’t have to.
• Central bank digital currencies (CBDCs) seem dead. They were never going to take root in the U.S. We fetishize privacy too much. But the deeper problem with CBDCs is that they're naturally not interoperable across borders, and we are not exactly living through a golden age of international cooperation. A corporate-run stablecoin that works across most countries on day one might genuinely be the more pragmatic path.
• Shouldn't the DOJ be at least a little curious about this? A bunch of companies that are normally fierce competitors are teaming up to reshape a foundational and fast-growing part of the global payments industry. At what point does a payments consortium of this size and composition become an antitrust question?
• Are we letting public, non-finance companies issue stablecoins now? As Amanda Fischer points out, this structure would seem to tiptoe around the GENIUS Act's restrictions on non-finance public companies issuing stablecoins. How exactly do those restrictions apply to a consortium in which public non-finance companies like Google are members but not issuers?

Now, let's talk about some specific companies:

• Stripe is playing the Meta role, and playing it far better than Meta did. Bridge's CEO is the temporary CEO of Open Standard. Tempo is supported on day one. Stripe has a much better reputation than Meta did when it attempted Libra, and Stripe is strategically smart about stuff like this. I'm sure Patrick and John have studied how Dee Hock pulled the original Visa consortium together and are taking those lessons to heart.
• Coinbase being involved is very interesting. Coinbase has Circle trapped in the 2026 business equivalent of a medieval vassalage arrangement, and I doubt it plans to end that anytime soon. But joining OUSD at the founding suggests Coinbase views this initiative as sufficiently threatening to Circle's long-term viability that it wants to hedge its bet. And if it didn't get in at the beginning, it wouldn't get the same control by joining later.
• Speaking of Circle, this obviously isn't great for them. The company’s stock price dropped sharply after the announcement. Here’s the essential question: Can Circle build out the Stripe side of its business (an integrated suite of payments capabilities) before Stripe can build out the Circle side of its business (a highly liquid and well-distributed payment stablecoin)? When Stripe's strategy for building its own Circle was to have Bridge issue its own stablecoin and grow adoption slowly, that was an even-ish contest. What Stripe seems to have realized is that building its own Circle is necessary but not profitable — so why go it alone? By switching to the consortium model, Stripe just fast-tracked the Circle side of its business. Circle can't run the same play in reverse. It needs to own the Stripe side, because that's where the money is.
• I think Tether will be fine. In the U.S., we talk about the unbanked a lot. Globally, the analogous population is the "undollared," and it's a massive market that Tether owns outright. A consortium of American payments companies isn’t a threat to Tether’s business.

Erebor, the brand-new and badly-named digital bank launched by Palmer Luckey, is reportedly in the process of raising additional capital, at double its last valuation:

Erebor Bank is in talks to raise money at a valuation of at least $8 billion, as deposits grow rapidly at the politically connected startup that received its bank charter only five months ago.

The bank’s deposit base now totals $4.05 billion, nearly four times the amount it held at the end of March, when the bank disclosed $1.1 billion to regulators, according to other people familiar with the matter, who asked not to be named discussing non-public information. The bank, which added nearly 400 customers in the past three months, expects to be profitable by the end of this year, the people said.

“Zero percent of Erebor’s deposit growth this quarter has come from my own companies,” Luckey said in a statement. “All the growth is from hundreds of new customers choosing Erebor because they want what we’ve built.”  

A large portion of the bank’s growth has come from hard technology companies and defense manufacturers. The bank is funding companies that are intent on “rebuilding America’s industrial base,” said Luckey. Erebor has expanded credit facilities for those firms, including for equipment financing and venture debt.

Every new Erebor story produces a very loud and excited reaction among tech folks on Twitter because A.) Erebor is a bank owned by a famous tech entrepreneur and built, explicitly, to serve the tech ecosystem, and B.) most tech folks don’t understand how banking works.

So, they look at a reported statement like, “Erebor grew its deposits to $4.05 billion, nearly 4x what it had three months ago,” and assume that it means that Erebor is a rocketship, bound for the stars.

Maybe!

But some nuance would, I think, be helpful.

According to the bank’s Q1 call report, it had roughly $1.7 billion in assets. Those assets came from investors (roughly $600 million) and from depositors ($1.1 billion). Based on Luckey’s statement to Bloomberg about the bank’s Q2 growth coming from companies he doesn’t own, I think we can assume that most of the deposits that were sourced in Q1 were sourced from his companies. According to the call report, the bank basically didn’t do anything with this money in Q1. It mostly just kept it in cash (it could have paid back all of its depositors with the cash it had on hand) and put the remainder in highly liquid securities.

Now Bloomberg is telling us that the bank has brought in an additional $3 billion in deposits in the last three months. These deposits come, according to Luckey, from hundreds of independent companies that have chosen to work with Erebor, though I think we can assume that Luckey’s extended network likely played a big role in influencing that choice.

So, what can we conclude from this information?

Well, essentially, Erebor is currently a pile of $600M in capital (maybe a bit less today considering the bank’s burn rate) that has demonstrated the ability to attract $1B in deposits from companies that its founder also owns and $3B in deposits from other customers who likely know and trust its founder and/or investors. Now Erebor is reportedly going to be increasing that pile of $600M to something bigger (the Bloomberg article didn’t specify how much it might raise) at roughly double the valuation that it last received.

I should stress at this point that it’s early. We don’t know what kind of bank Erebor will eventually become. However, with that said, the current situation is weird as hell.

An investment round that values Erebor at $8B would likely put its price to tangible book value (P/TBV) somewhere between 7x and 13x, depending on the amount of new capital raised. That’s significantly higher than the premiums that the most highly-valued public banks get (JPMorgan Chase, for example, usually trades between 2x and 3x).  

From my vantage point, there are two ways to explain that number:

1. Erebor’s investors don’t give a shit about getting a return on this investment. They just want to create certainty for all of their other investments. The bank’s existing backers include Lux Capital, 8VC, Andreessen Horowitz and Founders Fund. If the new investors coming in for the next round are also VC firms focused on tech startups, then perhaps the easiest way to think about Erebor is that it’s a low-risk, minimally-profitable utility that provides its investors’ portfolio companies with guaranteed access to regulated banking services, which they may perceive as highly valuable depending on how scared they are about potentially being debanked in the future.
2. Erebor’s investors do give a shit about getting a return on this investment and they see a path for the bank to become much more profitable than it is currently (Erebor wasn’t profitable in Q1 and Bloomberg reported that the bank expects to reach profitability by the end of the year, which implies that it was still unprofitable in Q2). The question is how will it accomplish this? Even though, as Bloomberg reports, Erebor has started lending money, that can’t be the way that the company is planning to earn its $8B valuation. First, Luckey has explicitly said that the bank is planning to maintain an extremely conservative loan-to-deposit ratio (50% or less). Second, and more importantly, even if Luckey changed his mind about that, banks have capital requirements. For every dollar that a bank wants to grow its assets, it is required by regulators to also hold a certain percentage in equity on its books. For Erebor, right now as a de novo bank, it has a minimum Tier 1 leverage ratio of 12%. That means that every incremental $1B of assets that the bank brings in requires roughly $120M of fresh equity. That’s an unusually high leverage ratio for a de novo bank (which, I’m guessing, reflects the OCC’s concern with Erebor’s novel business plan) but even at a lower ratio, this requirement would still be a huge drag on earnings. This is why banks don’t trade at a significant premium; they structurally can’t compound the way that software companies do. Thus, if Erebor’s investors believe that the bank will someday be worth a P/TBV of 7x – 13x, they must believe that the bank has a way to radically grow the non-interest income side of its business (Crypto? Stablecoins? BaaS? I really don’t know!)       

Finally, if we set the valuation aside, it’s worth pointing out that the timing of this reported capital raise probably isn’t random. 

A Tier 1 leverage ratio (unlike risk-based ratios) applies to all assets — including cash. Erebor doesn't get credit for being extremely conservative with its assets; $1 parked at the Fed consumes the same 12 cents of capital as $1 lent to a drone factory. 

The reported Q2 numbers — $4.05B of deposits plus $600M of equity — imply roughly $4.6B of total assets. Multiply that by 12% and you get $550M of required Tier 1 capital. And remember, Erebor raised roughly $635M from its investors. By the end of Q1, after operating losses, the call report showed $600M remaining. More has undoubtedly been burned in Q2. How close is Erebor's remaining equity to that $550M floor? How much more can it even grow in deposits without raising additional capital?

This is, again, where it’s dangerous to apply a tech mindset to banking. Tech investors are trained to hear, “we literally can’t keep up with customer demand unless we get more money,” as an unambiguously good thing. 

In banking, customer demand is both an asset and a liability, and not being able to keep up is rarely a good thing.

The Department of Education finalized the implementing regulations for the higher ed accountability provisions of last summer’s One Big Beautiful Bill Act:

Today, the U.S. Department of Education (the Department) announced a final rule establishing a long-overdue postsecondary education accountability framework. Under the new Student Tuition and Transparency System (STATS) and Earnings Accountability rule, undergraduate programs will be required to demonstrate that their graduates earn more than the typical high school diploma holder, and graduate programs will be required to demonstrate that their graduates earn more than the typical bachelor’s degree holder.  

If a program fails to show at least this modest financial return on investment for its graduates in two out of three consecutive award years, it will lose eligibility to participate in the federal Direct Loan program. After three years of consistently failing the earnings premium measure, the Department could also terminate eligibility for Title IV of the Higher Education Act (HEA), including Pell Grant eligibility, for all of an institution's low-earning outcome programs.

I know — education policy story, fintech newsletter. Bear with me. 

Federal student loans are the second-largest category of consumer debt in America ($1.7 trillion), and this rule is, functionally, the government deciding how to underwrite the programs that debt flows into. And it does so using a data workaround that will feel very familiar to anyone who has spent time in fair lending.

Some context. The government has been trying to police low-value college programs since the Obama administration, through a series of “Gainful Employment” rules built around a debt-to-earnings test, which is essentially a DTI check applied to degree programs instead of borrowers. These “Gainful Employment” rules haven’t produced the intended result (lowering the cost of higher education) for a number of different reasons, but one of the biggest is that they were toothless: Not a single program has ever lost eligibility under them.

STATS is the reboot. It scraps the debt-to-earnings metric and replaces it with a single earnings premium test: Do a program’s graduates out-earn a counterfactual person who never enrolled? 

Undergrad programs get benchmarked against high school graduates; grad programs against bachelor’s holders. Earnings come from actual IRS data, four years after completion. Fail two out of three years, lose access to federal loans. Keep failing, lose everything, Pell included. And it applies to everyone — for-profits, community colleges, Harvard’s masters programs — where the old rules (theoretically) had teeth only for the for-profit sector.

The underwriting logic here is one that anyone working in consumer lending knows well: The government shouldn’t originate loans into programs that demonstrably leave borrowers no better off. That’s the ability-to-repay doctrine, migrated from post-2008 mortgage lending into federal education finance.

However, there’s also a mechanical issue here that I find fascinating.

This whole apparatus runs on data the government mostly doesn’t have. Federal tax privacy law (IRC Section 6103) is one of the most airtight statutes in the U.S. code. The IRS won’t hand over tax records; the Department of Education will send over cohort lists and get back only median earnings figures, and only where the IRS can produce a statistically safe match. The result: the Department of Education lacks complete earnings data for roughly 76% of programs, and its impact analysis simply assumes the unmeasured three-quarters fail at the same rate as the measured quarter.

Interestingly, it’s the same mechanical weakness that has historically hamstrung post-origination fair lending analysis (disparate impact). ECOA bars lenders from collecting race on most credit applications, so disparate impact analysis runs on BISG — a probabilistic proxy built from surnames and geography — and enforcement consequences flow from the estimate rather than the measurement. 

In both cases, a privacy-protective law creates a data vacuum, the government fills it with an assumption, real consequences attach to the output, and the assumption is effectively unauditable.

The first real eligibility consequences from STATS won’t hit until 2028 at the earliest. Which gives us about two years to find out whether anyone forces the Department of Education to defend that missing-data assumption before colleges and universities start losing access to $1.7 trillion worth of lending.

Sponsored by MX

In this Data Takes spotlight, MX research puts a number on every bank's coming deposit problem: Millennials are up to 10x more willing than Baby Boomers to let AI move their money autonomously.

MX's research shows generational AI trust is already outpacing what most financial institutions have built for.

48% of Millennials and 46% of Gen Z trust AI to track and manage their finances, versus 32% of consumers overall.

More than half of both cohorts would hand off proactive reminders, spending breakdowns, customer support, and personalized recommendations to AI today.

When it comes to higher-stakes actions (prioritizing a bill, investment guidance, moving funds to a better savings/checking account without asking first), Millennials are 3–10x more willing than Boomers to say yes.

The window to get ahead is still open. MX's data tells you where.

2 READING RECOMMENDATIONS

Smart, nuanced writing in tokenized deposit networks. What a treat!

Nothing to do with fintech, but I enjoyed this piece immensely.

Rules can tell you the cheapest flight, but they can't tell you it's the wrong one. Unlike every technology wave before it, agentic AI may hit business software first, because convenience is a consumer export. Context, the thing agentic AI runs on, is not. Read my new deep dive.

*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. 

Can someone explain Robinhood’s new 7% DeFi earn offering to me? 

If yes, please remember to speak to me as if I were a small child, or a golden retriever.

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  

Join 2,444 other finance and fintech leaders in the Fintech Takes Network

The conversation doesn't have to stop here
Keep learning and connecting in the Fintech Takes Network

EVENTS  |  FEED  |  LIBRARY  |  DIRECTORY

Workweek Media Inc. 
1023 Springdale Road, STE 9E
Austin, TX 78721

Want to ruin my day? Unsubscribe.