In partnership with

Happy Monday, Fintech Takers!

And May the 4th be with you (and especially with Cordelia!)

Lots to get to today, so let’s jump right in.

- Alex

P.S. — Every week, 95 million consumers in 6,500+ stores across the country walk past a largely untapped marketing channel.

For some reason (probably because folks are already in “wallets out” mode), fintech companies are seeing 49% more reach when they add in-store ads to their marketing strategy.

I'm joining Daniel Murray of The Marketing Millennials next week in a virtual event to explore why this channel is landing for fintech companies, and how to build a repeatable acquisition playbook around it.

If your marketing teams are looking for new channels, please send them our way!

Was this email forwarded to you?

Subscribe Here

Sponsored by Fundbox

B2B platforms that want to offer SMBs capital have two options.

The first: build embedded lending in-house. That means state licensing, dedicated loan capital, a risk team, and years of development before you originate a single dollar. Most platforms doing that math decide not to build.

The second: find infrastructure that keeps up. Fundbox's embedded capital infrastructure gives B2B platforms full-stack, white-label, embedded credit products, with underwriting, compliance, and capital all handled.

The platform owns the customer. Fundbox takes care of everything else.

Case in point: Relay wanted term loans inside their banking platform. They went from kickoff to launch in under eight weeks.

For platforms debating their two options, Fundbox has made the math pretty simple.

See how Relay did it

Toll Gate by Henri Rousseau.

3 FINTECH NEWS STORIES

#1: Please Let This Be The End

What happened?

The U.S. Senate may finally be in a position to pass the Clarity Act, based on some newly proposed language on the issue of yield-bearing payment stablecoins:

The new section of proposed Digital Asset Market Clarity Act text released Friday revealed that the compromise hashed out by U.S. Senators Thom Tillis (R-N.C.) and Angela Alsobrooks (D-Md.) would ban stablecoin issuers from offering yield based on just holding stablecoin reserves. It contends that "depository institutions provide financial services that are integral to the strength of the American economy," and stablecoin issuers offering similar services "may inhibit" these institutions.

In its legalese, the new text reads, "No covered party shall, directly or indirectly, pay any form of interest or yield (whether in cash, tokens, or other consideration) to a restricted recipient — (A) solely in connection with the holding of such restricted recipient's payment stablecoins; or (B) on a payment stablecoin balance in a manner that is economically or functionally equivalent to the payment of interest or yield on an interest-bearing bank deposit."

This restriction does not apply to incentives "based on bona fide activities or bona fide transactions" that are different from yield generated by interest-bearing bank deposits, the text said, maintaining an approach to rewards that's similar to what financial firms offer on credit card activity. The restriction does apply to loyalty programs or similar efforts. 

So what?

I’m so tired and I just want this to be over. The endless wrangling over an issue that probably doesn’t matter that much in the grand scheme of things has become unbearable.

I mean, just look at that legal language! 

No interest or yield in connection with holding a stablecoin … but if it’s based on activities or transactions, then it’s OK … except if those activities are tied to loyalty programs or subscriptions that function like yield on deposits.  

Brendan Pedersen, the reporter who broke the news of this new compromise, does his best to summarize it here:

This is confusing! What does “resemble yield” mean? What options does crypto have, exactly? Would, for example, a subscription service that rebates fees only to users with an average daily balance greater than $X be considered a loyalty program or a "yield-equivalent" on a deposit account?

I don’t know. No one knows. If the Clarity Act is passed with this compromise included, this question (and many many others) would need to be answered by federal regulators, who would be responsible for writing the implementing regulations. However, such regulations would be immediately challenged in court, given the vagueness of the bill's language and the diminished deference courts now show to regulators..

But will it pass?

Coinbase is apparently onboard:

"Mark it up," Coinbase CEO Brian Armstrong wrote in a posting on social media site X. His company had been at the center of the talks and potentially had the most to lose from restrictions on stablecoin rewards.

Coinbase's chief legal officer, Paul Grewal, said in a separate post that this language "preserves activity-based rewards tied to real participation on crypto platforms and networks, which is what the bank lobby said they wanted," adding that "we’re focused on getting a bill done and are satisfied that this language should not be the basis of any objection."  

This full-throated support, coming on the heels of Coinbase’s persistent intransigence on this issue, is likely to give the banking lobby pause. However, at this point, it may be too difficult for them to stop. This seems like a compromise that is specifically designed to drive a wedge between community banks (which are freaked out by the specific threat of deposit displacement) and the big national banks (which seem more concerned about allowing any new forms of competition into the market).

And that brings me to a question that I hadn’t thought to ask until just now: Where is the Electronic Payments Coalition on this issue?    

The EPC is a trade association representing Visa, Mastercard, and the big credit card issuers. It exists, primarily, to oppose any legislative or regulatory threat to the credit card market, such as the Credit Card Competition Act.

If yield-bearing payment stablecoins were the all-encompassing threat that the banks have claimed that they are, why hasn’t this latest compromise — restricting stablecoin yield on the holding of funds but allowing stablecoin rewards on payments activity — triggered an angry response from the EPC? I mean, the compromise explicitly allows for cashback on stablecoin transactions! Isn’t that a massive threat to credit card issuers?!? 

Apparently not.

My guess as to why the EPC isn’t interested in this issue is that credit card issuers are confident in the value their product provides. They know that consumers LOVE credit card rewards and they know that the revenue that pays for those rewards (annual fees, late fees, interest from revolvers, and interchange fees reluctantly paid by merchants) is stable and sufficiently diversified. They rightly believe that stablecoins are not a threat to their business (and may actually be a boost).

Objections to yield-bearing payment stablecoins tell you a lot about the confidence that the companies lodging the objections have in the value of their own products.

#2: OpenAI is Ushered Inside the Banking Perimeter

What happened?

Customers Bank revealed, in dramatic fashion, a new partnership with OpenAI:

Last Friday, about one-third of the way into the bank's first-quarter earnings call, Customers CEO Sam Sidhu revealed to analysts that the prepared remarks he had delivered just a few minutes prior were made by his "AI clone." He said the action "may be a first in the history of public company earnings calls."

"The execution of this call itself is a live demonstration of what we mean when we say AI is not an experiment at Customers Bank," Sidhu said. "We will be using it to transform our company."

Customers Bank is partnering with OpenAI to deploy artificial intelligence across its commercial banking business, a move that will position it to be "an AI native bank," the company announced … The multiyear collaboration between the West Reading, Pennsylvania-based regional bank and the creator of ChatGPT will focus on end-to-end workflow management in lending, deposits and payments, Customers said in a press release. The rollout involves placing OpenAI technical teams onsite at Customers to build custom AI capabilities, such as AI-enabled document collection, digital account onboarding, post-closing portfolio and collateral monitoring and AI-driven risk tools.

So what?

I’m not shocked that Customers was the one to take this step. It has always been willing to take innovation-flavored risks. Customers was one of the earliest and most aggressive banks to move into crypto. It operates a deposits and tokenized-payments franchise for 100+ digital-asset firms (Coinbase, Circle, Paxos, etc.), which has been a significant source of growth and profitability for the bank (as well as a source of compliance trouble … the bank got hit with an enforcement action from the Fed in 2024 over AML and sanctions deficiencies tied to its digital-asset business).

This most recent earnings call, on which the AI clone was used, was actually the very first earnings call that Sam Sidhu had done since taking over the Customers Bancorp CEO role from his father Jay at the beginning of this year.

That is one hell of a way to introduce yourself to investors, and it leaves me with a few questions:

• What does Customers Bank get out of this? The bank laid out some very specific goals: an improvement in its efficiency ratio from 49% to the low 40s, a reduction in the time it takes to close a commercial loan from 30-45 days down to 7, and reducing commercial deposit account opening from a day down to less than 20 minutes.

• What does OpenAI get out of this? OpenAI is deploying some of its engineers to sit inside Customers Bank and co-develop AI solutions for them. Why is it doing that? Because, in exchange, it gets years of supervised access to a chartered commercial bank's lending and onboarding workflows, a case study OpenAI can hand to the next bank's vendor-risk team, bank-specific domain fluency that's hard to acquire any other way, and (I’m guessing) deal terms permitting OpenAI to package up the co-developed tools and sell them to Customers' peer banks. Given OpenAI’s focus on growing its enterprise revenue, this arrangement seems like a no-brainer.

• How does this relationship fit into regulators' traditional third-party risk management frameworks?Interagency TPRM Guidance from the OCC, FDIC, and Fed was not built for a world where the vendor's engineers sit inside the bank. OpenAI engineers will be operating inside Customers' lending, deposit, and payments businesses, co-developing tools the bank will run on, and OpenAI likely retains the right to resell those tools to Customers' competitors. Are OpenAI engineers covered by Customers' BSA/AML and fair-lending programs when they're touching a credit decisioning workflow? Whose model is the agentic system when both sides built it? Is OpenAI a significant service provider under the BSCA?

• What’s the long-term impact of this deal? The most obvious outcome is that OpenAI uses Customers Bank to build a productized "commercial banking harness." Basically a vertical agentic AI system, pre-trained on real bank workflows, that OpenAI sells to every other commercial bank in the country, with Customers as the marquee reference customer. The less obvious path is that once you've built the agentic system that runs commercial lending, onboarding, and payments operations inside a bank, you've also built most of what you'd need to be a bank, or at least to be the customer-facing intelligence layer between businesses and the banks they use. I wouldn’t say that’s the most likely outcome from this partnership, but it is definitely a possibility that this deal opens up.

#3:How Much Risk Will Erebor End Up Taking?

What happened?

Erebor Bank filed its first call report, seven weeks after opening its unadorned virtual doors, reporting nearly $1.1 billion in deposits:

So what?

This news has been celebrated on fintech VC twitter as unprecedented growth for a de novo bank or fintech neobank. I won’t comment on the veracity of that claim, although it does seem entirely possible. A billion dollars in seven weeks is FAST.

What I will say is that, in banking, the rapid accumulation of deposits is universally seen as a risky thing by regulators because it is highly correlated with bank failures. This phenomenon has been studied extensively by the FDIC. Acquiring lots of deposits quickly usually comes with a cost, either monetary (i.e., paying higher rates) or risk (i.e., taking on higher-risk customers who other banks don’t want) or both. And if banks are willing to absorb those costs, it’s usually because they have a plan to deploy that liquidity to generate above-market returns, which also often entails a lot of risk. The combination of risky loans + expensive, non-sticky deposits frequently results in bank failures.

Now, if you wanted to defend Erebor, you could point out that unlike the most prominent bank failure in recent history (Silicon Valley Bank), most of Erebor’s deposits are insured. While roughly 90% of SVB’s deposits were uninsured before the bank failed, Erebor reports that only 23% of its deposits are uninsured. This difference matters. Insured depositors are much less likely to run, assuming that they know they are insured.

However, if we’re going to use SVB as our counter-example for Erebor (and I think it makes sense to do so given that they are going after roughly the same customer base), then we should also talk about the composition of Erebor’s deposits versus SVB’s deposits.

Prior to its failure, SVB's deposit base was 46% non-interest-bearing demand deposit accounts. These were operating accounts for startups and startup founders that cost the bank nothing. Erebor's deposit mix is 12% non-interest-bearing and 88% interest-bearing, which means Erebor's cost of funds is significantly higher than SVB’s was.

And what is Erebor doing with all of these expensive interest-bearing deposits? 

Nothing, yet. 

The call report shows zero loans on the bank's balance sheet, zero loan commitments outstanding, and 83% of total assets sitting at the Federal Reserve earning IOER. As currently structured, Erebor is functionally a money market fund with FDIC insurance. It earns a thin spread between Fed reserves and what it pays its depositors, and right now that spread is essentially zero. The bank reported a $6 million net loss in its first 52 days, and salaries alone ($5.6M) exceeded total revenue ($4.6M).

The bigger issue, though, is what the lending side will look like once it spins up. Palmer Luckey has said publicly that Erebor will run "the most conservative loan-to-deposit ratios of any bank in history," with early indications targeting around 50%, meaning the bank intends to lend out only half of every deposit dollar it takes in. At the same time, Erebor's stated mission is to serve early-stage startups, defense companies, and crypto-adjacent operators that other banks won't touch (i.e., higher-risk borrowers, often with non-standard collateral).

This is, ironically, very similar to SVB, which had an even lower loan-to-deposit ratio when it failed in 2023 than what Erebor is aiming for in 2026. That didn't mean SVB wasn't taking risks with the money it wasn't loaning out. As we know, it was, and that's what ended up sinking it.

Erebor seems to be betting that it can avoid the same fate by keeping its ratio of insured-to-uninsured deposits high, keeping its loan-to-deposit ratio low, and not doing anything stupid with the money that it has left over. 

It’s a sound idea, conceptually. However, in practice, there’s a reason why regulators worry about banks with this profile. The temptation to eventually do something stupid with the money is very very strong. 

Sponsored by MX

In this Data Takes spotlight, MX digs into the distance between what consumers report and what their spending shows.

62% identify as living paycheck to paycheck.

MX platform data puts 51% spending 90% or more of their income each month.

The consumers who know where they stand are less likely to be stuck there, and those engaging with MX's financial insights tools are 21% less likely to be living paycheck to paycheck than their peers.

The closer you are to your customers' financial lives, the more leverage you have to move the needle.

Read the data

2 READING RECOMMENDATIONS

#1: Partner Banking is: a) Back b) Broken c) Obsolete d) All of the above (by Jason Henrichs, Fintech Business Weekly) 📚

I really enjoyed Mr. Henrichs’ guest essay in Mr. Mikula’s newsletter yesterday. BaaS is in a fascinating state of flux right now.

#2: The Coming FinTech Liquidity Supercycle (FT Partners & Blue Dot Investors) 📚

Some interesting stats in this report from FT Partners and Blue Dot Investors.

*Bonus: Fast Rails, Frozen Books: What Banks Are Missing in the Era of Faster Everything (with Rouzbeh Rotabi at Qolo and Kiah Haslett of Fintech Takes Banking) 💻

Instant payments. Real-time subaccount posting. Stablecoin settlement. Your commercial clients can get all of that somewhere ... the question is whether it's from you. Rouzbeh and Kiah dig into why banks are losing ground on treasury management and what fixing it requires. May 20th @ 12 pm ET (virtual); register here.

*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’m interested in digging into the history of auto-pay. Do you know of any good resources (articles, books, people to interview, etc.) on this topic?

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