- Fintech Brainfood
- Posts
- New FIDC rule proposed. Plus does every Neobank need to become a bank?
New FIDC rule proposed. Plus does every Neobank need to become a bank?
Probably yes. Revolt is. Nubank is. Who's next is a matter of timing and priorities. The new FDIC rule is smart but only a start.
Welcome to Fintech Brainfood, the weekly deep dive into Fintech news, events, and analysis. You can subscribe by hitting the button below, and you can get in touch by hitting reply to the email (or subscribing then replying)
Hey, Fintech Nerds đź‘‹
Huge week!
The FDIC announced its proposed rule on Fintech Accounts, designed to ensure the FDIC can always see who owns what funds. It wouldn’t solve Synapse/Evolve, but it would help in most other cases.
Meanwhile, as one bank exits BaaS, Brex launched embedded expense management. Embedded is the future, but you gotta do the hard yards on compliance.
Affirm is on Apple Pay; Monzo Flex is on Apple Pay, and JP Morgan looks set to run Apple Card.
That’s before we get to the 50bps rate cut.
This week I’ve been in Madrid ranting at banks, playing with Replit and Google NotebookLM and wondering if every Neobank should become a bank if rates don’t go back to zero.
PS. Huge congratulations to Nik Milanovic on raising Fund 2. It’s a hard time for Solo GPs. Good ones are rare. Nik is exceptional. You should fight to be an LP in anything he does.
Here's this week's Brainfood in summary
đź“Ł Rant: Does every Neobank need to become a bank?
đź’¸ 4 Fintech Companies:
đź‘€ Things to Know:
đź“š Good Read: The future of credit risk decisioning
If your email client clips some of this newsletter click below to see the rest
Weekly Rant đź“Ł
Does every Neobank need to become a bank?
Probably yes.
Interest rates are returning to a more "normal" level of between 2 and 4% for the foreseeable future. In that world, the reward for becoming a bank is significant revenue and profit in the medium term. I predict we'll see more Neobanks acquiring smaller banks as a way into the US market. The path there is incredibly hard, but the juice might, finally, be worth the squeeze.
Two years ago, I wrote a piece called "Do you really want to be a bank?" I argued that in markets like the US or Europe, it is better to avoid being a bank until you really need to.
A lot can change in two years.
Back then, I cited Revolut as an example of a company who had avoided being a bank in the UK (its largest market) which had enabled it to extend its product set much faster than competitors. This has driven it to over 45m users.
Today, Revolut is on a path to a bank licence in the UK, as well as one in Mexico and Lithuania.
What changed?
What's changed is that interest rates went from a long term average close to 0, to 5%+. This made the prize much greater.
What changed is that digital banks like Nubank, SoFi, Moneylion, and even Monzo started to become much more profitable in this rate environment.
Being a digital bank is an amazingly profitable endeavor if interest rates are high enough.
The prize for success in western markets has increased with rising rates, and changing investor sentiment
Regulation is harder than ever, but the way through is persistence over time
Being a bank doesn't make sense for the embedded finance universe, therefore embedded finance will bifurcate into those that own the workflow (User experience) and those that own the balance sheet and infrastructure (headless banks)
I started ranting about this after seeing this piece by Alex and Justin at a16z
1. Being a digital bank is great if interest rates are high.
Low cost + high profitability = win.
Nubank is the perfect case in point.
Low cost: They dramatically decreased the cost to acquire and serve customers from a benchmark in Brazil of to $7 to acquire and $9 per annum to serve.
+ Profitability: They operate in a market that's wildly profitable by default. The interest rate in Brazil is 10.5%, and Mexico's is 10.75%. This makes the "low-income" or unbanked population potentially much more profitable, especially if the cost of funds is low.
= win: Nubank has over 100 million customers, a return on equity in the high 20s, and has only just started expanding its product and market.
Low cost from technology. High profit from higher margin lending products than ZIRP-era markets. Consistent execution.
Their Return on Equity performance is also astonishingly good for a consumer portfolio in a low-income segment. Most of their larger competitors serve higher ROE segments (like affluent, wealth, and corporate). Larger banks will profit more per customer, but they can't profit from the same customers Nubank did.
In Western markets, large banks deliver a higher RPAC and ROE by offering all the products to all the people. For example, while a consumer book might have an ROE of 10%, commercial banking (SMBs) is closer to 20%, corporate 25%, and wealth can be as high as 80 or 90%.
The "universal" model means banks can attract a huge pool of consumer deposits from the mass market to their balance sheets. This unlocks the low cost of funding for activities in the more profitable parts of the bank.
This is how those giant banks can remain profitable while carrying the legacy of 1000s of branches, 100,000+ staff counts, and legacy IT infrastructure.
The bull case for Nubank is, what happens when it too, offers all the products, to all the people?
In fact, if all digital-first banks have similar low cost economics, what happens when they stick the landing on product extension?
Great questions.
But before we get there, we need to look at what's changed in my thinking since the original piece.
2. The prize for successfully getting a charter has improved dramatically in the west.
Getting a charter is still near impossible.
But we have exited the "ZIRP era" of long-term, near-zero interest rates. For the foreseeable future, trade wars, inflation, and a higher base rate are more likely realities than heading back to 0.25% Fed funds levels.
The current 10-year US Treasury rate is 3.8%.
This means the market believes that a bank can make at least ~3.5% by holding deposits at the Fed for the next decade.
That feels like one heck of a prize for showing up!
Today, Neobanks benefit from some of this yield because in a competitive market, banks share some of their deposit yield to attract deposits in the first place. Yet even in this scenario, the Neobank marketing to the customer and attracting that deposit is leaving money on the table by not having a charter.
Neobanks have recently enjoyed a windfall from higher rates. Reports and data from companies like Robinhood, Wise, and Revolut all highlight deposits as a core driver for their record growth.
As rates fall below 5%, banks will get less generous, and Neobanks will have added incentive to chase a higher share of that sweet, money-for-nothing you get at the Fed. (And in the UK, Revolut is already ahead of this with their recent banking license with restrictions approval)
Meanwhile on the other side of the balance sheet.
The cost of funding for lenders is skyrocketing.
Affirm just reported that their cost of funds is 7.7%, vs. an average of 2.59% across the banking industry as a whole (according to FDIC), h/t Jason Mikula for this find.
If you want to lend to a consumer or business, you need to get a supply of capital to lend in the first place. Where a bank just invents a new deposit as an asset on its ledger when it issues a loan, a non-bank has to get that supply of capital from somewhere, and those suppliers are raising prices.
Affirm could raise its prices, but in a competitive market, that's a bad idea. If they raise prices too much, the core value proposition of BNPL vs. credit cards starts to dilute.
Want to know what's lower cost than buying your capital from 3rd parties?
Deposits.
Paying a yield on deposits like 4% is much better than paying someone else 7.7%. The only catch is you have to be a bank to take deposits.
3. Getting a charter is still stupidly hard in the US
Banking regulation is hard. There's scrutiny for having an MTL, and then there's the direct scrutiny a bank receives, and the difference is a chasm.
As a bank, you're scrutinized by every regulator at the federal and state levels. You're directly accountable to regulators for an alphabet of laws, including BSCA, ECOA, TILA, FCRA, CARD, and hundreds of others.
Oh, and you're also the police of money. Banks interface directly with law enforcement and the government to implement anti-money laundering, sanctions, and counterterrorism financing policies.
Sure, other participants do too. But not to nearly the same level.
That's why getting a charter in the first place is difficult. It's such a massive step up from almost any other angle. Unless you've been at a bank and had to live this reality, it's genuinely hard to explain how tricky this is. It's all achievable; there just isn't an easy path there. It's complex, and it takes time.
That's why:
De-novo charters are difficult. I'm fairly convinced the incumbent US regulatory bodies want fewer banks rather than more. So the solution is simple:
Close the front door.
Don't issue new charters.
Let the smaller banks fail or get bigger through M&A.
Small banks are risky and cause headaches for the FDIC, but big banks are stable (except for the global financial crisis, when the regulators hit them with plenty of fines and now Basel III. They must figure that's a pretty good punch on the nose).
The biggest problem with smaller banks is that there are just so many of them. And it doesn't matter if some are incredibly well run, with world-class teams, high return on equity, and ultra-conservative balance sheets. If you're a federal regulator, it's whack-a-mole with which bank in which state is doing the next risky thing.
While there's a whole rant about why that's a fundamentally broken assumption, it's also unlikely to change for the foreseeable future.
Regulation works on legal precedent, not first principles. That's why we don't have a fancy UK or Singapore-style path to becoming a bank for innovators in the US.
What about acquiring a bank?
But folks like SoFI and LendingClub have all acquired charters, and it appears to be working.
LendingClub acquired RadiusBank in 2020 for $185m to reduce its cost of funds when lending to its $4.9m customers. For context, this is a veteran of the early 2010s "P2P lending" cycle of companies who appeared with massive valuations only to plummet later as the market judged them as a "lending business."
They've now originated $90bn in loans, servicing high-FICO but high-debt consumers looking to consolidate their cards. This customer base is returning, with an average net promoter score of 80. They can do this with a cost of funds closer to 4.7%. While this is higher than many older banks, it's much lower than Affirm's 7.7%. This is a meaningful difference in their ability to make a profit from lending.
With 8.7m members, they delivered $567m quarterly revenue, adjusted EBITDA of $138m, and are now GAAP profitable (at over $7m). I couldn't find the cost of funds in their recent annual report, but they are offering 4.5% APY to consumers to attract deposits. They'd likely be in the ~5% ballpark when accounting for other costs and risks.
They're playing a pricing game to win consumers, grow their book, and attract deposits rapidly, and it's working. It's also a much lower funding cost than the 7.7% a mature, scaled lender like Affirm is having to take.
What's perhaps more impressive is their continued growth story. They're growing revenue, profits, and market share quarter after quarter. YoY, both companies are in the ~20% top-line revenue growth range. If they can sustain that momentum, that's huge.
For Neobanks the prize is becoming too great to ignore.
4. Will every Neobank become a digital bank?
Most will eventually.
I think some folks are concerned about being the next Varo, who by all accounts, have had a rough time getting their DeNovo charter and hitting profit. But the argument here is, don't do that. Buy a bank and slowly, steadily build the balance sheet. Varo had to go from 0 to 100 as a bank very quickly, and that's a cold start problem. Building a balance sheet business takes time. Companies with existing revenue and path to profit, have time.
(I don't mean to throw shade on Varo, I just think they're a victim of timing and path dependency. The views below are with the benefit of hindsight, which they did not have!)
The existing class of US Neobanks has other fish to fry in the short term.
But there's a case for each.
Chime is on the path to IPO and needs to stick the landing on that before it takes on anything else. To get there they need to prove they can be a consistently high-performing lending business. That said, I wonder if acquiring a bank while private and then IPO would be easier.
Robinhood has plenty of room to run as it expands its credit card, subscription, and retirement offerings. They're probably some distance from needing the efficiency of being a bank, but the temptation has to be there.
Things like CashApp or Venmo are in a much more interesting spot but might have to spin out from the group to make it work. CashApp is one piece of a larger conglomorate (Block). Block has bets like a Bitcoin-based decentralized exchange, a self-custody wallet, and the Tidal music service. It also has an Industrial Loan Charter (ILC) in Utah but doesn't use it for CashApp. There's a case to be made for spinning out CashApp and letting it be the bank for 24m monthly actives, but I doubt that's a management priority.
Younger consumer competitors like Current arguably have different questions to answer, like can they stand out in a crowded market, can they continue to drive growth, and if they do, what gives them a right to win? Weirdly, being a bank might be it.
Then there are the spend management companies like Ramp, Brex, and Mercury. Each of these companies has a feature depth and breadth that's moving ever closer to being a financial operating system for companies. They're starting to own workflows, and compete more with ERPs and finance software. These companies have so much growth and profit potential.
The natural mover here would be Mercury. They build their entire product around their account, and have arguably had more direct impact from sponsor banks issues with regulators out of everyone. Yet their proximity to the Synapse, Evolve et al, dumpster fire means they're probably best off avoiding opening a war on that front for now.
Like the smaller consumer Neobanks, I think the smaller spend management / CFO stack companies might be well positioned to counterposition by acquiring a bank. Imagine an Arc, Rho, or even Novo getting a charter. Would that help rather than hinder their next funding round?
5. This math isn't the same for embedded finance use cases
I don't think Shopify, Airbnb, or Uber want to be banks. Each of these companies offers debit cards and financial products to its users that are tightly woven into its core experience. It also drives substantial revenue from that business model.
If Shopify became a bank, that would likely distract from its mission and medium-term targets. Its growth isn't as dependent on the cost of funds or deposit yield. Sure, it benefits from those products, but it's not the workflow.
The bank service company act alone could force a company like Shopify to jettison a large part of its SaaS business because you can't both be a bank and do non-bank things.
If your objective is owning a workflow, your goal should be getting closer to the customer and owning more of that workflow. This makes you embedded in finance, and you don't want to be a bank.
If your objective is directly serving an account and building other services around that, then going down the stack and owning more of the infrastructure and balance sheet will yield ever more profitable results.
6. More innovative and efficient banks will benefit the economy and society
In the next five years, at least a handful of multi-million customer Neobanks in consumer and giant spend management companies will become banks. The fears of revenue multiples and the regulatory burden are outweighed by the size of the long-term revenue and profit potential.
The path there won't be easy, but sometimes the pull is too great to ignore.
In much of the world, Fintech innovation is a direct mechanism for growing GDP. Brazil, India, the UK, Singapore, China, Hong Kong, Vietnam, Colombia, and even the entire Middle East are actively promoting their Fintech industries through policy, state investment, regulation, and infrastructure to unlock growth.
Whether it's real-time payments in India and Brazil or neobanks like Nubank or Revolut becoming banks, these companies consistently create jobs, serve new social segments, and increase competition with stale incumbent offerings.
Last week, a conference for the fintech sector was held in India with 100,000 attendees. This week, I'm attending a conference in Saudi Arabia with 26,000. Compare that to M2020 Vegas and its 8,000. Sure, attendees aren't a proxy for quality, but they are a sign of the relative importance placed on the sector.
That's a good thing. The best incumbents step up, and get more competitive.
The US is currently an outlier here, and that's making it less competitive on the global stage.
The US has by far the most competitive and dynamic private sector in the world, and while it's unlikely to get India or Brazil-style regulatory simplification, I think there's a giant missed opportunity to grow GDP with some simple policy options.
What if the OCC, FDIC, Fed, and Treasury had a "charter with restrictions" where a de-novo or recently acquired bank spent 12 to 24 months under surveillance, with regular oversight?
In this world, you could become a bank while having regular feedback loops to ensure safety and soundness, measure progress on BSA/AML progress, lending fairness, or third-party risk management.
Innovative, digital-first companies faced with this constraint would be forced to adapt and adopt solutions to manage these risks responsibly.
But the dividend would be that it's digital-first. It's progress-first.
It's finance accelerated.
It's f/acc.
ST
4 Fintech Companies đź’¸
1. Skyfire - The commerce layer for AI agents
Skyfire lets AI agents pay and get paid for performing useful specialist tasks. For example, if you build a specialist language translation agent, it could charge a more general LLM or 3rd party for autonomously performing that service. The Skyfire SDK lets agents discover and pay for services, set budgets and demonstrate their "good actor" status by proving they have successfully worked with other agents.
🧠A commerce layer for agents is interesting because it enables microtransactions at scale and speed. The Skyfire team talks about "on-ramping" into the network with cards and paying by bank or stablecoins. Once you're in this network, each agent will have a wallet, so the cost of payments will be incredibly low. Each wallet also has a history, so you can see successful agents who have done lots of work before. Agents that can do economically useful things at speed and low cost feel instinctively like some kind of huge unlock.
2. Thera - Deel + Stripe Invoicing competing on price
Thera helps companies pay a remote workforce and contractors through a single platform. It offers instant payouts, low-cost FX, low-cost employee of record, and a "dedicated stack" at all pricing tiers. Their key differentiator is helping contractors "earn 3% more" by saving the employer costs in money sending. This becomes an employer perk when hiring.
🧠Deel, Rippling, and Remote.com are well-established and will be hard to dislodge, but the market is much more price-sensitive. In the age of AI, everyone is economically sensitive. Speed and efficiency are watchwords and Thera plays to this. If Thera can grab enough mind-share, the next generation of AI-first companies might default here instead of Deel.
3. ScribeUp - Subscription Management in Digital Banking
Scribe Up allows community banks, Neobanks, and companies that offer card programs to embed subscription management. It lets users cancel a subscription with 1 click, get alerted on upcoming annual subscriptions, and get offers.
🧠This is a fairly standard best-in-class feature that's still not widely adopted by banking apps. The vast majority of mid-sized banks, especially, are still so happy to have an app that they're not moving the experience forward at all. I'd expect every major core provider to have this kind of feature-as-a-business in 5 years.
4. Wavvest - Digital platform for RIAs.
Wavvest gives a digital CRM and portfolio management suite for RIAs and their clients. Clients get a modern mobile app with instant account opening and funding, and advisors get to reduce their admin burden. RIAs spend a lot of time on paperwork like account openings or portfolio management, but Wavvest promises to automate that.
🧠The pitch here is for wealth transfer. There are plenty of companies in this category, but because RIAs are a long-tail market, it's hard to tell which will win the distribution battle. This is the profitable core of many banks like Schwaab and Morgan Stanley. there's money to be made in helping the wealthy manage their money.
Things to know đź‘€
The FDIC announced its proposed rule on fintech accounts, designed to prevent a Synapse repeat. The rule would require daily reconciliation of “for benefit of” (FBO) accounts or direct access to a Fintech company ledger.
🧠The rule exempts broker-dealers, and Synapse had a broker-dealer subsidiary. However, for SVB and most other cases, this would be a helpful rule.
🧠Although that’s beside the point, the core issue is at the banks ledger. They are legally responsible for reconciling customer funds, and this rule makes that black and white, which is a good thing.
🧠The rule has unanimous support among FDIC board members. The election won’t impact its passing.
🧠The news of Five Star Bank exiting BaaS isn’t surprising. It’s no longer lucrative unless banks invest in compliance. To carry that cost of compliance needs either modern tech for efficiency, OR scale. Banks that have neither fail to see the business case.
🧠It is still possible to win at embedded finance. But you gotta come correct. Banks that have invested in compliance and hit scale have structurally higher ROTE than their peers. (Did I mention we’re running a Chatham house event on the fringes of M2020 on exactly this? HMU if you’re from a bank and want to join)
🧠Did you see that the FDIC has granted 8 De Novo charters this year? The ice is thawing. If you were going to launch a tech-first, API-first, digital bank for embedded finance, now could be a great time (ala Column, Lead Bank etc).
🧠Can we stop namechecking VCs for clicks now? Seriously it’s getting tired. Even Bloomberg is at it now. Imagine if we did this with public companies.
CNBC notes that discussions for Apple Card are still early and may falter on price and feature availability. The Apple Card has high losses, so JP Morgan would likely want a low price for the $17bn of loans. Affirm is now available as a BNPL option for Apple Pay users.
🧠Will any bank give a home to an Apple Card? Five years ago, this would have been every bank's dream, and Apple had all the leverage. Now, Apple is battling antitrust, opening up NFC, and everyone knows the Apple Card is a loss maker.
🧠Banks don’t like losing money. Apple’s historic insane leverage has gone. They’re realizing what everyone does: that lending is easy, but getting paid back profitably is hard.
🧠Perhaps the Affirm news is a sign of things to come. Apple’s can likely achieve much more through partnerships, open banking and wallet integration than they can trying to own financial services.
🧠Every brand with millions of customers should be looking at how access to NFC would change their offering. The tap-to-pay chip functionality makes secure commerce, ticketing, and identity much more possible. I’m excited by where this could go.
🧠Remember when Google Pay mattered? Google is too busy being a cloud service provider to play meaningfully in wallets, and that’s sad. They’re following Apple’s lead instead of innovating. Who’s thinking we see Affirm integration in about 12 months? lol.
Good Reads đź“š
In my favorite collab of the summer, Francisco of Chaos Engineering and Alex of Fintech Takes go deep into the history of credit underwriting, and how to build a data-driven credit model. This is a first-principles examination of using AI and ML to examine credit data. It runs through everything from what data to collect to how to build, evaluate, and deploy a model.
🧠This deserves a place as a textbook and recommended reading.
Tweets of the week đź•Š
a16z made this handy list of the top 50 AI Gen AI web products: a16z.com/100-gen-ai-app…
We checked, and turns out that 82% use @stripe.
We've been building a bunch of functionality that's useful for AI products, including usage-based billing (stripe.com/billing/usage-…) to handle… x.com/i/web/status/1…
— Patrick Collison (@patrickc)
11:26 PM • Sep 10, 2024
That's all, folks. đź‘‹
Remember, if you're enjoying this content, please do tell all your fintech friends to check it out and hit the subscribe button :)
(1) All content and views expressed here are the authors' personal opinions and do not reflect the views of any of their employers or employees.
(2) All companies or assets mentioned by the author in which the author has a personal and/or financial interest are denoted with a *. None of the above constitutes investment advice, and you should seek independent advice before making any investment decisions.
(3) Any companies mentioned are top of mind and used for illustrative purposes only.
(4) A team of researchers has not rigorously fact-checked this. Please don't take it as gospel—strong opinions weakly held
(5) Citations may be missing, and I've done my best to cite, but I will always aim to update and correct the live version where possible. If I cited you and got the referencing wrong, please reach out