Fintech 🧠 Food - Adyen's big miss

Plus; A Rant on Creativity in finance, Nubank earnings, & why the DoJ is coming for Visa

Hey everyone 👋, thanks for coming back to Brainfood, where I take the week's biggest events and try to get under the skin of what's happening in Fintech. If you're reading this and haven't signed up, join the 32,811 others by clicking below, and to the regular readers, thank you. 🙏

Hey Fintech Nerds 👋

What a week! 

Adyen lost 40% of its value in a day as it reported below-expected growth, NuBank crushed its Q2 earnings, Plaid is partnering with Pinwheel + Atomic, and the DoJ is coming after Visa again.

Adyen's growth couldn't last forever, but not growing by a staggering amount is a narrative violation. Payments are a race to the bottom, and the market likes balance sheets. But don’t forget, Adyen is a bank. Long term, they'll be just fine, and I'm not going to bet against them, but this is the first time they've seemed mortal. Be greedy when others are fearful, or the benchmark just dropped? (Things to Know 👀). 

Visa's tokenization pricing strategy is worth unpacking. They incentivize merchants to use tokens with lower prices and for better security, but that also creates a level of lock-in. How? See more in (Things to Know 👀). 

There are banks, and then there's NuBank, and Plaid partnering is a sign of the Fintech times. 

It's great to be a bank when rates are up. Infrastructure companies focused on profitability are more open to partnering. Put together; the Fintech industry is maturing.

So what's next?

I think it's more creative regulated financial products. We've disrupted distribution but not manufacture. What if that changed? This week's Rant is "Can we create financial products?" 📣 Weekly Rant.

PS. Congrats to Lex Sokolin of Fintech Blueprint fame for the launch of Generative Ventures (great name and timing). Lex has always been the "Fintech Nerd's Nerd." He has a habit of blowing my mind and making me laugh in the process. Super excited to see what comes next.

PPS. Congrats to Citizens Bank of Edmond having launched the neobank Roger to serve the needs of military service members. Who says niche Neobanks aren't needed? This will do well.

Here's this week's Brainfood in summary

📣 Rant: Can we create better financial products?

💸 4 Fintech Companies:

  1. Cascading - GenAI agents for Bank Processes

  2. Accend - KYB and EDD AI Co-Pilot 

  3. Trench - The Open Source toolkit for fraud prevention

  4. Lula - "Stripe for Insurance"

👀 Things to Know:

📚 Good Read:

Weekly Rant 📣

Finance is more creative than you think.

Fintech has led to a burst of creativity in financial services. 

Whether it's cashflow-based underwriting, 100% mobile servicing, or enabling e-commerce to become a default way to buy things. Fintech has reshaped the finance industry. 

It is 5% of payments by market cap, 40% of consumer checking accounts are opened at Chime or PayPal Apple just collected $10bn of consumer deposits in its savings product as most banks see deposit flight. 

Financial services remain the world's largest profit pool so the prize for innovating and succeeding is massive. It's also everywhere. Finance is a horizontal that cuts across every industry, and often things that don't look like finance are, in fact finance

But after a decade-long Fintech boom, folks who have been around for a while wonder. What's next? Is it more of the same and gradual maturation?

For critics, the argument against Fintech is that "all it has done is re-invent distribution" or "that is just regulatory arbitrage." 

We've barely scratched the surface of creativity in financial products and services. We've talked for decades about "personalization," but we haven't delivered. Regulation created a strait jacket over what a financial product is and how it must be manufactured and distributed. 

Any new technology like AI must be "explainable," any lending offered to consumers must meet a myriad of rules about fairness, and financial institutions need to balance consumer privacy with preventing money laundering. 

The job of building a novel financial product and getting it live is like trying to steal priceless art from a museum in a heist movie.

So I wanted to look back at the creativity. Where did it come from? How does it get delivered? And where might it appear next?

And will our GenAI buddy help us build more, faster?

  1. What is a financial product

    1. "Regulated financial product" vs. "Digital product"

    2. Manufacturing vs. distribution

  2. Innovations in distribution

    1. Cashflow underwriting uses better data to distribute lending

    2. Neobanks are new ways to distribute debit cards

    3. BNPL is a better way to distribute "lending."

    4. Spend management is a better way to distribute corporate cards

    5. Embedded finance is a better way to distribute everything

  3. Regulation makes financial product creativity stupidly hard

    1. Regulation is useful for company and customer

    2. But regulation creates costs and limitations

    3. Over time, limiting product innovation around a small core set of products

    4. Products rarely get combined or created because the complexity explodes

  4. Finance has been creative historically

    1. Bonds may have ushered in the industrial revolution

    2. Stocks may have also kickstarted the industrial revolution

    3. Derivatives offered a way to manage price shocks in the 1970s

    4. Collateralized Debt Obligations (CDOs) create more profit from loans and spread the risk across multiple banks

    5. Exchange Traded Funds (ETFs) are a cost-effective way to buy multiple stocks or assets simultaneously

    6. Tokenization uses software to reduce costs and increase automation

  5. What barriers do we have to overcome to get creative?

    1. New financial products have unintended consequences

    2. Combining financial products can go wrong if the incentives are broken

    3. New financial products don't have a market until they do

  6. Areas of opportunity

    1. Packaging existing private asset classes to new audiences

    2. New or alternative asset classes

    3. Thoughtfully combined products

  7. Don't hate; simulate.


1. What is a financial product?

a) Regulated vs. Digital products. The way tech people and bankers think about product is fundamentally different. So it's helpful to think about a "regulated financial product" vs a digital product. 

Two examples

  1. Regulated financial product: A checking account. Features include minimum deposit amount, payments in, payments out, autopay, and if you're looking for an APY yield. Pricing would be fees or subscriptions (e.g., $10 per month to bundle phone insurance and free ATM usage).

  2. Digital product: A bank's mobile app. Features include digital onboarding, transaction history, and request for a new debit card. Typically there is no pricing; all of the economics are buried in the regulated product.

The regulatory burden, skill set, and experience often differ between regulated and digital products. Some of the bigger banks often split digital products from "product" for this reason. 

As a bank, you could hire someone great at mobile to lead your mobile app, but that person would struggle to lead your checking product. Moreover, the systems, people, and processes were built or bought in different decades. The regulated financial product lives in a system of record or "core" system and has a team who builds and maintains that product. The digital product software and team were added later as a layer over the top. 

This distinction is important because it also shapes what gets regulated and what can get manufactured as a new product.

(This model is changing due to the Fintech boom in the past decade. In Neobanks and digital banks, the financial and digital product are usually considered the same. But the regulated financial product is still tough to change, even with better tech.)

b) Manufacturing vs Distribution. The financial vs. digital distinction can also be considered the separation between manufacturing and distribution. The digital product is a form of distribution that sits alongside branches or telephone banking as a way customers shop, buy, and are serviced. 

While we have seen plenty of innovation in distribution (and, to some extent, pricing) over the past decade, manufacturing innovation has been limited. For the most part, financial products are static, regulated, and hard to change.

As a result, we very rarely see innovation in product manufacturing. 

2. Innovations in distribution 

We have seen plenty of innovation in product distribution. Taking the same raw financial products that always existed and repackaging those. While this creates some regulatory grey areas, for the most part, a compliance team somewhere can get comfortable with it, and it provides a route to market for innovators.

From a pure "banker" perspective, these digital products aren't "new." 

But to a customer, they very much are.

And that's a mindf*ck for both sides of the TradFi / Fintech universe.

a) Cashflow underwriting is using better data to distribute lending. Petal pioneered in offering credit to low credit score consumers using their checking account data to score risk. Since the advent of open banking (and now payroll data APIs), companies have understood a user's financial behavior even if they have zero credit history. This has been so successful that now most of the major banks have started to, or will soon offer, the service.

b) Neobanks are new ways to distribute debit cards. A debit card typically comes with a regular checking account to simplify everyday spending in-store and online. Neobanks created a new category of financial products, where a non-bank builds a suite of features around the debit card as the core financial product (the debit card). Simplifying onboarding, offering mobile-only servicing, changing the fee structure, and innovating features like "get paid early" or instant payments. 

c) BNPL is a new way to distribute lending. Merchant-funded Point-of-sale lending always existed, involving a lengthy, agent-led application and form-filling process in-store. Because the merchant takes the risk of the consumer not repaying, often this includes "0%" offers for consumers. BNPL repackaged that at the e-commerce checkout to "pay-in-4" installments. By creating a mobile app and shopping experience, they have also been able to better track consumer buying behavior across multiple merchants and drive new transactions and repeat customers for their merchants.

d) Spend management is a better way to distribute corporate cards. Corporate expense cards were a necessary evil. Before the rise of digital spend management to travel on business, you'd have to collect receipts, take photos, and manually upload them to some god-awful SAP portal. Brex, Ramp, Novo, and a crop of global equivalents started to make that effortless. They issued credit cards to smaller growth businesses with much larger lines than personal credit cards. They would send a real-time alert whenever a transaction happened to take a photo of the receipt (or forward it to an inbox). For the first time, expenses worked without utter agony for the user. This became a wedge to build Payables, receivables and become a company CFO dashboard.

e) The strongest example of better distribution is embedded finance. Embedded finance is a better distribution of financial products relevant to a user's behavior. Putting lending at e-commerce makes more sense when buying the thing than it does as a separate website and application process.

We took the processes around the regulated financial product and gave them better distribution. 

But we haven't meaningfully altered the regulated financial products.

As a map, I think about financial product innovation coming to the bottom up the complexity curve. We started with unbanked consumers and have worked our way up through SMB, wealth, and corporate.

(for a full explanation, see this video from a few years ago discussing the model).

3. Regulation makes financial product creativity hard.

The regulation's primary purpose is to ensure that the profit motive of financial services doesn't exploit consumers. As a large Fintech company or financial institution, you know more about finance than your customers. Regulation exists to solve this information asymmetry and promote fairness and inclusion in the financial system. 

a) Regulation is helpful for customers and the company. For the company, being regulated can act as a seal of quality and confidence for customers. The regulated logos like FDIC insured (or FSCS protected in the UK) and even the small print act as a subtle mark of quality and confidence. For the company with those seals, that becomes a responsibility, but it also attracts customers and, therefore, revenue. (Let's set aside how various embedded finance products might have somewhat stretched this privilege for this Rant).

b) But regulation creates costs and limitations. Regulatory reporting defaults to paper, it involves exams, and the vast majority of it involves building a process to do XYZ and then documenting that process. During an exam, a financial institution will roll out its policy on (for example) fair lending that comply with the Fair Housing and Equal Credit Opportunity Act. This data is collected and reported to State and Federal bodies under the Dodd-Frank Act. Someone has to create the policy, set up the processes, report on its effectiveness, and adjust over time. 

Seemingly sensible things like collecting more data about a consumer to more accurately price lending risk might be illegal, even if the net result was more inclusive lending. Because regulation and legislation are often prescriptive and complex.

c) Over time, limiting product innovation around a small core set of products. The complexity of these rules and the diverse types of harms they're designed to prevent makes innovating within them challenging. A seemingly simple question like "Can we do this?" Doesn't have an easy answer. So the default becomes "no." 

d) Products rarely get combined because the complexity explodes. Imagine you wanted to have an investment account that could double as an everyday spending account. If you trade advanced products (like options) in that brokerage account, it may go into a positive balance. Wouldn't it be great if you could spend that instead of your paycheck? Yes. But. With options that account can go negative, if you spend temporary balances, you might take the account insolvent or liquidation. What works in a good market might not work in a bad one. Also, consumers should avoid day trading and save gradually over time. It's notable that Robinhood recently closed its "cash management" account that did this to new applicants.

That's just one example where the complexity of spending with a non-traditional product is harder. Imagine a table with every major product type combined with every other, and against that, every regulation that applies to each product.

On some level, embedded finance can help here because the underlying providers should help abstract the regulation complexity, but if you're customer-facing, you still bear responsibility for customer outcomes.

4. Finance's creative history has some successes

When I think about financial products, the following examples are net new manufactured products. 

a) Bonds may have ushered in the industrial revolution. Bonds are an IOU issued by a government or company. They became popular as countries grew and built infrastructure like railroads, bridges, and ships. Today across the globe, Bonds are worth more than $100trn, and with rising interest rates, are considered a "risk-free" way for investors to earn yield (because that yield will be paid so long as a Government continues to exist and doesn't miss a payment). Bonds powered much of the growth we unlocked in the 19th and 20th centuries and the industrial revolution. 

b) Stocks may also have brought about the industrial revolution. Stocks allow companies to raise capital in return for a share of profits. The Dutch East India Company is credited with the invention in 1602 with the investment used for shipbuilding that would be used to trade spices and assets from India. As international trips expanded, these voyages became riskier, and selling shares allowed companies to spread that risk and raise more capital for riskier ventures than might have been possible with bonds. The global stock market is valued at over $100trn.

c) Derivatives offered a way to manage price shocks in the 1970s. During the 1970s, the world was battling inflation, an oil crisis, and commodity price shocks. Derivatives are contracts that "derive" their value from an underlying asset. They allow companies to lock in commodity prices (like oil, wheat, or corn), currencies, or even interest rates. If you're a farmer expecting to sell your crops, you can lock in the price today without worrying about what it is tomorrow. An investor might believe the price will change, and they can profit from your crop so they would offer you the locked-in price today. Derivatives offer certainty on one side and the opportunity for profit on the other. This allows businesses to manage external market shocks more evenly. The global derivatives market is valued at over $600trn (although that's misleading, long story). 

d) Collateralized Debt Obligations (CDOs) create more profit from loans and spread the risk across multiple banks. Imagine a bucket, and in that bucket, I take some very high, medium, and low-quality loans and sell them as a bundle. By bundling loans and selling them as CDOs, banks could earn more than selling them individually. The rise of the CDOs came in the mid-1980s to create new liquidity for banks to lend with. This new lending meant that businesses could grow and the economy could expand. They became a more efficient way to sell loans, creating more room on a bank's balance sheet to sell new loans. Today the CDO market is worth <$10trn.

e) Exchange Traded Funds (ETFs) are a cost-effective way to buy multiple stocks or assets simultaneously. An example of a popular ETF is an index. One example is considering an ETF, like a sample size of every stock on a particular market. For example, the NASDAQ QQQ fund tracks the top 100 stocks on the NASDAQ in proportion to their size. So if you invest $100, and the #1 company (e.g., Apple) is 10% of the overall value of the top 100, then $10 you'd hold $10 of Apple stock through that fund. ETFs are an easy way to buy traditionally harder-to-purchase products like Gold. Today the ETF market is estimated at over $10.3trn.

f) Tokenization uses software to reduce costs and increase automation. Different software providers could operate many processes that would be a department of a financial institution. Some tokens represent existing instruments (like cash, stocks, or bonds). Some tokens are much newer, representing assets like intellectual property or energy (kWh). The space for innovation and automation here is significant but still early. 

Notice how far newer financial products are from the distribution?

I wonder if this starts to meet in the middle more?

5. Challenges building new regulated financial products

a) New financial products have unintended consequences. The stock crash of 1929 followed enthusiastic speculation and a lack of regulation and led to the great depression. The CDO was called a "weapon of mass financial destruction" for its role in the Global Financial Crisis of 2008. The "bundled" loans meant banks didn't know what housing loans were good or bad. ETFs and high-frequency trading worsened the flash crash of 2010. As a shock hit the market, the ETF funds began selling together quickly, automatically hastening the decline. 

b) Combining financial products is often hard and can go wrong. The UK had a scandal for "Payment Protection Insurance" (or PPI). This insurance was bundled with mortgages to protect the homeowner from missing a payment. However, most consumers were sold PPI without being aware of it. It became a way to "default" in more revenue (although perhaps the issue was more distribution than manufacture.) 

c) New financial products don't have a market until they do. Building any new widget is risky because there's no guarantee it will have a product market fit. Venture is the natural way to de-risk that (if it can avoid the kids-soccer style group think that sometimes overtakes some elements of itFoom.)  

6. Areas of opportunity

The next battleground is two-fold.

  1. A cost battle in the areas disrupted so far. A push for better unit economics (typified by Nubank and the digital banks).

  2. A push into the more complex financial products up the complexity curve.

Fintech is the natural creative. Growing by copying the incumbents and playing their game is nearly impossible. They have to get creative.

Some early signals of where that creativity might come from.

a) Existing asset classes to new audiences. Many assets were only available to a few ultra-high net worth or institutional investors. Examples would be private debt, equity, or venture capital. Some companies have started to aggregate lots of smaller investors into a fund. While this represents the new distribution, they're a sort of private market "ETF." Not in the sense that the bucket has multiple stocks together, but they are a fund that is lower cost and easier to buy.

This market is big and growing.

b) New or alternative asset classes. Institutions have begun to purchase recording artist back catalogs (IP). For example, the Red Hot Chili Peppers sold everything pre-2020 to Hipgnosis, a specialist IP acquisition firm with $3bn in net asset value. They also count Mark Ronson, Shakira, and The Chainsmokers among the IP they own. They offer this as an ETF on the London Stock Exchange, but it's often harder to find in modern investment platforms. Now imagine this for whiskey, art, collectibles, carbon, and every other asset that isn't commonly traded.

c) Combined products. Baking insurance into lending isn't a bad idea in principle. In fact, almost no financial product is a bad idea in principle. It's the implementation that matters. I spoke to TruStage, who's something called Payment Guard. It's a group insurance product that makes some number of payments on the borrower's behalf if they lose their job or become unable to work due to disability. The crucial thing is that the lender pays the premium, not the consumer. 

I'm sure there are 1000s of other examples of this type of thing (and the same thing done by others). The takeaway is these products can be combined if we balance incentives well and rely on specialism. That's not a guarantee of success but a good north star. It's also interesting how partnerships play a role here. The lesson is thinking about how one financial product lives alongside the other and the risks that it presents. This feels like a GTM exercise that could have countless other places it works.

What happens if you use a savings yield as a discount against your mortgage rate? This isn't available in the US due to several tax laws but is a standard practice in the UK. There are likely many other legal tripwires when products are combined, but it's an interesting thought experiment.

7. Closing thoughts 🤔

Getting creative in financial services is incredibly hard due to the legal and regulatory complexity. 

Often the biggest blocker here is the unknown. "What regulations might apply if we combined x with y" is a tough question. Fortunately, we now have countless GenAI agents and buddies that can accelerate this process.

We need to be able to experiment more as an industry.

NVIDIA has a product called Omniverse. It allows companies building cars to accurately simulate their new car and then run it on millions of road surfaces for billions of miles before a prototype is ever created. 

What if we did something similar for financial products?

A digital twin for the financial industry where we can simulate launching a new product.

That's one reason I'm so excited the UK has a permanent digital sandbox that lets companies build a new widget and simulate what running that product with 50m UK accounts would look like. The simulation is based on data synthesized from major UK banks and has the full visibility of the regulator.

There's a long way to go from having a data set to being able to simulate a digital twin, but damn, that idea is exciting, no?

Don't hate. Simulate.


4 Fintech Companies 💸

1. Cascading - GenAI agents for Bank Processes

Cascading helps financial institutions "hire GenAI" agents for processes like onboarding, customer service, and back office automation. During onboarding, the agent can email customers, request further information, and analyze it vs other data sources. Customer service can manage complaints and surface responses from bank policies and regulatory guidelines. 

🤔 From outsourcing to AI sourcing. Many of these tasks are situational and don't follow a simple workflow. Where every Fintech company and Neobank develop these features in-house, small and mid-sized financial institutions may lack the expertise. Even with a great tech team, they're constrained by rules, existing policies, and providers. Hiring an "AI agent" is a great mental model for where GenAI fits. What was UIPath, or an offshoring firm, is now GenAI.

2. Accend - KYB and EDD AI Co-Pilot 

Accent helps risk operations teams summarize the key points of potential risk about a customer drawn from sources like Google, LinkedIn, OpenCorp, and Crunchbase. It produces automated Know Your Business (KYB) and enhanced due diligence (EDD) reports formatted and explained in your company's style and within its risk appetite. Accend can also help with chargeback disputes, collating evidence, filing for arbitration, and preparing reports for SAR filing. 

🤔 Where's the moat here? Many of these businesses look like OpenAI + Vector databases. Perhaps the most interesting thing about Accend is the setup process that involves baking in the Fintech or banks processes and risk appetite into the system. That means the model outputs will be much closer to what a human would create. But is that enough to justify an entire SaaS business? I fear Fintech companies can build this with minor tinkering on the weekend, and financial institutions are hard to sell to without market access and distribution. My gut says the moat is actually in the ability to collect the most data and/or provide the most accurate data to train these models with. Which raises the question, what data can we trust? (Comments apply to the previous company too)

3. Trench - The Open Source toolkit for fraud prevention

Trench is a self-hosted fraud platform for merchants and marketplaces that allows them to review merchants or payments through its dashboard. The service also features network analysis, location intelligence, and behavior analysis. The goal is to compete with traditional vendors who own the stack and data and often can't help solve an individual's merchants' use case. 

🤔 I like orienting as "self-hosted" as an alternative to self-build. Every marketplace would build tools to solve its users' use cases and risk profiles in a perfect world. Positioning as open source is a good proxy for "it's yours, and you can run it and own it." That wasn't true of traditional vendor solutions, but more modern platforms like Alloy, Unit21, and Sardine* are much more configurable, allowing open data access. The strength of Trench will come down to how well users can use the product and the results they can get. Vendors never solve everything, but clients often lack the knowledge to get the results they expect. 

4. Lula - "Stripe for Insurance"

Lula provides "insurance infrastructure" for Auto, trucking, and freight companies and says they'll "soon" support embedded insurance. Lula provides risk assessments, policy management, claims management, and a network of insurers. 

🤔 The only companies raising decent growth rounds are Insurance-as-a-Service? European Insurance-as-a-Service startup Qover has focussed on the embedded finance use cases and just landed its series C with clients like Revolut and Deliveroo (think Doordash for the UK). Lula has focussed on more traditional industries like Auto and Freight, forcing it to build a perhaps more enterprise-grade solution set with APIs. I'm curious about where the long-term revenue and margin are. Is this in the heavy industries B2B2X or embedded B2B2C? Maybe both.

Things to know 👀

Adyen reported 739.1 million euros ($804.3 million) revenue in H1, up 21% from 2022 but below the expected 40%. EBITDA is down 10% from H1 22. Adyen says talent costs increased as it ramped hiring and customers became more cost-conscious. 

🤔 Investors wanted "up only" from Adyen; these results are a narrative violation. Adyen has higher structural costs with hiring and fewer macro tailwinds. The party may be over for them as the untouchable growth story.  

🤔 Adyen benefited from macro tailwinds like the shift to e-commerce during the pandemic. It now faces macro headwinds as its biggest customers like Netflix, Meta, and Spotify reign in their spending. Adyen had a concentration of customers and segments.

🤔 Trend or blip? The end to the growth was expected just maybe not so soon. This has people worrying about how bad can it get? Now it must diversify and face pricing competition in the crucial US market, which could further erode profitability.

🤔 The market likes balance sheets, not transactions. Payments companies have had an incredible run in the past decade. The switch to higher interest rates could change that (although you don’t see it in bank stocks). Companies that can connect net interest income (NII) from deposits and lending activity are booming (see: Wise, Bunq, etc). Don’t forget, Adyen is a bank.

🤔 Did the benchmark just drop? Many saw Adyen as the benchmark for Fintech. If that drops, what does it mean for Fintech? That’s an oversimplification. The future is clouded for sure, but I’m not gonna bet against either Adyen or Fintech.

🤔 Adyen was famously lean, but it's hiring to grow. Adyen's headcount was less than 3,500 vs. Stripe's peak of nearly 8,000, and incumbents like Fiserv with far higher. It is pushing into more markets and extending its product base as its customers go global and demand more. This, in turn, creates cost. I hypothesize that there's a natural ceiling to size and market share a payments company can reach before it kills its margins to grow again. 

🤔 Let's not forget Adyen is a beast. They delivered a 39% EBITDA margin plus 21% revenue growth. A "growth company" is anything hitting the rule of 40. The rule of 40 is margin + revenue growth. For Adyen, that's 60. Fiserv reported 8% H1 revenue growth and 36% margins. Even in a bad patch of Macro, Adyen is still a beast. Payments companies generally have super high margins, but that's a long story. 

🤔 The challengers are still taking market share from incumbents. Zoom out from these results, and you see a picture of big challengers facing the same macro head and tailwinds as incumbents but adapting faster. They're also continuing to win market share. The structural landscape shift remains true.

The DoJ is investigating Visa for charging more to merchants who do not use its "tokenization" system. The system swaps 16-digit card numbers for "tokens" tied to an individual device and merchant (e.g., an iPhone and Apple Pay). Last year Mastercard agreed to share customer account information for debit cards with competing networks.

🤔 Tokens have benefits for security and convenience for users. The token replaces the 16-digit card number and is tied to a specific device. 1 in 5 US consumers uses Apple Pay every month, and 73% of GenZ use Apple Pay in-store. It is becoming a default. It also means if you change your card, the network can update this with the merchant, and your new card just works. You don't have to enter new card details if you have a favorite e-commerce site (like Amazon) and use Apple Pay.

🤔 But tokens create "network lock-in" for merchants. Network tokens like those issued by Visa can only be used on the Visa network. Merchants in the US can direct a card transaction to a competing network to secure a better price. By attracting consumers and merchants to the security and convenience of tokens, Visa (and Mastercard) also created a level of network lock-in.

🤔 Lock-in dressed up as security is a classic incumbent trick. You'll notice banks' big issue with Open Banking is "security." The reason Apple is pushing Apple Wallet for identity is "security." On the surface, that's not wrong (and often better than what came before). But it also creates lock-in. Lock-in creates a moat and a higher barrier to entry for the competition and can be viewed as anti-competitive.

🤔 Tokens aren't fraud-proof! Far from it. Many merchants and issuers see high fraud rates coming from Apple Pay users without the classic signs of fraud. The token hides many data points operators would use to detect potential fraud (like the Bank Identification Number or velocity checks). Fraudsters and scammers know this and are exploiting the weaknesses. This is solvable, but the first step is knowing the token isn't a protection from fraud.

🥊 Quick hits

🤔 There are banks, and then there's Nubank. Damn, this is impressive. Nubank now has 83m customers adding 1.5m per month. The cost to operate an account is running at $0.8 per month with an ARPU of $9.3 per month. This tells me two things 1) They're investing heavily in product/geographic expansion, and 2) Owning their tech infrastructure makes them incredibly efficient. The operator talent from Nu will feed the Brazil and LatAm ecosystem for decades. 

🤔 Average Revenue Per User Momentum. If most customers are new, the cross-selling engine is just starting. Even when they saturate home markets, there's a ton of revenue headroom. Does it make you wonder if they'll expand beyond markets like Brazil/Mexico? Also, would they ever spin out / re-sell the platform like Starling has.

🥊 Plaid partners with Pinwheel on income verification. Plaid had offered income verification previously but has added the ability to validate against payroll data and streamline direct deposit switching. 

🤔 Partnership season is upon us. Plaid was once the canonical "but we can build it" company. To see them partnering is a sign of focus, maturity, and how the industry cycle has changed. I always found "big Fintech" naively optimistic in its "everyone is a competitor" posture. Market pressures have created a much-needed reality check. Especially when revenue and profitability matter more than market share land grabs.

Good Reads 📚

The problem with GenAI is that it's bad at edge cases. Getting to 80% is fine; getting to 99% is much harder, so will we just fall back to manual work? And couldn't software have done the first 80%? It's also bad at being correct, and for some things, humans are cheaper. BUT. The best use cases are where correctness isn't binary, and there are plenty. It can and is being used in every market and is already better than humans at some high-value tasks like engineering on a $ basis. GenAI brings the marginal cost of creation to 0.

🤔 If creation goes to 0, what do creators do? The argument goes that microchips created an order of magnitude cost reduction in computing, the internet in distribution, then GenAI will do that to creation. You can see this in the media industry strikes with writers, actors, and artists.

🤔 Professions like law, finance, and healthcare are more creative than they're given credit for. They make great TV shows because the drama and creativity involved in these professions usually come from full contact with humans, incentives, and risks. 

🤓 Extra Credit: Should consumers trust GenAI is the wrong question by Alex Johnson. 

That's all, folks. 👋

Remember, if you're enjoying this content, please do tell all your fintech friends to check it out and hit the subscribe button :)

Disclosures: (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 (3) Any companies mentioned in Rants are top of mind and used for illustrative purposes only. (4) I'm not an expert at everything you read here. Some of it is me thinking out loud and learning as I go; please don't take it as gospel—strong opinions, weakly held.