• Fintech Brainfood
  • Posts
  • Fintech 🧠 Food - 6th June 2021 - The EU Digital Wallet, CBDCs vs Privacy and what Fintech can learn from Operating Systems

Fintech 🧠 Food - 6th June 2021 - The EU Digital Wallet, CBDCs vs Privacy and what Fintech can learn from Operating Systems

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 6,222 others by clicking below, and to the regular readers, thank you. 🙏

Weekly Rant 📣

Operating systems and finance

In 1991 a small open source project released the source code for a free operating system kernel based on the Unix code base.

It quickly became apparent that a kernel by itself "gets you nowhere," as the inventor Linus Torvalds once said. That system, called Linux, only worked on Intel 386 chips and hard drives created by AT&T.  

To run the operating system yourself, you needed a compiler, a shell (UI), a library, a file system, and crucially the ability to deal with the varied range of hardware on the market. Hard drives, processors, graphics chips, and human input devices needed to be supported.

By 1993 there was a thriving community of at least 100 maintainers of the Linux kernel, contributing various core components that helped the code base quickly adapt, and the Linux architecture developed as a set of concentric circles 👇.

Linux drives computer hardware with its kernel (core bit of code), the kernel has a shell (the UI), and around the shell are the various applications. By open-sourcing the Kernel, any engineer or developer could take these primitives and build (compile) an entire operating system, unleashing a wave of creativity and specialism for different use cases.

Before 1991 an operating system was proprietary, and the code that ran was as good as the company that built it. In many cases, this meant that as an engineer, if you wanted to create, say, high-performance communication networks, you either needed to write specialist software to drive the hardware yourself or rely on the operating system's performance. The commercial operating systems are highly opinionated about how an operating system should work. This made them easier to use for the mass market but less flexible for developers.

Linux created a community around the primitives as an open-source project, from hardware to kernel and even the shell. The open-source approach dramatically increased the quality of what was previously hidden and non-differentiated code. Developers could pick up battle-tested code for whatever their use case was.

Linux moved code quality from being an afterthought to the core concern. The developer enthusiast community adopted Linux with zeal, as the increased granularity of what it offered allowed them to build what they wanted, how they wanted. Linux was one of the first projects to empower developers completely.

As early as 1993 distributions, pre-compiled versions of Linux emerged. This complete pre-packaged software could be installed by any developer or user on almost any hardware. Throughout the early 2000s, Linux quietly replaced Unix and Microsoft as the default choice for high-performance servers and computing. Symbolically in 2009, Red Hat's (a Linux distro) market cap was larger than Sun Microsystems.

In effect, developers had taken the open-source primitives and created a version of the Linux operating system that handled different use cases. Red Hat became more suited to enterprise, while Debian and Ubuntu are often preferred by the enthusiast/developer community (for example the popular Raspberry Pi is based on Debian).

Perhaps most symbolically, in 2012, Microsoft announced support for Linux on its Azure cloud platform. From open source primitives, developers were free to create distributions for various use cases.  

B2B Fintech looks a lot like the server market Pre-Linux.

Historically if you wanted to launch a financial product, you either had to work with the deep financial infrastructure yourself or use a highly opinionated bit of software (AKA Core Banking software) to interact with the financial infrastructure.  

The infrastructure is where the primitives of finance—things like storing value, moving value, lending value are managed. The complexity of writing your software to interact with this infrastructure is significant; it would be like writing your own graphics chip driver to create a video game (this made sense when the market was nascent but is almost unheard of today).  There’s not just one infrastructure, developers need to interact with (or drive) specific financial rails (e.g., ACH, Wire, Cards).

The creators/vendors built core banking software (often written in the 80s or earlier) before the advent of modern software engineering practices. As a result, the software was never built for performance or adaptable to a developer's requirement.

There are two solutions to this

  1. Better packaged software

  2. Better primitives

Better packaged software has already transformed finance.

In the past five years, the B2B fintech space has exploded with creativity. If you want to build a fintech product, there is a clean, robust API for just about everything. Modern core banking platforms are emerging (e.g., Mambu, Thought Machine, 11:FS Foundry) and vertical fintech SaaS platforms that remove the need to run software entirely (e.g., Amount, Blend).

Better software is the on-ramp to better consumer and customer outcomes, empowering builders to get to market quickly and at a low cost.

It has never been easier to build and launch the most basic level of financial product in some markets like the US. Where "the finance bit" is non-core, and you want to differentiate on user experience, modern B2B fintech software, and APIs win for you. 

Better packaged software also comes with opinions. While it’s possible to do incredible things with the new B2B fintech providers, inevitably as a business that does fintech scales, it will hit limitations.

Better primitives are about to reshape finance at scale.

There's a point in every business where the value flips from speed to market to scale.  

As a fintech business scales, it bumps into all of the beautiful quirks of an aging financial system. Banks will randomly reject payments for no apparent reason; the payments rails will reject your payment if your formatting isn't correct. Well-designed software or SaaS will handle many of these quirks on your behalf but for a price.

At scale, unit economics matter.  And what if you want to do something the provider doesn't?

The great thing about opinionated software is it simplifies the complex to improve time to market. The frustrating thing about opinionated software is it makes trade-offs and choices that might not suit your use case.

Finance at scale is more like a high-performance server-side operating system than a consumer-grade operating system. As a business scales to 10m or 100m customers on their finance product(s), it becomes increasingly important that they can bend the software to their use case.

Maybe there's an excellent API for onboarding consumers, but it doesn't handle my use case. Maybe there's an excellent fraud platform, but it can't yet prevent this type of fraud we keep seeing. How do you orchestrate all of those new APIs to do exactly what you need them to do and in a way that differentiates you to your users?

This is historically when a team would self-build.

That was before the beginning of the open-source movement in financial services. Moov and FinOS take all of those complex, close to the metal use cases and make them high performance.

In effect, the kernel of financial services is slowly getting open-sourced.

The hardware, the kernel, the shell, and the application.

How would we ever get the best of both? Is it possible to have time to market and massive developer control?

I think the answer is yes.  

We're already seeing the evolution of proprietary operating systems in fintech. Companies like Unit or Bond create a developer-controlled dashboard to install "fintech apps" on their operating system (BaaS platform). These companies massively improve the simplicity of working with different fintech apps for anyone building or launching financial products.


I think there's room for a Linux of fintech operating systems. One where there could be many distros for specific types of company or use case.  

A Linux-inspired financial operating system for scale, high performance, and the enthusiast would be interesting. It has never been easier to do "another" Neobank, but if you want to do something at scale, there's still so much heavy engineering developers have to do.

Each operating system distro would need pre-built apps (e.g., a clock, a file system), but crucially, the developer or user is empowered to change any of those whenever they want, right down to the metal. The operating system would also enable developers to change with the kernel, the shell, and even how it interacts with the hardware.  

If you were building an OS in 2021 instead of 1991, you'd also probably want a marketplace or app store.

Something like that would be cool.


4 Fintech Companies 💸

1. VoPay https://vopay.com/  - Single API for Account to Account payments (🍁)

  • VoPay is a "pay by bank app" solution and API for the Canadian market.  Users can choose VoPay at checkout and are directed to their bank app (e.g., BOMO, RBC, Scotia) to authenticate the payment. The merchant gets to check funds are available before completing the transaction, and the API will pull relevant checkout data like name and address from the bank app (if available).

  • VoPay also has solutions for a business like bulk payments, payment requests, and payments tracking dashboard. The Canadian market is ripe for this kind of disruption, with a concentrated banking market and a limited appetite for regulatory change.  Aside from Wealthsimple, there aren't many well-known Canadian fintech companies, but that could be changing. Also, account-to-account payments are the killer app of open banking waiting to break out.

2. Sumatra https://sumatra.ai/  - Real time fraud prevention API

  • Sumatra provides tools to detect and block fraud and scam attempts across e-commerce, mobile banking, P2P payments, and iGaming platforms.  Scammers and fraudsters will try various tricks to gather cash that changes over time and is non-obvious (e.g., promo abuse or "gnoming"). 

  • Sumatra has an operations team that helps its ML identify new threats as they arise in the market; this allows them to respond to new threats much faster than algorithms trained on historical data alone.  In effect, Sumatra is a mix of human and machine fraud prevention that can be an outsourced specialist fraud team for companies that don't yet have one or are looking for a specialist to improve the impact of that team.

3. Verified.ly https://verified.ly/kyb - Onboarding as a Service

  • Verified is an identity, onboarding, and AML compliance platform that supports both KYC and KYB.  In addition to onboarding customers, Verified.ly will check OFAC lists, prevent fraud through IP detection, and can run KYC with just a phone number. Unlike most of its competitors, Verified.ly also offers a Know-Your-Business (KYB) onboarding tool.

  • There aren't many API-driven providers who offer KYB; the most well-known is Middesk. There's a vast opportunity for good KYB flows as SMB banking heats up, especially outside the US.  Interestingly the Verified.ly team page only has three folks on it, and none of the clients they list are live. The opportunity for what they do is massive. Perhaps this is early, with early-stage clients but, a little odd 🤷‍♂️.

4. Jeeves - Expense Management for Global Startups

  • Jeeves provides expense management, invoice reconciliation, and virtual cards for startups.  Much like how Brex solved the "it's too hard to get a credit limit if you're an international founder" problem initially, it's now too hard to be a startup that operates internationally. Jeeves is live in Mexico, Canada, and the USA and allows startups in those countries to "spend like a local" but settle with Jeeves in local currency (e.g., pay in $USD, but send Jeeves $CAD). 

  • Emerging from stealth with YCominator and a16z investing, Jeeves has all the hallmarks of a big-name player in years to come. It's still way too hard for businesses to operate internationally, but making that possible opens up pools of talent and commerce that could drive massive growth.

Things to know 👀

  • The Chinese Communist Party has taken steps to weaken its currency by moving the peg of its currency against the dollar. The RMB has seen weakness in offshore trading as global commodity prices soar. International FX experts believe China is intentionally weakening the RMB to help its struggling exporters. 

  • China's foreign currency reserves are shrinking, suggesting the Chinese economy is in trouble.  The likely response will be that many Asian economies will also weaken their currencies, making commodities (like oil) more expensive, weakening Chinese demand for those commodities. If the Chinese demand for commodities falls, the Chinese economy may suffer short term, but commodity prices may also cool.

  • 🤔 My Analysis: We're in a full-blown currency war, and everything happening in finance policy flows from that. Maybe historians will look back at this time as the currency cold war.

  • 🤔 My Analysis: The pandemic has impacted the supply of commodities, which has raised their price. In response to the pandemic, the US Government has printed trillions of dollars. Commodities are priced in dollars, and when there is a larger supply of dollars, the dollar's value against that commodity decreases.  The price of stuff is going up, while the value of money used to buy stuff goes down.

  • 🤔 My Analysis: Attempts to create Central Bank Digital Currencies (CBDC), regulate stablecoins, or issue COVID passports should all be seen in the context of Geopolitics. (e.g. 👇)

  • The wallet will store payment details and passwords to allow citizens from 27 member states to log into local government websites or pay utility bills using a single recognized identity. The wallet will all store official documents like a passport or driver's license and will offer "great flexibility for post-pandemic life."

  • In the same week, the EU-wide system for verifying COVID vaccinations and test results went live. The pass uses QR codes and digital signatures, and all data is stored in the user's home country. The announcement is at pains to point out "no personal data is exchanged or retained" during the verification process. 

  • 🤔 My Analysis: The EU now has certificates to go back to everyday life and "soon" have a wallet for all passports and identity data. If you wanted to implement bloc-wide digital identity, the pandemic certificate, QR code usage is a handy wedge product.

  • 🤔 My Analysis: The EU tends to be aggressive on privacy protection (as the home of GDPR, for instance) and has an anti-Big Tech posture. The announcements point out that Governments could not sell data from these wallets for advertising.  But governments are often pro-privacy from commercial companies, not themselves.

  • 🤔 My Analysis: If the EU ever did launch this service, it could be ultra-convenient for consumers and businesses. But the big word is if.  The EU has a history of long, slow, top-down, bureaucratic efforts that disappoint and underdeliver. SEPA, PSD2, Target 2.

  • 🤔 My Analysis: I worry that even if this wallet delivers, it's another step to a more balkanized internet and world.  The Indian Aadhaar model, the Chinese model, and the EU passport all feel quite different and not built for a global citizen.  

  • 🤔 My Analysis: If these wallets had open standards and APIs, it could be another story. Imagine making your "real identity" compatible with privacy-preserving online pseudonym tools like idx and verida. The alternative is there are two worlds: the world of the nation-state and the online citizen of nowhere. 

Good Reads 📚

  • The major fintech companies like PayPal, Square, and Stripe are in a war for merchants. The battle isn't limited to who can accept payments but extends deep into HR software, marketing, sales, accounting, and even credit.  For example, Paypal's acquisition of Honey added to its sales and marketing capabilities for merchants.

  • These platforms will eat into banks lending business over time, but in this piece, Ron stipulates the best opportunity for a bank is to be a balance sheet lender to those platforms.  In return for an almost zero cost of acquisition (CAC), banks would have to accept a lower margin.  As banks stop patting themselves on the back for the success of PPP lending, they need to find a posture to deal with the new goliaths. 

  • 🤔 My Analysis: It's clear banks can't compete with these platforms on experience, so Ron is right; they have to recognize the new market and react accordingly. There will be no BAU as branches re-open, the world changed.   The question in my mind is, why would a platform partner with your bank?  There are thousands of banks the platforms could choose from.

  • 🤔 My Analysis: The partner banks that are winning are those that have productized partnerships and compliance. Banks like Cross River, Metabank, Sutton, Evolve, and many more have intentionally gone on the front foot to partner after initial successes with Durbin Debit. The mega-banks may get some partnerships done (e.g., Citi & Google, Goldman, and Apple), but they're not as hungry as the smaller or mid-sized banks.  

  • Paytm is the largest fintech in India, readying itself for IPO. It has an interesting history of pivots from mobile marketing to phone top-ups to e-commerce and more. In 2014 Paytm gained a PPI license from the reserve bank of India, allowing its users to download an app and load it with cash from their bank accounts.  

  • With the invention of UPI in 2016, the competition in mobile wallets increased massively, to where PhonePe (backed by Walmart) now has the largest market share.  Paytm intends to move into a more full-service credit and banking suite over time.

  • 🤔 My Analysis: Any summary I give above won't do justice to Marc's work profiling Paytm and their story, so this is well worth your time reading.  

  • 🤔 My Analysis: What stood out to me was that Paytm's major shareholder is Alibaba, which influenced much of the Paytm strategy. And now, both Ant Financial and Paytm are struggling with local regulations in China and India, respectively.  Government oversight has come to the big fintech wallets.  

  • 🤔 My Analysis: India's strategic assets like Aadhaar and UPI allow it to create a competitive landscape for big fintech and reign them in where needed. Far more effective than senate hearings.

Tweets of the week 🕊

That's all folks. 👋