Why haven’t we arrived at the AWS Era for Fintech?

20230720 LaunchPad GA SQ 2
Gil Akos
Co-founder, CEO

What the architecture of the Cloud predicts for the next wave of Financial Technology.

With key infrastructure providers maturing, simultaneously “Everything is Fintech” and financial technology is transitioning into the Third Wave of technical innovation. If we trace this progression from data to infrastructure to platforms within the Cloud, we can see a clear parallel to the early days of Amazon Web Services and look for an equivalent for Fintech today. But this thought experiment leads to more questions than answers. How much of AWS do we have in place for Fintech? What are the core technologies it would include?

This article evaluates the key technological elements of AWS and places them in the context of financial services, drawing three conclusions:

  • We have yet to leverage the power of both abstraction and virtualization in Fintech.
  • We have only started to build out the Infrastructure layer of the tech stack.
  • The AWS Era for Fintech is not here... yet!

Don’t be discouraged though – the fact that an AWS Era isn't here yet is actually an optimistic position! It means the architecture of the Cloud should give us a detailed model for how we might enable fintech products and services that we haven’t even yet imagined.

Although "the Cloud" and Cloud Computing have origins that reach back to the early days of the Internet, since its launch in 2006, Amazon Web Services has become the standard for building and scaling applications on the web. It has the volume of Google Cloud Platform, Microsoft Azure, and IBM Softlayer combined; it's grown to a $40B run rate; and, were it to be a standalone company, it could be one of the 10 most highly valued companies in the world.

Not only has Amazon Web Services achieved a high level of enterprise and technological success, but it has made reliable web services so abundant and affordable that an explosion of internet-based applications has radically changed almost every part of our modern lives. Or in other words, “software is eating the world” – arguably this is only possible because of platforms like AWS. With the optimism around what similar positive second-order effects might be enabled by a true platform for finance and with all the success achieved by Amazon, it’s natural to seek an equivalent to Amazon Web Services for Fintech.

As the clear leader in enabling other applications on the internet, Amazon Web Services is the perfect model for evaluating where we are in the evolution of financial technology and articulating what we still need to really power the next generation of products in finance. Although the technical mediums vary between AWS and Fintech, there is a strong analogy between the layers of the technology stack and the opportunity a Fintech platform might enable. So what would the AWS Era look like?

Let’s start with the impact of AWS by rewinding 20 years to the turn of the century. The internet was growing and great companies like Amazon.com, Netflix, and Google were finding early success. But to connect to the web and operate a website where consumers could engage your product, you had to buy hardware, then develop infrastructure for that hardware, then scale that infrastructure as you found success. You had to own and manage your whole stack, at great effort and expense. In short, you had to follow the model of "one application, one server".

The Cloud broke that one-to-one requirement and the capital overhead of growing your application. Incrementally, cloud computing enabled you to move your "on-prem" servers to a Data Center to rent those resources instead of own them. You could then use early web infrastructure companies, like Opsware, to manage those servers at scale. If your business required operational excellence and you built an infrastructure that efficiently scaled, you might find an opportunity to expose that infrastructure to third parties as a service.

The Architecture of the Cloud | Adapted from diagram by MongoDB

At each step in this evolution, a technological service layer was built that abstracted the key function of the prior layer, generalizing it for use and enabling that layer to scale. Data Centers abstract your hosting needs, offering scale to your web applications by allowing you to rent more hardware as needed. Infrastructure-as-a-Service or "IaaS" abstracts the hardware, scaling your needs through an operating system for that hardware and so on.

The ability build on top of a service layer not only means that software can scale effectively and economically but also means that the provider of that service layer can generate wholly new sources of revenue while powering the growth of entire ecosystems.

At level i+1 of the stack, the newly valuable resource is that which emerges as scarce out of the transition from scarcity to abstraction to abundance at layer i.

Alex Danco

Amazon Web Services was the first to offer a robust infrastructure layer for the Internet and grew into the de facto Platform-as-a-Service juggernaut that it is today. As such, its success was inextricably linked to the growth of applications on the web. More applications mean more AWS clients. More clients equal more scalable infrastructure and cost savings. All resulting in better and cheaper service for all of those applications in the first place. And in the end, more applications are created – the perfect Amazon flywheel.

The Amazon Flywheel | Recreated from diagram by Ben Thompson, Stratechery

If you believe companies will build applications from scratch on top of the infrastructure services if the right selection (of services) existed, and we believed they would if the right selection existed, then the operating system becomes the internet, which is really different from what had been the case for the (previous) 30 years.

Andy Jassy, CEO of AWS

What key technical components enabled the AWS flywheel?

The start of the answer lies with the two core services comprising AWS at launch. The very first versions of Amazon Web Services included two key infrastructure elements, Simple Storage Service (S3) and Elastic Compute Cloud (EC2). S3 and EC2 offered a cost-effective solution to the two most important features needed in the cloud: storage and execution (aka compute). Both relied on the abstraction of the corresponding physical hardware in the data center to meet the needs of web applications. Abstraction allows for a common data model across devices and consistent accessibility for other technical elements to systematically interact with that infrastructure.

Virtualization and Abstraction for the Cloud and Fintech

But that's just half the story – virtualization is the other key component to what made S3 and EC2 so powerful. Through a "virtual" operating system, one storage device or processing unit can interact with data from multiple applications simultaneously. Similarly, to balance the load of high traffic, the needs of one growing application can be delivered through multiple servers, dynamically and on-demand. Together this changes the equation from "one application, one server" to "many applications, one (abstracted) server" to "many applications, many (virtualized) servers".

In concert, abstraction and virtualization mean that a web service can operate agnostically to the physical device in the data center, pool resources across hardware, and partition or expand capabilities as needed. This breadth of functionality defined the early success of AWS by ensuring that their service was extremely flexible, reliable, and scalable.

What would define an equivalent platform for financial technology?

In order to construct a true AWS model for Fintech that is able to leverage the technological advances that come with both abstraction and virtualization, we have to start with the core atomic units at the bottom of the stack. What is the financial equivalent to the data, the utility of storage, and the execution that makes up a Platform-as-a-Service? Using S3 and EC2 as a guide, we can see that the two key service equivalents are still to be defined:

The Cloud vs Fintech Architecture

The medium of finance is money, not pure data. So where the Cloud allows us to store and access bits, bank accounts act as the storage device to the dollars that are accumulated in bank accounts and processed through the financial system via transfers. But Fintech is missing these two key elements at the virtualized level above accounts and transfers.

If you believe this model from the Cloud to be true for Fintech, then the evolution of the capabilities of financial technology and what is need to seize the equivalent opportunity become clear:

A technical model for Virtualization and Abstraction in Fintech

1. We have yet to leverage the power of both abstraction and virtualization in Fintech

We may have converging data models for how a bank account should be consistently defined and digitally accessed through great APIs, like our partner Plaid, but we are missing the ability to logically partition or pool storage resources. Imagine a savings app that doesn't come with a new bank account but actually operates on top of your existing one!

Similarly, there are robust APIs, like that from our partner Dwolla, for facilitating the flow of funds between banks, but we are missing the ability to apply business logic to the execution of that processing universally across institutions. What might an operating system full of transfer routines enable for new financial services?

The Architecture of the Cloud vs Fintech

2. We have only started to build out the Infrastructure layer of the tech stack

Understanding the technical architecture of AWS gives us a clear lens to evaluate the evolution of financial technology. At the lowest level in the stack, Banking-as-a-Service is here and expanding from current players like Evolve to new entrants like Goldman Sachs. Key infrastructure pieces are in place with more and more options coming online from maturing fintech startups like Marqeta and Stripe. But we are still missing essential technical building blocks (see 1) to truly move up the stack.

A fun exercise to illustrate the room to run at the Infrastructure-as-a-Service layer: open the product page for AWS and try to match a fintech offering to the equivalent category, much less individual product. For instance, Compute (EC2) has 14 products listed in that category alone. There is still so much to build!

The Fintech Platform Stack | Expanded from diagram by Chris McCann

3. The AWS Era for Fintech is not here... yet!
The layers of the tech stack are really only relevant within the ecosystem they power. Mapping that of web applications and our model of AWS in an ideal case shows us what might be. Currently, a developer building a Fintech app needs to engage down to at least the Infrastructure layer and most likely the Banking layer just to create the first version of their application.

Only when we have a fully matured Infrastructure-as-a-Service layer and the Platform that abstracts it, will we be able to empower a similar abundance of products and services in financial technology as we’ve seen via the Cloud. We will explore more of what that means for consumers and businesses in our next post.

Amazon Web Services provides us with a robust framework for what we might achieve with financial technology. AWS as a model clearly articulates the scope and ambitions of what might be, from the perspective of the creation of enterprise value, the possibilities for technical innovation, and the range of future businesses that it might empower. There is plenty of potential energy accumulated in our industry, but we are missing the key elements that will make all that’s possible a reality.

Platforms will get us there. But we are just getting started!