Long Take: Marqeta's $300MM of revenue & Ethereum's $20B in ann. transaction fees highlight opportunity and industry structure
Hi Fintech futurists --
This week, we cover these ideas:
How market structure determines the types of companies and projects that succeed
A walk through Marqeta’s economics and business model, and how Square’s Cash App and DoorDash were needed for success
The emerging $10B transaction revenue pool on Ethereum, MEV, and the changes to mining and gas
For more analysis parsing frontier developments every week, podcast conversations on operating fintechs, and new mental models, become a Blueprint member below.
In Partnership with:
Get One Month of Extracting Investment Signals Free with a Year Subscription
With Accern. You can easily draw out key themes, sentiment and insights from unstructured and structured data such a news, financial filings, investment research and more- efficiently and accurately. Check it out here.
There are several fantastic opportunities at ConsenSys. Join the team as:
DeFi Protocol Economist: design tokenomic networks
DeFi Asset Management Product Manager: build DeFi asset management
There’s a lot to be said for sticking around long enough.
You can say it like a capitalist. It’s important to have exposure to your opportunity and asset class.
You can say it like a shaman. If you listen to what is happening, you will learn what you have to do.
In that spirit, let’s look at how industry structure generates business opportunities — and in particular at Marqeta, and its S-1, and Ethereum mining, and the issues around trading and upcoming software changes.
So the Marqeta SEC filings are out and public, and fintech analysts are digging through the document to understand how this company is trading at $16 billion on the private markets. The company did about $300 million in revenue last year, and is annualizing to $450 million this year. Let’s round all the numbers, and say that is a forward looking revenue multiple of about 30x, and a backward looking one of 50x. All things equal, that range is consistent with our prior dives on payment and embedded finance darlings like Checkout.com, Rapyd, and Affirm. In the public markets, we would expect to see 10x on revenue, and in the private markets, things could float up to 100x.
Also, let’s kick the tires on economic engine. Below, you can see how American interchange splits out for debit card issuance, with 20 bps going to Marqeta courtesy of James Ho.
You can also take a naive pass, like this.
In 2020, there was $60 billion of volume processed, and in 2019 it was $20 billion or so. The associated revenue numbers are $290 million, and $140 million. That’s about 50 and 70 basis points. About 60% of revenue is paid out as “costs of revenue”, which “consist of Card Network costs, Issuing Bank costs, and card fulfillment costs.” So we get a sense that the take home rate is around 20 to 30 basis points.
What is Marqeta again, and why did it have to wait around for the market?
Marqeta issues cards programmatically on behalf of third parties. Those third parties have users as clients. While the firm claims to be the first of its kind, we recommend you take a listen to our conversation with Anil Aggarwal of TxVia and Clarity Money about the evolution of issuing processing. Here’s a useful definition:
Acquiring processing, the easiest way to think about it is that is a merchant or retailer accepting a payment. How do they do that? How do they get money basically in? It could be through a payment network, traditional one, or some other method. Issuing processing is kind of the other side of the transaction where that's where your balance is stored. That's where you make a payment from.
Issuing processing generally refers to more of the cardholder or consumer side of the transaction where they have a form factor of some kind to make a payment, and then they interact with the acquiring side to facilitate that payment. They're two really distinct sides of the transaction.
So companies like Stripe help businesses accept payments by giving developers ways to integrate checkout options through APIs. That’s not where Marqeta comes in, at least not in the beginning, despite having lots of pictures of APIs on its website. Rather, it helps developers with card issuing inside of their applications on behalf of people. It is a feature of your digital wallet — the digital card feature.
Their large relationships are Square, Doordash, and Klarna. In fact, Square is about 60-70% of the entire business. But isn’t Square for merchants, you say? Well, yeah. That’s a different part of Square. We are talking about Cash App, which is either peer-to-peer payments or a mobile wallet. And the Marqeta card is that little bit of magic that allows Cash App to generate what immediately feel like bank accounts. Are they bank accounts? Or some sort of cash account? We understand it to be creating pre-paid debit cards, which are sometimes loaded with debt that another start-up is willing to assume. Here’s an integrated flow:
Of course, you can’t just do this — you need a bank. In this case, an issuing bank behind the card.
Good thing there are interchange fees!
The operating flow with DoorDash is also really interesting. You’ve got people ordering food and paying DoorDash, and then you’ve got contractors that work for DoorDash that have to get to a restaurant, pick up the food, and match the payments between the restaurant and DoorDash, and deliver it to the consumer. Marqeta creates digital card numbers for the contractors that are funded exactly to the amount of the order: “card issuing works in the background as money moves from the app to the delivery driver’s payment card, allowing the driver to pay for exactly what you ordered, and nothing else.”
And this takes us to the core point of the write-up, which is not that Marqeta is an awesome company (which it is), but that it is a good example of what has to go right for a company to be successful.
Marqeta was founded in 2010. Here’s a chart of eCommerce from then until now.
There are two sides to the commerce equation: consumer and merchant. Robots, like Stripe, have infiltrated the merchant side. As merchants become cyborgs with web and mobile storefronts, they need digital. And robots, like Marqeta, can now infiltrate the consumer side, which has also cyborged itself into neobanks and roboadvisors. What perhaps is one of the more telling examples is that JP Morgan and Goldman Sachs Marcus, two financial institutions with a lot of capability, are using Marqeta for the issuance of virtual cards to their own customers.
This company could not exist without the flourishing of the digital economy, the peer-to-peer payments apps that it powers, the desperate digital transformation of the banks, or the Silicon Valley borganization of food and shopping services during the Covid pandemic. Nor could it exist without the American economics of interchange payments, which power everything from Chime to Aspiration. Interchange hides the cost to the customer inside transaction flow, thereby structurally incentivizing money-in-motion companies that monetize on turnover.
Marqeta had to have fantastic execution. But more importantly, it had to understand the field it was playing in, and provide the right service to the right company in the right form factor. It’s large clients didn’t even exist when it was founded; Cash App and DoorDash both launched in 2013, 3 years after Marqeta was born. So don’t feel bad about spending some time in the valley of despair.
The Crypto take
We sort of think that putting an API on a traditional bank is a stop gap to real money innovation.
That means in the long run, APIs like Marqeta should create crypto addresses, not bank accounts. But like with 2010 Marqeta, where the market has not articulated yet what certain fees were for, there are certain parts of the crypto value chain whose value is only now being articulated and clarified.
Let’s focus on Ethereum gas fees. Gas is paid for computation on the Ethereum network, and as price rises and usage accumulates, fees go up. There is some portion of the fee that is subsidized by the network through issuance of new ETH, and some part of the fee that derives from a transaction prioritization auction. You can see the key statistics below.
Over the last three years, transaction fees have become a much larger source of economics for miners: from about 5-10% at the height of 2017, to around 50% today. Users are bidding up how much they are willing to pay to use Ethereum — in large part because they have become wealthy through recent capital gains, or represent institutional interests. About $1.5 billion of fees per month, or nearly $20 billion of fees per year, are generated by the network.
ETH holders are both proud and horrified by this statistic. That it is so high makes the network more valuable under a cash flow analysis. A $400 billion valuation on $20 billion of fees is a modest 20x on revenue!
And yet, it is so high that new users cannot interact with the network. To this end, we see scaling solutions like Polygon, Optimism, and others plug into the base layer and expand its functionality.
The first learning is that there are likely to be multiple lanes for different types of financial participants. Polygon today is worth around a $10 billion marketcap, and it provides much more affordable transactions while inter-operating with Ethereum. Without Ethereum’s hyper-popularity and increasing usage, Polygon would not be valuable, nor would it have financial activity to absorb. Like Marqeta, which needed its Cash App, Polygon (and similar projects) allows a non-institutional Ethereum user to remain a participant.
This implies a value settlement lane for institutional or high-net-worth players, and a second, slightly less secure one, for retail. Ironically, it recreates the types of distinctions we saw when evaluating the Federal Reserve — and its ACH and FedWire systems — as a payment processor.
The second learning is that there is about $200 billion of enterprise value currently attributed to Ethereum that is derived from transaction fees — and those relate in part to transaction ordering and arbitrage. We are being a bit simplistic, true. But if 50% of onchain revenue is from transaction fees, and those transaction fees are really high because users are bidding up fees to go faster, then … we are not entirely wrong.
Without going down the rabbit hole, we point to (1) the Flashbots maximal-extractable-value project, and (2) the change in fee generation due to the shifts to proof-of-stake via ETH2 and the upcoming implementation of the EIP-1559 proposal. Flashbots highlights the emerging market for algorithmic trading around the “re-ordering, insertion or censorship of transactions within a block being produced”, with value estimates below.
Enormous companies can and will be built out of this attribute of the Ethereum economy, as trading and payment volumes continues to shift on-chain, and as these systems become mission critical for the world.
Further, as the software itself changes and transforms to look more like a financial instrument with 1-5% annual rates of return, rather than an army of overclocked server farms, we are going to see changes in the aforementioned revenue pools. The implementation of gas is splitting into a “base” fee that is removed from circulation to offset ETH inflation, and a “tips” fee that rewards miners/validators for accelerating the ordering within blocks.
The willingness to pay for arbitrage won’t go away, just because you re-format the structure of the market. But like our previous discussion of interchange, the shape of money flows will determine what kind of financial super-structures you can build on top.
Stick around long enough, and you will start seeing and capitalizing on these opportunities.
For more analysis parsing 12 frontier technology developments every week, a podcast conversation on fintechs, and food-for-thought essays, become a Blueprint member.
Our NFTs, embedded with premium subscription access, are available for purchase.