The Golden Age of Middleware: Why Visa Has Always Been a Great Middleware Company [And Why Buying Plaid Makes Sense]

Travis May
5 min readJan 17, 2020


Author’s Note: I am a huge fan of middleware companies. Datavant is a middleware company connecting health data; my last startup, LiveRamp (now NYSE:RAMP), is the largest marketing middleware company; and I am an investor in a number of middleware companies across industries. The Visa/Plaid combination is a great example of what makes the middleware space so compelling, which I’ll walk through below.


Visa’s acquisition of payments company Plaid for $5.3 billion this week has cast a spotlight on an often-ignored part of tech: middleware. Though little understood, middleware companies make strong, defensible businesses, and can accelerate innovation across entire industries.

Plaid is a great middleware company. Wait, what is middleware again?

Middleware companies are connectivity companies. They build the interfaces, integrations and solutions that make it easy for different players in an industry to connect and collaborate around data. Middleware remains one of the least visible and least sexy classes of technology companies, but, as I noted after Salesforce’s acquisition of middleware company Mulesoft, middleware is entering a golden age.

For new financial technology companies, Plaid has built the underlying connectivity tools that make it easier for new entrants to “plug in” to the existing financial ecosystem, including exchanging transaction information with banks, credit card companies, and merchants. As Visa CEO Al Kelly noted in the conference call after the acquisition announcement, Visa aspires to be focused not only on payments, but “on the movement of funds for any purpose around the world.”

Life without Plaid would be slow (and full of friction)

Today, when a new fintech company like Venmo is founded, it can use the Plaid technology to avoid needing to build a direct technical integration with every bank in the country. To understand the value of middleware, it might be helpful to imagine a universe without Plaid where a new company is trying to launch its own financial application.

In the Plaid-less universe, that company would need to form individual relationships with and build technical connections to every bank, credit card company, and merchant where its customers would like to transact. The task is enough to keep many dozens of business development managers, lawyers, operations managers and engineers busy for years, which requires capital and time that is not available to most start-ups.

In that world, each new financial application would aspire to build its own proprietary version of Plaid. This is bad for everyone:

  • It’s bad for the application company, which is primarily focused on providing value to customers, not building out a back-end financial technology ecosystem
  • It’s bad for the financial institutions that have to work with all of those applications (each new application is an independent relationship and a new technical implementation)
  • And, most importantly, it’s bad for end customers, who would see much slower innovation as companies struggle to independently build the necessary underlying ecosystem (and spend less time on their customer-facing applications)

Despite these dynamics, this is how many industries have worked prior to the emergence of dominant middleware companies, from finance to education to healthcare.

Visa Has Always Been a Middleware Company. Plaid is just Visa 2.0.

Looked at from the right angle, Visa has always been a middleware company.

Unlike a bank, Visa does not actually extend credit to consumers. Instead, Visa acts as a middleman between the financial institutions that issue credit cards, the merchants who sell to consumers with credit cards, and the merchants’ banks. The key to Visa’s business model is that it has built a vast network of merchants and financial institutions, and it specializes in the payment mechanics (“plumbing”) necessary to enable those relationships. That includes things like authenticating borrowers, settling transactions, and facilitating the flow of funds.

It’s important to understand that not only would it be extremely expensive and redundant for each major bank to build its own payment network, but any given bank is also strategically ill-suited to do so. Whereas banks compete directly with the other banks that they would like to work with, credit card companies remain neutral, allowing them to work with everyone. No one bank would be able to obtain the ubiquity that Visa has obtained today.

Why Plaid is Dramatically Undervalued

$5.3 billion is almost certainly a hefty revenue multiple for Plaid. But the acquisition was well worth it for Visa, and if anything undervalues the power of Plaid’s network.

For many years, credit cards were the only real way to conduct a cashless transaction. Today, however, the market is shifting because of applications like Venmo, which allow customers to transfer funds directly from one account to another. New ways of transferring funds have also led to new payment networks springing up, which is where Plaid comes in.

According to the announcement, one-in-four people with a bank account have used Plaid (usually without knowing it) because Plaid is working with thousands of financial application developers and over 11,000 financial institutions. For a company founded in 2013, that is a massive network, and given Visa’s business, it makes complete sense for Visa to take it over.

The key thing to understand about middleware companies is that they are, in the truest sense, network effect businesses. Great technology is necessary, but not sufficient. Strong financials are necessary, but not sufficient. Middleware companies only work when they are ubiquitous among the relevant companies in an industry. Once a company obtains that position, it sets off a positive feedback loop where the company’s network makes its connectivity technology more and more valuable until it becomes table stakes for any new entrant that wants to work in the sector.

Visa built that flywheel for credit cards, and is up over 500% since its initial public offering in 2008. Plaid has now built its own flywheel for financial technology on a complementary network, and is well-positioned to follow a similar growth path.

Why Middleware Is Eating the World

While companies like Mulesoft and Plaid (and older companies like Visa) have proven the middleware model in various markets, more and more examples of large middleware companies are popping up — and I believe the next $100 billion enterprise technology company will likely be a middleware company.

For the last 40 years, one of the fundamental problems in enterprise technology has been the connectivity of data within an enterprise. Companies like Oracle and Salesforce emerged to satisfy that need. The next generation of problem is about the connectivity of data across enterprises: the core function of middleware, and an exponentially stickier, more defensible, and more valuable use case. The explosion of data across sectors and the explosion of point-solution SAAS applications has made cross-organizational data connectivity a huge challenge in every major part of the economy.

Thanks to Bob Borek for helping to draft this post.



Travis May

Entrepreneur, Investor, and Board Member. Founder & Fmr CEO of LiveRamp and Datavant.