You have almost certainly used Plaid without knowing it.
Every time you connected a bank account to Venmo, Robinhood, Chime, Cash App, or any of thousands of other financial applications, a small consent screen appeared asking for your bank credentials. You typed them in. The app verified your account and linked to it. The whole process took maybe thirty seconds. The invisible technology making that thirty seconds possible was almost certainly Plaid.
The Plaid success story is a specific kind of story: the infrastructure play. Not a consumer brand, not a product people talk about at dinner, but the plumbing underneath the faucets that the consumer brands depend on. The closest analogy in this series is Stripe for payments or Supabase for databases: deeply technical, deeply necessary, almost entirely invisible to the people whose financial lives it enables.
Zach Perret and William Hockey built that infrastructure from a realization they had while trying to build something completely different. They were going to be consumer fintech founders. They discovered they couldn’t be, because the infrastructure to support consumer fintech didn’t exist. So they built the infrastructure instead.
Two Bain Consultants and the Problem Nobody Had Solved
Zach Perret grew up in Clemmons, North Carolina, a small town where he saw first-hand what limited access to financial services looked like for the people around him. He went to Duke, studied chemistry and biology, and went to Bain & Company’s Atlanta office as an associate after graduation. William Hockey grew up on a farm in central California, studied computer science at Emory, and also landed at Bain in Atlanta.
They met at Bain and recognized each other as the specific type of person who was constitutionally unsuited for large-organization life: both technically-oriented, both restless with consulting work, both enamored with how technology could have a large impact on the world. Perret has described finding Hockey as finding a “kindred spirit.” Hockey has described their relationship as two people who liked building things.
They left Bain and headed to New York, the natural destination for young people who wanted to work in finance but not for banks. They couch-surfed, ran up credit card debt, used borrowed office space. Hockey slept on a friend’s couch for nearly a year. Perret moved in with his now-fiancée after four months together, gambling that the relationship was solid enough to survive the stress of building a startup.
Their initial idea was a consumer personal finance application. The category was active in 2012 and 2013: Mint.com had demonstrated that people wanted tools to understand their financial lives, and there was a generation of startups building budgeting, savings, and financial planning products.
The problem was the plumbing.
Every consumer finance app needed to connect to the user’s bank account to be useful. To see your transactions, you needed to pull data from your bank. To verify your balance, you needed to pull data from your bank. To initiate transfers, you needed to authenticate with your bank. The standard way to do this in 2012 was screen scraping: your app would log into your bank’s online portal using your credentials and extract the data from the HTML of the pages it found there.
Screen scraping was slow, fragile, and broke whenever banks updated their websites. Different banks had different login flows, different HTML structures, different session management. Building integrations with each bank individually was a significant engineering project, and maintaining those integrations over time was a continuous burden. The whole system was also a security nightmare: users were handing their bank credentials to third-party applications and hoping the applications would protect them.
Perret and Hockey hit this wall personally while building their consumer finance app. They realized, in Hockey’s words, that “we were struggling so much of the time because we couldn’t connect anything to financial services.” The fundamental building blocks that should have made it easy to build financial applications simply didn’t exist.
The Pivot from Consumer App to Financial Infrastructure
The insight that became Plaid was sharp and counterintuitive at the time: the problem they had hit while trying to build a consumer finance app was not their problem. It was everyone’s problem. Every developer trying to build fintech applications was struggling with the same fractured, fragile, insecure mess of bank integrations. If you fixed the infrastructure problem rather than just solving it for your own app, you could enable an entire ecosystem of applications that were currently impossible to build well.
Hockey has described the framing this way: “Our mission is, how do you convince a college kid—instead of making Instagram for cats or whatever—to make a fintech app? You do that by making it really easy; you do that by making all the complexity go away.”
The vision was a single API that abstracted away the complexity of connecting to financial institutions. A developer building a fintech app would integrate with Plaid once. Plaid would handle the connections to each individual bank, the authentication flows, the credential management, the data normalization. The developer got a clean, consistent API. Their users got a consistent experience regardless of which bank they used.
Before writing a line of code for that API, they validated the idea by winning the 2013 TechCrunch Disrupt hackathon in Manhattan. The app they built, called Rambler, mapped consumers’ banking activity using the underlying connectivity technology they were developing. Winning the hackathon did two things: it proved the technology worked, and it got them in front of venture capitalists.
They moved to San Francisco in the fall of 2013, raised a $2.8 million seed round from Spark Capital, Google Ventures, and New Enterprise Associates, and started building Plaid properly.
70 Investor Rejections and the First Customer Who Got It
Despite the hackathon win, the fundraising process for Plaid’s initial rounds was difficult. The founders pitched approximately 70 investors. Most said no.
The common objection was category skepticism. Banks were notoriously resistant to sharing data with third parties. The regulatory environment for financial data access was unclear. The consumer fintech market was nascent, which meant Plaid’s potential customers, fintech app developers, were not yet clearly a large and growing market. The business model, charging API calls, was unfamiliar to investors accustomed to SaaS subscriptions or marketplace models.
The investors who did back Plaid early were making a bet on infrastructure ahead of the ecosystem that would need it. NEA partner Aaron Jacobson, who backed the seed round and joined the board, described it as conviction that Plaid had “potential to be a leader in the space, connecting banks and financial institutions to new technologies and applications.”
The first major customer who validated the bet was Venmo.
Venmo was growing fast in 2013 as a peer-to-peer payment app, but the engineering team had a specific problem: verifying whether a user actually had sufficient funds before processing a transaction required real-time access to bank balance information. Venmo was relying on batched transaction settlement, which meant verification happened with a lag. The head of Venmo’s engineering happened to be a friend of Perret and Hockey, and he wanted a better solution.
Plaid’s API gave Venmo what it needed: real-time balance verification at the moment of transaction initiation. For Plaid, Venmo was more than just a customer. It was proof of concept at scale. Millions of Venmo users were connecting bank accounts through Plaid, generating transaction volume that validated the infrastructure worked under real load and demonstrating to other fintech developers that Plaid could handle production-scale usage.
Perret described the strategic significance of Venmo internally as a “great proof point to see tonnes of people using Plaid at scale inside of Venmo, which then helped launch us into a lot of other applications.”
The Developer-First Strategy and the Sell Through the Basement
Plaid’s go-to-market strategy was deliberately bottom-up in a sector that had always sold top-down.
Traditional financial technology sold through enterprise sales: relationships with bank executives, formal procurement processes, contracts that took months to negotiate. This made sense for the large banks and financial institutions that were the traditional buyers of financial technology. It made no sense for the developers who were building the next generation of fintech applications.
Plaid’s strategy, which CEO Zach Perret described explicitly as “selling through the basement,” was to identify the lowest-level engineers at potential customer companies and create the best possible experience for them. If a developer could integrate Plaid in a few hours using public API documentation, build a prototype in their sandbox, and see it working with their bank account, they would advocate for Plaid within their organization. The bottom-up adoption would flow upward to engineering leads, CTOs, and eventually product and business decision-makers.
The API documentation was publicly available on Plaid’s website rather than gated behind a sales inquiry form. This was uncommon among financial infrastructure companies at the time. Most competitors required potential customers to contact sales before accessing documentation, ensuring the sales team could capture the lead and control the evaluation process. Plaid’s approach traded lead capture for developer experience: any developer could evaluate Plaid autonomously, without interacting with a salesperson, before deciding whether to integrate.
The developer-first design philosophy extended beyond documentation. The API was clean and consistent. The sandbox environment allowed testing without connecting to real banks. The error messages were informative rather than cryptic. The integration experience was designed to work in hours rather than days. These were choices that cost engineering effort but produced a product that developers talked about positively to other developers.
Plaid also used an unusual hiring approach during its early period: the company hired by posting positions on its API. Developers who were curious about Plaid’s technology could discover the job postings by interacting with the API itself. The signal it sent was clear and specific: Plaid wanted to hire the kind of developer who used APIs for fun.
Building the Connections to 11,000 Banks
The practical challenge of building Plaid’s infrastructure was enormous and largely invisible to the developers who used the resulting API.
Every bank in the United States had a different online banking portal with a different authentication flow, a different session management approach, a different data format, and a different security model. Getting permission from banks to integrate with their systems was not always possible. The initial approach, screen scraping, required Plaid to manually map each bank’s login flow and data structure and maintain those mappings continuously as banks updated their websites.
Perret described the problem with characteristic directness in several interviews: “Paper is our biggest competitor. It’s the loan officer that you’re bringing your shoebox full of documentation that you printed out to apply for your mortgage.” The digital banking ecosystem was advancing rapidly, but the connection layer between that ecosystem and the institutions holding the actual money was still fundamentally broken.
Plaid’s team spent years doing the manual, unglamorous work of building and maintaining those connections. They developed rigorous security practices to justify the trust of both banks and consumers: tokenizing credentials on the user’s device before transmitting them, building detailed audit logs, maintaining compliance with financial regulations. They engaged proactively with regulators: Perret made early visits to the Consumer Financial Protection Bureau and other regulatory bodies, long before most Silicon Valley companies considered talking to financial regulators. Those conversations eventually led to Plaid hiring its first policy person in Washington.
By the time of the Visa deal announcement in January 2020, Plaid had active connections to over 11,000 financial institutions in the United States. That network of connections was not just a technical achievement. It was a competitive moat that would take years and enormous effort to replicate.
Robinhood, Acorns, and the Fintech Ecosystem That Grew Around Plaid
As the fintech ecosystem exploded in the 2015 to 2020 period, Plaid’s infrastructure was underneath almost all of it.
Robinhood used Plaid to let users link their bank accounts and fund their trading accounts. Acorns used Plaid to connect to the checking accounts from which it pulled spare change. Chime used Plaid to verify banking credentials for its checking account users. Coinbase used Plaid to let crypto buyers fund their purchases. Credit Karma used Plaid to verify accounts. SoFi used Plaid to streamline loan applications.
The growth trajectory of each of these companies was, in a structural sense, dependent on the quality of Plaid’s infrastructure. When Robinhood added millions of users in a short period, those millions of users went through Plaid’s bank connection flow. When Coinbase surged during cryptocurrency price spikes, the bank connection demand flowed through Plaid.
This created a specific dynamic in Plaid’s business: revenue was correlated with the growth of the fintech ecosystem overall. If fintechs grew, Plaid grew. If a specific fintech went from one million users to ten million users, Plaid’s API calls from that company increased proportionally. The per-call pricing model meant Plaid’s revenue was naturally tied to the activity on its network rather than requiring active upselling.
The model also created switching costs for customers that compounded over time. A fintech that had integrated Plaid deeply, built their user onboarding around Plaid’s connection flow, and stored years of transaction history in Plaid’s format had real reasons not to switch providers. The integration cost and the data migration complexity were both deterrents to switching.
NEA’s description of Plaid’s go-to-market as a “beautifully efficient business model” reflected this dynamic: the company was benefiting from demand that was pulling it in from a growing market, with customers that were naturally expanding their own usage, and with switching costs that preserved the customer relationships once established.
Visa’s $5.3 Billion Deal and the Department of Justice That Changed Everything
On January 13, 2020, Visa announced it was acquiring Plaid for $5.3 billion, approximately double Plaid’s then-current private market valuation. Visa CEO Alfred Kelly described the acquisition as “an insurance policy” to deal with what he called a “threat” to the company’s debit business. That description would prove to be the central exhibit in what followed.
Visa’s logic was sound from a strategic perspective. Plaid had access to banking data from over 11,000 institutions. The company was developing technology that could enable direct bank-to-bank payments, bypassing payment networks like Visa entirely. If “pay-by-bank” payment options became widespread, merchants could initiate transactions directly from customers’ bank accounts without the interchange fees that Visa charged. The DOJ’s complaint estimated Plaid’s potential “pay-by-bank” platform could save merchants and consumers hundreds of millions of dollars per year in debit fees, directly at Visa’s expense.
Visa’s solution was to acquire Plaid and control how that technology was used.
The Department of Justice sued to block the acquisition in November 2020, arguing that Visa was a monopolist in online debit transactions and that the acquisition would eliminate a nascent competitive threat before it could succeed. The complaint drew on Kelly’s own words about the “insurance policy” framing, arguing it was direct evidence that Visa was acquiring Plaid specifically to neutralize a competitive threat.
The deal dragged into early 2021. By the time it became clear that the DOJ intended to litigate rather than settle, something significant had changed: the pandemic had accelerated digital finance adoption dramatically, and Plaid’s business had boomed. The $5.3 billion acquisition price that had looked generous in January 2020 looked cheap by December 2020. Both parties agreed to abandon the deal on January 12, 2021.
In April 2021, three months after the deal collapsed, Plaid raised a $425 million Series D round that valued the company at $13.4 billion, more than 2.5 times the Visa acquisition price. The investors who had expected Plaid to be acquired at $5.3 billion were now looking at a company worth more than twice that amount as an independent entity.
Perret described the acquisition period and its collapse as the hardest stretch of his professional life. He also said later that Plaid “wouldn’t have entered into the deal” had it known of the antitrust scrutiny that was coming.
The Valuation Correction, Layoffs, and the Rebuilding
The $13.4 billion valuation in April 2021 was set at the peak of a zero-interest-rate environment in which private technology companies were routinely valued at multiples that would later look unsustainable.
In late 2022 and into 2023, Plaid underwent the same valuation correction that hit most high-growth software companies. The company conducted layoffs in December 2022, cutting approximately 20% of its workforce. Reports suggested the internal valuation for equity purposes had been marked down significantly from the Series D peak.
The correction was painful but not existential. The core business, connecting financial applications to banking data for tens of millions of consumers through thousands of fintech customers, was not broken. The fintech ecosystem had slowed its growth in the post-2021 funding contraction, which meant some of Plaid’s largest customers were growing more slowly. But the fundamental need that Plaid served, enabling financial applications to connect securely to banking data, was not going away.
Plaid used the period to diversify its product surface beyond the original bank connection API. The company launched Plaid Transfer, an ACH payment initiation product that let fintechs move money directly through Plaid rather than through separate payment rails. It added real-time payment capabilities through the RTP network and FedNow, enabling instant fund disbursements. It added identity verification, income verification, and credit-related products that used the banking data it already had access to.
Each new product served customers who were already in the Plaid ecosystem, expanding the average revenue per customer without requiring new customer acquisition. The platform was evolving from a single-product bank connection API to a multi-product financial data platform.
CFPB Rule 1033 and the Open Banking Moment Plaid Was Built For
In October 2024, the Consumer Financial Protection Bureau finalized Rule 1033, a regulation requiring financial institutions to make consumer financial data available to third parties upon consumer request, at no charge. The rule formalized open banking in the United States, establishing a legal framework that had previously been absent.
Plaid had been advocating for this regulatory outcome for years. The company’s head of policy had been working with the CFPB and other regulatory bodies since the early years of the company. Perret had made transparency and consumer data portability central to Plaid’s public positioning throughout the company’s history.
The practical effect of Rule 1033 was to shift the framework for bank data access from a fragile, relationship-dependent, technically messy status quo to a legal entitlement. Consumers would have the right to share their financial data with applications of their choice. Banks would be required to support that sharing through standardized APIs. The screen-scraping workarounds that had been necessary because banks wouldn’t expose proper APIs would become less necessary as banks were legally required to build them.
This was the regulatory moment Plaid had been building toward. The infrastructure for open banking that Plaid had been constructing since 2013 was now being mandated. Every financial institution that needed to comply with Rule 1033 needed exactly the kind of financial data connectivity infrastructure that Plaid had spent a decade building.
Perret described the company’s positioning relative to Rule 1033 with characteristic understatement: the rule creates opportunities in credit scoring, insurance, and other financial services areas where consumer-controlled data portability could drive the same efficiency improvements that it had driven in banking.
The Invisible Plumbing That Became a $13 Billion Company
By 2025, Plaid was processing connections to banking data for over a billion consumer accounts, operating in the United States, Canada, the United Kingdom, and Europe, with relationships across more than 12,000 financial institutions. The company employed over 1,200 people across seven countries.
The fundamental product was unchanged from 2013: a developer integrates Plaid once, and their users can connect their bank accounts through a secure, consistent flow. The technology underlying that flow had been rebuilt multiple times, expanded to support dozens of data types and use cases, and extended to support payment initiation, identity verification, and income verification alongside the original transaction data product.
Perret has described the long-term trajectory in terms that suggest the current state of open banking is still the beginning: “I hope that the industry can get together and build tools [that] create a safer financial ecosystem. Because the trend is towards digitization. Consumers love it. We just need to ensure that we’re in a safe ecosystem as that digitization takes over.”
The IPO that Plaid has been working toward since around 2023 would be one of the more significant public offerings in fintech history: a company that is genuinely critical infrastructure for the US financial ecosystem, invisible to consumers but embedded in the daily financial lives of tens of millions of Americans who use the apps it enables.
What the Plaid Story Is Really About
The Plaid success story is about a problem that was too boring to be interesting and too important to be ignored.
In 2013, the problem of connecting financial applications to bank accounts was a technical nuisance that every fintech developer complained about and nobody had fixed. It was boring infrastructure work. It required dealing with banks, regulators, and legacy authentication systems rather than building consumer-facing features. It generated no end-user enthusiasm. Nobody was going to write a blog post about being excited to link their bank account.
Perret and Hockey chose it anyway, because they had experienced the problem personally while trying to build something they cared about, and because they recognized that if the problem didn’t get fixed, the financial applications that could improve people’s financial lives would remain fragile and hard to build.
The infrastructure play rewards patience in a specific way: the market grows around you as the ecosystem that depends on your infrastructure grows. Every new fintech app that launched using Plaid’s API was evidence that the infrastructure was useful. Every Venmo user who connected their bank account through Plaid was a transaction on the network. Every Robinhood account linked through Plaid was a data point in the growing picture of consumer financial behavior.
The Visa acquisition attempt and its collapse turned out to validate the infrastructure play in the most direct possible way: the DOJ argued, correctly, that Plaid was so strategically important to the financial ecosystem that acquiring it to prevent it from competing was an antitrust violation. The company that had been pitched to 70 skeptical investors in 2013 had become, eight years later, a competitive threat so significant that the largest payment network in the world was willing to pay $5.3 billion as an “insurance policy” to contain it.
The deal collapsed. The company raised at more than double that value independently. The boring infrastructure that two Bain consultants built because they couldn’t find it when they needed it had become the plumbing of American fintech.
Leave a Reply