Part IThe Story
The Wrong Column
The meeting in Addis Ababa was supposed to go the way these meetings go. By the summer of 2025, William Hockey had made a habit of touring African capitals in search of local banking partners for Column, the company he owns, runs, and has never once asked a venture capitalist to fund. The choreography had settled into something close to liturgy. A local ambassador would line up a few introductions. Hockey and his team would arrive in suits. Their hosts would remark, with varying degrees of courtesy, on how young the Americans looked. The Americans would answer by reciting arcane facts about the host country's financial system, facts that most of the country's own bankers would have to look up. There would be a formal dinner. There would be a handshake. That was the deal. It was usually the deal.
This time he sat down and knew from the faces across the table that something had gone badly wrong. The bank's leader put an article in front of him. In Hockey's telling, he slammed it down. Then came the question: how dare he show his face after laundering money for drug cartels?
Hockey read quickly. The piece was an opinion column about TD Bank, which had paid billions of dollars in fines for precisely that kind of malfeasance. Somewhere in the machinery of syndication, the headline had picked up the word "column," lowercase and entirely innocent. A bank-monitoring service, doing what such services are paid to do, had matched the word against the name of his company and flagged it. The hosts had done their diligence. The diligence was wrong. Nothing he said could get the room back. "We tried," he would say later, "but I got up after 45 minutes and had to walk out."
"It was so physically uncomfortable," he told the reporter Alex Konrad. That September, sitting in a planning session with his top lieutenants, he raised the incident without being asked. "I'm still traumatized," he said, and then turned to explain to anyone who hadn't heard. Column later spent months lobbying the outlet to amend its headline and eventually succeeded, which helped not at all on the day in question. Column does not work with that bank. Perhaps, Hockey jokes, they still think he's a money launderer.
The episode has a particular comedy and a particular cruelty, and both deserve attention. Hockey is thirty-seven. He has spent his entire adult life building things that ordinary people do not see. The first was Plaid, the software that sits between your bank and the app on your phone, which roughly one in three American adults have used, according to the company, and which almost none of them could describe. The second is Column, a nationally chartered bank that sits beneath Plaid's layer, beneath the fintechs, beneath nearly everything. In Addis Ababa this builder of invisible machinery was undone by the most visible thing he owns, his company's name, and by a common English noun that he could not trademark out of the language.
There is a further twist that a novelist would cut for being too neat. This was not the first time TD Bank had crossed his path. In October 2020 TD sued Plaid, accusing the company of trying to "dupe" its customers. The bank that once called Hockey's first company a deceiver had now, by syndicated accident, gotten his second company mistaken for a criminal enterprise. It is difficult to build plumbing for the money of strangers without someone, somewhere, wondering what is in the pipes.
By the Numbers
Two Companies, Two Theories
70Investors who reportedly passed on the original Plaid idea
$5.3BVisa's January 2020 offer for Plaid, later blocked by the DOJ
$13.4BPlaid's April 2021 Series D valuation, its all-time high
$1B+Plaid shares Hockey pledged to secure a $70M loan to buy a bank
~$200MColumn's reported annual revenue, per Upstarts Media (Feb 2026)
25%Share of Column's annual earnings used to buy back employee equity
$0Outside venture capital raised by Column
He tends to start with the farm. "So I grew up on a farm out here in central California," he told Harry Stebbings on 20VC in December 2022, "and I grew up building everything." He was born on September 21, 1989, in San Luis Obispo, a county where the Pacific fog gives way to vineyards and grazing land. The adults around him did things with their hands for a living: farmers, welders, craftspeople. Emory's alumni magazine, which has written about him with the warm, slightly breathless pride of a school that knows a good story when it graduates one, describes a boy fascinated by the fact that if you broke something down, you could figure out how it worked.
The detail matters because Hockey keeps coming back to it, the way some founders keep returning to a dorm room or a garage. To him, the magazine reported, computer science never felt much different from farming or welding. Planting a seed, fusing two pieces of metal, writing a line of code: it was all the same work, making a thing that was not there before. On 20VC he described programming, which he took up in college, as a socially acceptable way to keep building things. It is a revealing way to put it. It suggests a young man who sensed that the welding torch was not going to follow him into the professional class and who went looking for an equivalent the professional class would accept.
He wanted out of California. He had spent little time on the East Coast, so he visited Atlanta, found Emory's campus beautiful and its computer science program strong, and applied there on what the school calls a whim. It was his only application. "I got in and never looked back." For a man who would later borrow $70 million against nearly everything he owned to buy a bank that no investor would have let him buy, the single application reads in retrospect like an early rehearsal. Concentrate the bet. Don't hedge. Go.
At Emory he pursued computer science alongside a business degree from the Goizueta Business School and graduated with a BBA in 2012. He remembers the place as permission. "You don't have to worry about anything actually being pragmatic or 'real world,'" he said. "You could just kind of build things and create things just for the sake of it, which was really beautiful."
Then came a summer internship at Bain & Company, in the Atlanta office, where he met Zach Perret. Perret was a Duke graduate and a fellow junior consultant who shared Hockey's interest in where financial services and technology overlapped. He would become Plaid's chief executive and remains so today. Where Hockey was the one who later disappeared into code and board seats, Perret became the public face: the one who signed the letters, sat for the CNBC segments, and wrote the note announcing the Visa deal. The two became close friends before they became co-founders, and both facts are plainly important to Hockey. Years later, he would call Perret "one of my best friends" in the same sentence where he described writing Plaid's first lines of code during his senior year of college.
It is a little funny, and not unlike Hockey, that the record cannot agree on exactly when Plaid began. Wikipedia's company infobox says May 2013. Its history section says 2012. TechCrunch says 2012, and Contrary Research says 2013. A PYMNTS article, describing the company in 2021, puts its founding in 2016, which is certainly wrong. The confusion is understandable. Plaid did not so much launch as gradually stop being something else.
Seventy Investors and a Hackathon
The something else was a consumer app. Perret and Hockey wanted to build tools for ordinary people: budgeting software, bookkeeping software, products born from their own irritation at how little visibility they had into the bills they paid. They kept hitting the same wall. Every bank was a sealed box, every credit union its own dialect, and connecting any of them to a piece of software meant building a custom integration from scratch, again and again, with no assurance it would still work the following week.
"We started to realize we were struggling so much of the time because we couldn't connect anything to financial services," Hockey said in 2018. "The fundamental building blocks that make it really easy to build an application, we just couldn't find."
So they decided to make the building blocks. Investors were not enthusiastic. By Contrary's count the pair pitched the idea to seventy of them, and none were interested. A unified API for bank data was not a story anyone wanted to hear in 2013. It sounded like plumbing, and plumbing, in a season of social apps and photo filters, sounded like a lifestyle business at best.
Their answer was to stop explaining and start demonstrating. In April 2013 they entered the TechCrunch Disrupt hackathon in Manhattan with an app called Rambler, which mapped a user's banking activity, and on April 28 Rambler won the grand prize. The repository is still public on Hockey's GitHub page, filed under its full name, Rambler.io. It is JavaScript, with five stars and six forks. Rambler itself went nowhere. The pipework underneath it became Plaid.
Their early backers at NEA remember the pitch as three words: Make money easier. Aaron Jacobson, the NEA partner who joined Plaid's early board, later wrote that his firm had first invested in "a team of three people (and a dog!)." In September 2013 the three people and the dog closed a $2.8 million seed round from Spark Capital, Google Ventures, and New Enterprise Associates. That fall they moved the company to San Francisco to find engineers.
Their first important customer came through friendship. Venmo's head of engineering knew the founders and had a problem. Venmo users saw their payments happen instantly, but the transactions actually settled in large batches the next day, which meant the app had to trust that a sender had the money without being able to check. Plaid let Venmo check. A balance verified in real time is a small thing. It is also, multiplied across tens of millions of transactions, the difference between a payments company and a fraud magnet.
Hockey was frank, almost cheerfully so, about how little the founders expected. "You should never start a business to make money," he told a packed room at the Emory Entrepreneurship Summit in April 2019. "That never entered our consciousness. We were like, we're going to build something that's going to fail and then we're going down." They found the work interesting, fun, different. "And we kind of fell in this hole of financial service."
He meant the hole as a compliment. People regularly rate financial stress as the greatest stress in their lives, above health, he told the Emory audience. "Yet nobody's ever said, 'Oh my God, the bank changed my life.' And we wanted to change that." The ambition was enormous and the method was self-effacing. Plaid would not change anyone's life directly. It would change the lives of developers, who would change the lives of users, who would never know Plaid was there. "We don't directly build applications for consumers," Hockey explained. "Instead, we create the building blocks for other people to build amazing things on."
He thought about his customers the way a farmer thinks about a field. "Tens of thousands of applications have been built off it now," he said in 2020. "One percent of them are going to be successful, but that one 1% are going to totally change how the world operates."
The company grew on that thesis: a Series A of $12.5 million from NEA in November 2014, a $44 million Series B led by Goldman Sachs Investment Partners in June 2016. Forbes put both founders on its 30 Under 30 list for enterprise technology in 2015 and later ran a profile under a headline that stuck: "Plaid: Fintech's Happy Plumbers." The founders personally recruited and interviewed the first hundred employees. More than ten percent of the company eventually worked in the People function, the highest share Jacobson said he had ever seen. And there was the matter of taste. Jacobson remembered Plaid's obsessive attention to design, in its website, its documentation, its office, and its swag, as "something that struck me as unique and unexpected for a tiny fintech infrastructure company." He declined to say which founder deserved the credit. "I'll let you decide which of the founders is the tastemaker," he wrote, and added a winking emoji, which is about as close as a venture blog post comes to a raised eyebrow.
A Logo on the Pipe
The conventional wisdom about infrastructure is that it should disappear. Jeff Lawson of Twilio liked to boast about how many people used Twilio without realizing it. Whitelabel is a selling point. The customer owns the experience, and the plumber stays in the basement.
Plaid started out that way. When an app like Venmo or Square Cash or Chime or Coinbase asked users to connect a bank account, each built its own screen for doing it, and Plaid ran quietly behind every one of them. Users had no idea the company existed. Hockey came to see this as a serious problem, and his reasons are worth hearing in full, because this was the decision he believes made Plaid's fortune.
— William Hockey, on the Cartoon Avatars podcastI think the most successful decision we made was actually owning the interface - the physical design and owning the client side... We made the application insert our branding, our logo and our experience into the application. That was extraordinarily controversial, as you can imagine, because these applications want to control the experience.
Controversial is putting it mildly. Plaid told every customer, including the venture-backed darlings of the fintech boom whose growth teams treated onboarding screens as sacred ground, that they had to replace their carefully tuned flows with a screen Plaid designed, carrying Plaid's name. Hockey has said the rollout took eighteen to twenty-four months. There was, he said, "a lot of pushback." It applied to everyone. Today, he says, one hundred percent of traffic flows through it.
His argument had two parts, and they pointed in different directions. The first was about the customer. Every app thought its own design was best, and every app was optimizing for its own conversion, but Plaid was the only party looking across all of them, "the only one focused this hard on conversion." A uniform screen, it turned out, converted better. Users began to recognize it: I know this screen, I've seen this before. Plaid could fine-tune the messaging for particular banks and push updates instantly. The second part was about Plaid. Without a relationship with the consumer, Hockey said, "we also would have got commoditized and it would have been really easy for these applications to switch it out." The screen was a moat dressed as a courtesy.
There was a third benefit that he says matters most to the banks. Under the new flow, sensitive credentials never touched the application. If a bank wanted users to accept new terms of service, Plaid could deploy the change everywhere at once. Hockey has said this is "one of the only reasons that we have good relationships with the banks." His own estimate of what the decision was worth, offered on a podcast in a way that does not seem like a boast, was ninety percent of the company's value.
What the estimate leaves out is that the same screen became the center of Plaid's worst years. To connect a bank account immediately, users typed their bank login credentials into an interface that was not their bank's. Critics said it looked like one. Plaid was accused of scraping data, impersonating bank login pages, and failing to disclose the privacy risks. TD Bank's 2020 lawsuit made the "dupe" accusation. In 2021 Plaid paid $58 million to settle a consolidated class action, five lawsuits combined into one, alleging that it used consumers' credentials to collect and pass along detailed financial data without consent. The plaintiffs said Plaid had "exploited its position as middleman." About 98 million people fell within the settlement class. Claimants could have their payments sent automatically through PayPal or Venmo. That is to say, through the very apps that ran on Plaid.
Both stories are true at once, which is what makes them interesting. The decision to put a logo on the pipe made Plaid hard to replace and pleasant to use. It also made Plaid visible, and therefore something people could blame. In October 2020 the company announced an "all-new Plaid Link" with fewer steps. It signed data-sharing agreements with JPMorgan Chase and, in May 2021, with U.S. Bank, under which some users would authorize access through their own bank's portal and Plaid would never see their credentials. The plumber had come up from the basement. Some residents were grateful and some called the police. Hockey's lasting lesson was not to go back downstairs.
The Art of Leaving the Room
By the end of 2018 Plaid was, by any external measure, a triumph. On December 11 it announced a $250 million Series C at a $2.65 billion valuation. The round was led by Mary Meeker, whose annual Internet Trends report had for two decades served as Silicon Valley's state-of-the-union address, and who joined the board. Andreessen Horowitz and Index Ventures came in as new investors. Visa and Mastercard, Plaid later disclosed, had both invested, an arrangement that looks in hindsight like two lions taking a stake in the gazelle. In January 2019 Plaid paid $200 million for its competitor Quovo. That spring it launched in the United Kingdom.
By June 2019 Plaid employed about 300 people, with offices in Salt Lake City and New York. It said it had integrated with 15,000 banks across the United States and Canada and that a quarter of people with bank accounts in those countries had connected through at least one Plaid-powered app, up from thirteen percent a year earlier. It was the kind of growth curve founders describe in their sleep.
That month, William Hockey, co-founder, president, and chief technology officer, announced that he was leaving.
TechCrunch broke the news on June 18 and did not hide its puzzlement. Its reporter had been on a panel with Hockey and Brex CEO Henrique Dubugras a month earlier, and Hockey had given no sign of what was coming. Leaving just as the company ramped up international expansion was, the outlet wrote, "strange timing for a founder to depart." The usual explanations did not apply. This was not the Lime model, where a co-founder makes room for a seasoned executive, and it was certainly not the SoFi model, where a chief executive departs amid scandal. Hockey's reports would go to Perret, to other executives, and to Eric Sager, a chief operating officer who had joined earlier that year. Sager is still Plaid's COO, still speaking for it in press releases about government benefits fraud seven years later. Hockey would stay on the board.
His blog post, shared with TechCrunch, was calm to the point of seeming defensive. "This conclusion was neither a rash nor a recent decision," he wrote. "Over the past couple of years, I have known that there would come a point at which I would choose to move to a purely strategic and advisorial role." He then turned the announcement into an argument. "In tech, it has historically been taboo to talk about founders or executives transitioning to different roles inside companies," he wrote. "Leadership transitions need to become a bedrock of any company that desires to endure across decades."
Notice the word endure. It will come up again. A man who would spend the next decade building a company designed to compound slowly was, at twenty-nine, already thinking in decades and already willing to remove himself from the frame if the frame would hold better without him.
The letter also had a passage of open tenderness, the kind founders rarely put in writing.
— William Hockey, open letter on leaving Plaid's operating roles, June 2019Plaid has been the central guiding force for what has now been the entirety of my adult life. I was immensely lucky to start Plaid with one of my best friends Zach, write its first lines of code my senior year of college, learn how to maintain conviction and keep pushing despite seemingly infinite challenges, and even meet my wife through friends at a Plaid party!
The exclamation point is the only one in the letter, and it falls on the wife. Her name is Annie Hockey. She is not a footnote to this story. When Hockey went on to build his second company, he built it with her, and she now helps run it alongside him. The company he left had, in its own way, given him his partner for the next one.
What he left to do, he did not say. TechCrunch did not know. For nearly three years almost no one did.
An Insurance Policy
On January 13, 2020, seven months after Hockey stepped down from operations, Visa announced it would acquire Plaid for $5.3 billion. Hockey was thirty, which, Wikipedia records with characteristic deadpan, made him the youngest person ever to potentially sell a company for more than $5 billion. Potentially turned out to be the key word.
Alfred Kelly, Visa's chairman and chief executive, tweeted that the company was "extremely excited about our acquisition of Plaid and how it enhances the growth trajectory of our business." Perret published a letter on Plaid's website that went back to the beginning: "What started with two founders building in a cramped conference room has become an incredible network that enables millions of consumers to interact with over 2,500 digital finance products." Emory posted congratulations on LinkedIn. Someone named Jared Thoma commented, "That's awesome Will! Congrats bud!" Someone else wondered what the founders would do next.
The Department of Justice also wondered what Visa would do next. On November 5, 2020, its Antitrust Division sued to block the deal, arguing that Visa held a monopoly in online debit and that buying Plaid would remove a nascent competitor before it could grow into a real threat. The complaint's most damaging evidence came from Visa's own executives. Kelly, the government said, had described the acquisition as an "insurance policy" to neutralize a "threat to our important US debit business." The New York Times editorial board weighed in with a headline that read like an indictment: "Visa Is Doing What Big American Companies Do to 'Protect This Business.'"
The phrase did enormous damage because it was so honest. For years Hockey and Perret had described Plaid as plumbing, unglamorous and supportive, a set of pipes for other people's ambitions. Here was the world's largest payments network saying privately that the pipes might one day carry water around Visa's dam, and that the safest course was to buy the pipes.
On January 12, 2021, Visa and Plaid called off the merger. Neither company wanted to spend years in litigation. Eighty-five days later, on April 7, Plaid raised a $425 million Series D led by Altimeter Capital at a valuation of $13.4 billion, more than two and a half times what Visa had offered. Forbes ran the story under a headline proclaiming the founders "Fintech's Newest Billionaires." That October, Plaid launched a pay-by-bank payments product, with third-party processors including Square, Stripe, and SoFi moving the money: a direct transfer from a user's bank account, offered at checkout next to the cards. It was very close to what the government had said Visa wanted to prevent.
Hockey has since discussed on podcasts what the blocked deal taught him about government intervention and motivation. Whatever conclusions he reached privately, the public outlines are instructive enough. A founder had prepared his succession, let go of the controls, and seen the exit arrive on schedule. Then the federal government voided the exit. The company was worth more as a result, at least for a while.
The while did not last. In April 2025 Plaid raised $575 million from Franklin Templeton, Fidelity, BlackRock, NEA, and Ribbit Capital at a $6.1 billion valuation, less than half its peak. In February 2026 an employee share sale valued it at $8 billion, thirty-one percent higher than a year earlier and forty percent below the high. In May 2026 Plaid and OpenAI announced that ChatGPT users could link their bank accounts to the chatbot through Plaid. The pipes still mattered. Their price had simply come back down to earth. Hockey, a board member with his name on all of it, watched the valuation chart from the outside, and by then he had already drawn his own conclusions about where the value in financial infrastructure really sat. He had bought a bank.
Buying the Bank
The idea, as Hockey describes it, came from six years spent at the edge of something. Plaid had worked with financial institutions and fintechs at enormous scale, and the closer Hockey looked, the more he saw that Plaid, like nearly every company in Silicon Valley's fintech wave, operated outside what he calls the regulatory perimeter. The fintechs were tenants. Underneath their apps were sponsor banks, and underneath those were cores, processors, and correspondent relationships, an entire supply chain of money, each link taking a cut and adding delay. Silicon Valley's instinct was to cover the mess with another layer of abstraction. Hockey's instinct, from the farm, from the welding shop, from writing screen scrapers, was to go down to what he has called the bare metal.
On 20VC he called it the money supply chain problem. The answer was to become the bank.
How he paid for it is the most telling financial decision of his career. According to his March 2026 conversation with Patrick O'Shaughnessy on Invest Like the Best, he pledged more than $1 billion in Plaid shares to secure a $70 million loan at SOFR plus ten percent, a loan-to-value ratio of five percent. The arithmetic is almost a joke. A borrower offering twenty dollars of collateral for every dollar of debt is still charged a rate fit for a distressed credit, because the collateral is illiquid private stock in a company whose acquisition the government had just blocked. Hockey took the deal. It was, as one summary of the conversation put it, a move no venture investor would have approved. That was the point. No venture investor would ever have to approve anything.
The target was Northern California National Bank, a small community bank with a national charter. In 2021 Hockey acquired it. Its old technology and operations were rebuilt from the ground up, and it was renamed Column National Association. The FDIC lists it as an insured national bank. Its primary federal regulator is the Office of the Comptroller of the Currency. The charter was the asset. Everything else could be rewritten.
The bank's existing staff seem to have taken the transformation with the slightly dazed goodwill of people whose small-town institution had been bought by someone famous. In April 2023 a Column employee named Amy Morin posted on LinkedIn about an evening event at which "our CEO and the Founder of Column" had talked about the future of banking. "As one of the cofounders of Plaid," she wrote, "William is a bit of a Fintech Legend." He had explained why he bought "our bank," talked about how important community banks are to the economy, and described "a huge technology gap between big banks and community banks." She mentioned that the team was led by William and someone named Todd Lewis and was working "to combine the best in technology with the relationships of a community bank." It is the only glimpse in the record of Hockey seen by people who were there before him. It reads like a church newsletter, and it is lovely.
Column came out of stealth on April 21, 2022. Hockey's LinkedIn announcement was characteristically modest: "After three years of building, today we're announcing Column! We're a nationally chartered bank built from the ground up for developers and anyone creating new financial products. We've been quietly helping power lots of incredible companies like Brex, Plaid, PointCard™, Oxygen and others!" Three years. He had left Plaid in June 2019. Whatever he told TechCrunch then, he had evidently been sketching a bank since roughly the moment he left the room.
One customer on that list deserves a second look. Plaid. The company he co-founded, whose board he sat on, was a customer of the bank he owned. Hockey had gone one layer below his own first company.
His one-sentence description of what he built is the clearest thing anyone has said about Column: "we are a software company that also owns a bank and what we do is we say okay we have this interesting regulatory moat we have a bank that most other people don't have and we're going to build just incredible software behind it." Note the order of the words. Software company first. The bank is the moat, not the identity. The charter keeps competitors out. The code is the castle.
By early 2026, Konrad reported in Upstarts, Column was doing about $200 million in revenue and powering, deep behind the scenes, a roster that reads like a fintech yearbook: Bilt, Brex, Mercury, Ramp, Wise. Unsolicited offers had pointed to a valuation, maybe, of around $6 billion, roughly what Plaid itself had been valued at in its most recent priced round. Hockey also sits on the board of Scale AI. Among fintech founders he has reached what Konrad called "cult-like 'if you know, you know' status." Which is to say that the plumber has fans, and none of them are the people who drink the water.
Slower Than People Think
Silicon Valley's favorite story about speed is that it kills. Ship before you're ready, raise before you need to, hire ahead of the plan. Hockey has spent the Column years arguing, often to rooms full of venture capitalists, that this story is mostly an artifact of how startups are financed, and that a company can be built much more slowly than people think.
The mechanics are simple to describe and hard to copy. Column funds its growth from its own earnings, and Hockey treats each year's profit as an internal funding round. Because no one is waiting for a markup, the company can take on projects with ten-year paybacks. Because no one is waiting for an exit, it can pass on the consensus trend of the quarter. On 20VC he described hiring with almost aggressive patience, waiting for exceptional candidates rather than filling seats and keeping the team deliberately small. One AI-assisted investor database lists Column at roughly 110 employees in early 2026, against about $200 million in revenue. He hires experienced people. He prefers those who have already been through a startup or two, people who, in his phrase, self-select for the pace.
What he offers those people is the most heretical piece of the model. In a LinkedIn post announcing the program, Hockey wrote: "The incentives, dilution and outcomes in company building have changed dramatically in the 13 years I've worked in Silicon Valley - however, how we give equity to employees hasn't. And it should have." Column, he said, was building an equity structure that is "liquid, non-dilutive, no pref stack and no exercise costs." Each year the company sets aside a quarter of its earnings to buy back employee shares in a tender offer. There are no outside preferences to sit ahead of the common stock because there is no outside capital. There is no exercise price to finance because there is nothing to exercise.
The accompanying essay, quoted by a commenter under the post, did not soften the critique. "Our industry hasn't been entirely intellectually honest about how tech capital structures have evolved," it read. "Most early-stage equity now requires a parlay of optimistic, outdated assumptions to pay off for employees." Under the post, someone named Dan Kahn left a joke: "What do you call a profitable startup that pays their employees fairly? 🦄"
Behind the program sits a moral argument that Hockey made at length with O'Shaughnessy. Early employees, he contends, often take more real financial risk than founders do. A twenty-four-year-old who leaves a $400,000-to-$500,000 package at Google for $90,000 at a startup is accepting a sacrifice of four or five years with no safety net. The founder, whatever happens, keeps the title of former CEO, can raise money again, and can often sell secondary shares regardless of how the company performs. Hockey's conclusion, applied to people like himself, is severe: second-time founders who already have liquidity should put all of their personal capital into the new venture. He did. The $1 billion in Plaid stock pledged against a $70 million loan was not a metaphor for skin in the game. It was the skin.
The irony, which Hockey surely appreciates, is that the loudest skeptic of venture capital in fintech is also one of its busiest angels. Under the banner of Moving Capital, data aggregators count roughly fifty investments, checks of $5,000 to $50,000, in companies with names like Slash, Town, Maza, Vooma, Portal, and Double. In May 2023 he invested in a startup called Boom alongside Zach Perret, who now writes checks through a fund called Mischief. The two former roommates in the cramped conference room are now co-investors in other people's cramped conference rooms. Hockey does not seem to think venture capital is bad. He thinks it is a tool built for a particular kind of company, and that founders have stopped asking whether they are that kind of company.
Two Thousand Pages for One Idea
O'Shaughnessy opened the Invest Like the Best episode with an anecdote that seemed to delight him. Hockey, he said, had tracked down an obscure book about an ancient bank and found, buried somewhere in its two thousand pages, a single idea that became a simple feature in Column's product. The host offered it as an example of a founder "winning because he is willing to do everything." The precise bank and the precise idea do not appear in the excerpts I have. That seems fitting. The anecdote is about the reading, not the result.
Hockey describes this as deep specialization in boring niches. Read the history of nineteenth-century banking systems; read thousands of pages that no generalist competitor will ever read; accept that most of it will yield nothing. The payoff is not knowledge in general but a few specific facts that nobody else has, which is how a team of young Americans in suits ends up reciting details about Ethiopia's financial system to Ethiopian bankers.
The two books he recommended on 20VC are worth looking at closely, because they rhyme with him. Dan Morgan's Merchants of Grain is about the five giant trading houses that quietly controlled the world's food supply, companies with names few consumers knew that moved the commodity everyone ate. Javier Blas and Jack Farchy's The World for Sale is about the commodity traders who broker oil, metals, and grain through the hidden passages of global trade. Both are about intermediaries. Both are about the enormous power and frequent infamy of firms that stand invisibly between producers and the world. A man who builds the plumbing of the dollar reads, for pleasure, about the people who built the plumbing of wheat and crude. Hockey does not explain the choice. He doesn't need to.
The reading suits his temperament. Konrad describes him as a natural introvert who prefers to spend conferences answering his engineers' pull requests. Asked on 20VC to name his strengths and weaknesses, he offered optimism, pain tolerance, and social anxiety, a list that sounds less like a self-assessment than a recipe for a certain kind of founder. Optimism to start. Pain tolerance to stay. Social anxiety to make the solitary work of reading two thousand pages feel like relief rather than duty.
Konrad asked him the simple question every profile eventually asks: what drives him to start companies? The answer took four words.
— William Hockey, to Alex Konrad, Upstarts Media, February 2026I'm petrified of being bored.
It is a strange confession from a billionaire who bought a community bank. Banking is perhaps the most boring industry Americans can name. Hockey himself has called it "very archaic," "not that sexy, it's super old." But boredom, for him, does not seem to be a property of a subject. It is a property of shallowness. The bank is boring from the street. From inside the ledger, from inside the two thousand pages, from inside the sealed box he spent his twenties trying to pry open, it is anything but.
The Dollar's Back Office
The African tours that ended badly in Addis Ababa did not begin as a sales strategy. They began as a way of seeing.
Hockey goes to places like Kinshasa on purpose, he told O'Shaughnessy, because constrained environments produce different products and different talent. In the Democratic Republic of the Congo, by his figures, mobile penetration is under twenty-five percent and banking penetration under five, the kind of gap that allowed China to skip from cash straight to mobile payments without ever really passing through the checkbook. He has noticed that mid-level banking executives in such markets often outperform their Western counterparts, because the best local talent flows into the established banks rather than into startups. The best person in the room in Kinshasa may work at a bank that a San Francisco founder has never heard of. He goes to find that person.
He also goes to escape something. In the conversation he repeated an observation from someone else that he said was "quite accurate actually": that "the two most consensus societies he's ever been to is San Francisco and Beijing." It is a pointed line from a man who has lived in San Francisco since the fall of 2013. The suspicion of consensus runs through everything: buying a bank when software was supposed to eat banks, refusing venture capital in the era of the mega-round, forcing a logo onto an industry that prized invisibility. Kinshasa is not a market so much as a corrective lens.
The larger argument, the one that has given his recent interviews an unexpected geopolitical edge, is about the dollar. Roughly seventy-five percent of global trade is still denominated in U.S. dollars, he points out, including trade between countries hostile to the United States. Russian gas sold to China is still largely priced in dollars. In his view, financial infrastructure, not military hardware, is the main instrument through which American interests are enforced abroad, and the infrastructure carrying the dollar is therefore a matter of national security. Konrad framed Hockey's "bigger fight" as a new financial Cold War with China. The plumber, it turns out, thinks the plumbing is load-bearing for the republic.
He is less romantic about the system at home than that might suggest. On 20VC he argued that "the US financial system can function like a protocol," and that its problems lie in the implementation layered on top, not in the underlying rails. With O'Shaughnessy he went further: the Federal Reserve's existing infrastructure can already move money instantly, around the clock. The friction consumers experience, the holds, the delays, the forms, is not a technical limitation. It is deliberate, engineered to protect the five to ten percent of users vulnerable to fraud and elder abuse. The pipes are fine. The valves are set conservatively on purpose. Anyone who wants to open them had better understand who might drown.
That is the knowledge he carries into the rooms with the suits and the ambassadors and the formal dinners. It is also why the rooms matter so much. To walk into a bank in a foreign capital and ask to connect it to the dollar system is to ask to be trusted with the most heavily policed flow of value on earth. The trust has to be earned in person, by a man who prefers not to be in person, under a name that turns out to be a common noun.
The bank in Addis Ababa still does not work with Column. Back home the company keeps compounding, and its employees collect their annual tender. Plaid's screen appears in apps a few hundred million times, and almost no one who taps through it thinks about the farm kid who decided, at the age when most people are deciding on a graduate program, that the pipe should carry a name.
And at the parties, the ones a successful man in San Francisco cannot always avoid, if there are more than four strangers in the house, you will not find William Hockey in the living room. Konrad says you will find him in the bathroom, door shut, coding on his Android phone.
Part IIThe Playbook
What follows are the operating principles that can be drawn from Hockey's two companies: one built the Silicon Valley way and then partly abandoned by its builder, the other built against nearly every Silicon Valley assumption. They are not maxims he has published. They are patterns in his decisions, read against the evidence. Several of them only make sense together. Owning the bank only works if you are willing to pay for it yourself, and paying for it yourself only works if you are willing to build slowly.
The principles
- 1.Descend a layer when the layer above won't hold.
- 2.Infrastructure should sign its work.
- 3.Prove it before you pitch it.
- 4.Design your own exit from the operating chair.
- 5.Own the license; don't rent it.
- 6.Finance the unfundable bet with your own balance sheet.
- 7.Treat each year's profit as the funding round.
- 8.Pay the people who take the real risk.
- 9.Read the two thousand pages.
- 10.Go where the constraints are.
- 11.Fix the implementation, not the protocol.
Principle 1
Descend a layer when the layer above won't hold
Hockey's career is a sequence of descents. He and Perret started at the consumer layer, with budgeting and bookkeeping apps, and kept hitting the inability to connect to banks. Rather than build a better app on a broken foundation, they went down a level and built the foundation. Ten years later he did it again. Having seen from Plaid that the fintech layer depended on a fragmented, rent-extracting chain of sponsor banks and processors, he went down to the charter itself. Plaid became one of Column's early customers, which means Hockey built beneath his own previous company.
The pattern is non-obvious because the instinct in software is to go up, toward users, brands, and higher margins, and to abstract away whatever lies beneath. Hockey's view, stated on 20VC, is that Silicon Valley too often prefers abstraction to solving the underlying problem. Abstractions inherit every defect in the layer they sit on. The founder who controls the lower layer controls the failure modes for everyone above.
Tactic: When the same obstacle blocks your product three times, stop working around it and seriously price what it would cost to own the layer that produces it.
Principle 2
Infrastructure should sign its work
The conventional wisdom says infrastructure should be invisible. Hockey bet Plaid against that view, requiring every customer to adopt a Plaid-designed, Plaid-branded connection screen over an eighteen-to-twenty-four-month rollout, against heavy pushback. He estimates the decision accounts for ninety percent of the company's value. The logic was defensive as well as generous. A uniform screen converted better and gave users familiarity and security. It also made Plaid impossible to swap out, because the consumer had come to know Plaid instead of the app's homegrown flow.
The cost of visibility is accountability. The same screen drew the TD Bank lawsuit and became central to the $58 million settlement. A signed piece of infrastructure gets credit and blame alike. The lesson is not that branding is free. It is that being anonymous is worse, because an anonymous supplier is a commodity, and a commodity's customers renegotiate.
Tactic: Find the single surface where your infrastructure touches the end user and insist on owning its design, even if it costs you a year of customer friction.
Principle 3
Prove it before you pitch it
Seventy investors passed on the idea of a unified bank API. The founders did not respond by refining the deck. They entered the TechCrunch Disrupt hackathon in April 2013, built Rambler, a working consumer app running on their pipes, and won the grand prize. Five months later they closed a $2.8 million seed round. Their first major customer, Venmo, came through a friend with an urgent operational problem, not through a sales process.
Infrastructure is notoriously hard to pitch because its value shows up in other people's products. The cure is to build one of those products yourself, quickly, as a demonstration. Rambler was disposable, and its repository still sits on Hockey's GitHub with five stars. What it demonstrated was not.
Tactic: If you are selling a building block, build one visible thing on top of it in a weekend and let that do the pitching.
Principle 4
Design your own exit from the operating chair
Hockey left his operating roles at Plaid in June 2019, at a $2.65 billion valuation, with the company growing fast and its UK launch only a month old, and with no scandal and no board pressure involved. TechCrunch called the timing strange. His own explanation was that the decision had been coming for years and that "leadership transitions need to become a bedrock of any company that desires to endure across decades."
Two things make the move instructive. First, he stayed on the board, keeping strategic involvement while giving up operational control, a role he deliberately designed rather than had forced on him. Second, the timing let him start something new from a position of strength. Column's three years of quiet building began almost immediately. On 20VC he talked about detaching his identity from his company's success as a way of building a secure base for taking bigger risks. Leaving well is a precondition for starting again well.
Tactic: Write down, now, the conditions under which you would move to an advisory role, and share them with your co-founder before anyone else has to bring it up.
Principle 5
Own the license; don't rent it
Most fintechs rent their regulated status from sponsor banks. Hockey bought his outright: Northern California National Bank, a national charter regulated by the OCC, rebuilt and renamed Column National Association. He describes the result plainly: "a software company that also owns a bank." The charter is a regulatory moat that "most other people don't have," and the software behind it is the differentiator.
The point is about where durable advantage sits in regulated industries. Software can be copied. Charters, licenses, and the operational record that regulators reward cannot be copied quickly. The founder who treats regulation as terrain to occupy rather than a cost to minimize ends up with something competitors cannot simply raise money to replicate.
Tactic: In a regulated market, figure out which license your competitors are renting and model what it would take to own one outright.
⚖
Two Companies, Two Theories of Building
Hockey's second company inverts nearly every structural choice of his first.
| Dimension | Plaid (2012–2019) | Column (2019–present) |
|---|---|---|
| Position relative to regulation | Outside Software layer on top of banks | Inside OCC-chartered national bank |
| Capital source | ~$310M in venture capital by mid-2019 | $70M loan secured by $1B+ of founder's Plaid shares |
| Growth funding | Successive priced rounds, seed through Series C | Annual earnings treated as the funding round |
| Visibility to consumers | Branded Link screen in partner apps | Deeper still, beneath Brex, Mercury, Ramp, Wise |
| Employee liquidity | Paper equity, preference stack, later secondary sales | 25% of earnings to annual buyback; no pref stack, no exercise cost |
| Key risk | Risk Commoditization; privacy litigation | Risk Founder concentration; name confusion abroad |
Principle 6
Finance the unfundable bet with your own balance sheet
Buying a regulated bank is the kind of capital-intensive, slow-payback, compliance-heavy move that venture investors are structurally built to avoid. Hockey avoided them instead. He pledged more than $1 billion of Plaid stock for a $70 million loan at SOFR plus ten percent, a five percent loan-to-value ratio at an interest rate fit for junk credit. On paper the terms were terrible. In practice they bought total freedom from the next fundraise.
His broader argument is that second-time founders with liquidity should put all of their personal capital into the new company. This is not bravado. It follows from his observation that founders are usually insulated from downside in ways their employees are not. A founder whose net worth is pledged against the company has joined the employees' side of the risk ledger.
Tactic: If your thesis is too slow or too strange for venture capital, see whether your existing assets can collateralize the first step, and accept bad loan terms in exchange for owning your timeline.
Principle 7
Treat each year's profit as the funding round
Column funds its expansion from its own earnings. Hockey describes each year's profit as an internal round, and because the company is not optimizing for a markup, it can invest in projects with ten-year paybacks. The pace that results is deliberate: patient hiring, a preference for experienced people who self-select for the tempo, and a headcount reportedly around 110 against roughly $200 million in revenue.
The non-obvious claim is that urgency and speed are separate things. Hockey argues that companies can be built much more slowly than the venture narrative suggests, and that the decisions requiring real speed are fewer than founders think. A company without fundraising deadlines can sort its decisions by how reversible they are rather than by the investor calendar.
Tactic: List your company's next twelve months of decisions and mark which ones would actually be harmed by waiting ninety days. Move fast only on those.
Principle 8
Pay the people who take the real risk
Hockey's critique of startup equity is specific. A twenty-four-year-old who leaves $400,000 to $500,000 at Google for $90,000 accepts years of foregone income, and in a world of long IPO timelines, heavy dilution, and stacked preferences, the paper equity meant to compensate for that often loses much of its value. Column's answer is an annual tender funded by twenty-five percent of earnings, with "no pref stack and no exercise costs."
The program also recruits. It appeals to second-time employees, people who have been through several funding rounds and never seen a payout, and who therefore value certainty over lottery tickets. A compensation structure is a filter. Column's selects for people who think like owners because they actually get paid like owners, every year.
Tactic: Calculate what your earliest employees gave up in cash compensation to join, and design at least one mechanism, whether buyback, secondary, or profit share, that returns some of it before an exit.
Principle 9
Read the two thousand pages
Patrick O'Shaughnessy's favorite story about Hockey is the obscure history of an ancient bank, two thousand pages long, from which Hockey drew a single product idea. The habit runs deep: histories of nineteenth-century banking, Merchants of Grain, The World for Sale. In practice it means a team that can recite details of a foreign financial system to that system's own bankers.
Deep specialization in unglamorous material is an advantage generalists cannot buy, because the cost of it is time spent mostly unproductively. Most of the two thousand pages yield nothing. The founder willing to absorb that waste gets the few ideas nobody else in the market has. Hockey's own explanation of his drive, that he is "petrified of being bored," suggests that for him the depth is the cure for tedium, not its cause.
Tactic: Pick the oldest, most boring history of your industry you can find and read it all, writing down the one idea that does not exist in any modern competitor.
Principle 10
Go where the constraints are
Hockey travels to Kinshasa and other emerging-market capitals deliberately, because constrained environments produce different products and different talent. In the DRC, under twenty-five percent mobile penetration and under five percent banking penetration create the possibility of leapfrogging, the way China went from cash to mobile payments. He has found that mid-level bank executives in those markets often outperform their Western peers, because local talent flows into the incumbents.
The deeper motive is escaping consensus. Hockey has quoted the observation that San Francisco and Beijing are the two most consensus-driven societies, approvingly. A founder surrounded only by peers drifts toward their opinions. Constraints elsewhere break the pattern. The trips carry real risk, as the walkout in Addis Ababa shows, but they are a research method, not a vacation.
Tactic: Once a year, spend a week in a market where your product category barely exists and write down which of your assumptions failed to survive the trip.
Principle 11
Fix the implementation, not the protocol
Hockey's most counterintuitive view about American finance is that it mostly works. The U.S. system, he has argued, "can function like a protocol." The Federal Reserve's infrastructure already supports instant, round-the-clock money movement, and much of the friction consumers hate is engineered on purpose to protect the five to ten percent of users vulnerable to fraud and elder abuse. The problems lie in the implementation built on top, not in the rails.
This puts him at odds with both the crypto camp, which wants to replace the protocol, and the fintech camp, which wants to wrap it. His position is to go inside, gain the license to touch the rails directly, and rebuild the implementation with modern software, while respecting the reasons the valves were set conservatively. Seen from Column's vantage, the dollar's global reach, three-quarters of world trade, is an asset to modernize, not a relic to disrupt.
Tactic: Before declaring that your industry's infrastructure is broken, find out who its friction is designed to protect, and design a replacement that protects them better.
⏱
The Hockey Timeline
Two companies, one long descent toward the bare metal.
2012
Graduates from Emory with a BBA; writes Plaid's first lines of code as a senior.
2013
Rambler wins the TechCrunch Disrupt NY hackathon (April 28); Plaid raises a $2.8M seed (September); moves to San Francisco.
2018
Plaid raises a $250M Series C at $2.65B, led by Mary Meeker.
2019
Plaid acquires Quovo for $200M (January); Hockey leaves his operating roles and stays on the board (June).
2020
Visa agrees to buy Plaid for $5.3B (January 13); DOJ sues to block (November 5).
2021
Merger abandoned (January 12); Plaid valued at $13.4B (April 7); Hockey acquires Northern California National Bank.
2022
Column emerges from stealth on April 21 after "three years of building."
2025
Plaid raises at $6.1B; the Addis Ababa walkout.
2026
Upstarts reports ~$200M in revenue for Column; Plaid's employee sale values it at $8B.
Free playbook
Get The Business Model Playbook
58 business models, one visual page each: how the money flows, the metrics that matter, and who runs it. Free when you join the Faster Than Normal email.
Free. No spam. Unsubscribe anytime.
Part IIIQuotes and Maxims
In their words
— William Hockey, 20VC with Harry Stebbings, December 2022So I grew up on a farm out here in central California, and I grew up building everything.
— William Hockey, Emory Entrepreneurship Summit, April 2019You should never start a business to make money. That never entered our consciousness. We were like, we're going to build something that's going to fail and then we're going down.
— William Hockey, blog post on leaving Plaid's operating roles, June 2019In tech, it has historically been taboo to talk about founders or executives transitioning to different roles inside companies. Leadership transitions need to become a bedrock of any company that desires to endure across decades.
— William Hockey, Invest Like the Best, March 2026We are a software company that also owns a bank and what we do is we say okay we have this interesting regulatory moat we have a bank that most other people don't have and we're going to build just incredible software behind it.
— William Hockey, LinkedInThe incentives, dilution and outcomes in company building have changed dramatically in the 13 years I've worked in Silicon Valley - however, how we give equity to employees hasn't. And it should have.
Maxims
- Go down a layer. When the foundation keeps breaking your product, stop building on top and own the foundation.
- Sign the pipe. Anonymous infrastructure becomes a commodity; put your name on the one screen the user sees.
- Demonstrate before you persuade. Seventy rejections lost to one weekend's working hackathon app.
- Leave on your own schedule. Plan your move out of the operating chair before anyone else has to.
- Own the charter. In regulated markets, the license is the moat and the software is the castle.
- Collateralize conviction. If no investor will fund the bet, pledge your own balance sheet and accept the bad terms.
- Profit is a funding round. A company that funds itself can afford ten-year paybacks and patient hiring.
- Pay the real risk-takers. Early employees give up more than founders do; return some of it every year, not just at exit.
- Read until it's boring, then keep reading. One idea from two thousand pages is a competitive advantage no one else can buy.
- Distrust consensus. Leave the room where everyone agrees, whether that means a party, a city, or a continent.
