on william hockey
The founding condition for Column is the Visa acquisition that didn't happen. In 2019, Visa announced it would acquire Plaid for $5.3 billion. Hockey and Zach Perret had built one of the more consequential fintech infrastructure companies in the last decade — the layer that gave apps access to bank account data and became embedded in most of the financial products people actually use. The DOJ blocked the deal in early 2021. No exit. No liquidity. Hockey left Plaid with paper wealth and essentially no cash, which is a particular kind of position to be in when you want to start a company that requires buying a regulated bank.
What he did next is worth understanding precisely because it contradicts the story people tell about second-time founders. The narrative assumes they're flush. Hockey pledged over a billion dollars of Plaid stock to borrow $70 million at SOFR plus 10 percent, 5% LTV. He bought Northern California National Bank for $70 million with borrowed money. Over the next three years he was margin called three times and came close to bankruptcy more than once. The bank required two to three years of non-revenue focus before they could take on clients. There was no path where a venture firm would have funded this. The bet worked because he couldn't afford for it not to.
The architectural thesis underneath all of that risk is specific. Most fintech is middleware on top of a bank. You build the product, the brand, the UX — and at some point you hit the charter boundary. The sponsor bank's risk tolerance becomes your risk tolerance. Their API surface becomes your ceiling. Their compliance posture shapes your roadmap in ways that have nothing to do with your product decisions. Hockey's argument is that middleware companies in financial services fail at that boundary predictably, and the only durable answer is to own the charter outright. Column is the bank. Not a fintech that works with a bank — the bank. The primitive that the middleware layer sits on.
The product surface that follows from that decision reads differently once you understand the founding story. Column offers ACH, wire, RTP, FedNow, check issuing, card issuing, and lending infrastructure — all direct, no intermediary in the path. The documentation is specific. The error messages are specific. The API is designed by people who've absorbed the underlying system rather than wrapped it. The signal is that the team built for the developer who's trying to do something real, not for the demo.
The model for how to run the company is as considered as the architecture. Column is 100% employee and founder owned. No external capital, no preference stack, no dilution. Growth funded by earnings. Every year, 25% of earnings go back to employees through a tender offer — actual liquidity, not a paper number on a vesting schedule. There's a $2,000 monthly housing stipend for employees who live within two miles of the office. The target hire is the second-company employee: someone who's been through six rounds and four pivots and has done the dilution math on their own outcome. The pitch works on them in a way it doesn't work on a new grad, because they've seen how the standard model resolves.
The VC dependency argument he makes is structural, not moral. His framing: venture money is like heroin — it feels good and the problem is you have to keep taking it. Very few companies raise a $100 million Series A and stop. The fundraising cycle shapes the strategy cycle. You optimize for what the next round needs to look like. You build a stablecoin strategy when stablecoins are interesting. You build an AI strategy when AI is interesting. The goal line moves every eighteen months because it has to. His version: if Column grows at 80% instead of 110%, it doesn't matter, because the company is profitable and building what it wants to build. The constraints from not having external capital are real. The freedom from not having external capital is also real.
The specialization argument is the part that generalizes most cleanly. Hockey studies banks in 1800s Japan. He reads 2,000-page books on 19th century Chinese banking history. The yield from that kind of reading is usually one small insight per book — something that only creates leverage because it sits on top of decades of domain knowledge. His claim is that this is where value actually hides: in the niche that's boring enough that nobody without a specific reason to care will go there. AI, geopolitics, the broad intellectual topics that make people feel smart — those spaces are crowded with competent people. The boring niche that requires a multi-decade commitment to find interesting is not. He describes himself as probably the best in the world at a small number of things that are very hard to explain on a podcast. That specificity is the strategy.
What I keep returning to is the relationship between the personal financial risk and the company's structure. The compression of outcomes — everything dependent on Column, everything dependent on Plaid, margins called, near bankruptcy, no diversification — is what made the long-term bets possible. Buying a bank during the Biden administration when it was non-consensus. Not raising money when raising money would have been easier. Building an employee model that doesn't optimize for fundraising cycles. Those decisions required a founder who couldn't afford to be wrong about the thesis, which is a different kind of conviction than the kind you can maintain while holding a diversified portfolio. He owns two things: Column and Plaid. That's the position. He says it's motivating. I believe him.