Stripe's first employee built the rails. Then he built his own. Now he's the rare fintech founder who buys community banks instead of burying them.
A friend of Darragh Buckley once sent a real-time payment to a car dealer. On a weekend. The dealer was so startled he assumed it couldn't be real - money does not move on Saturdays, everybody knows that. Except it does, it had, and the only thing missing was someone willing to wire it together. That small jolt of disbelief is the whole thesis of Buckley's career.
He runs Increase, a company that lets technology businesses store, move, and reconcile money through clean APIs - by plugging directly into the Federal Reserve's payment networks instead of stacking middlemen. Customers include Ramp, Check, and Pipe. The team is famously small. The volume is famously not: well over $100 billion a year.
The strange part is what he does with his spare conviction. While the fintech consensus declared community banks a dying species, Buckley went and bought stakes in them - including a position in Twin City Bank large enough that the Federal Reserve had to announce it. Most founders pick a side in the old-bank-versus-new-tech fight. He decided the fight was the wrong frame, and that the more interesting move was to own a piece of the thing everyone else was writing off.
He is not loud about any of it. There is no manifesto, no tour of hot takes. There is a small company moving an enormous amount of money, a handful of community banks getting a quiet patron, and an engineer who keeps choosing the layer underneath the one everyone argues about.
The lesson Buckley took from Stripe was almost geological. When you're stuck solving a business problem, a technical problem, and a risk problem all at the same instant, you don't push harder - you drop down a layer. Increase is that instinct turned into a company. Instead of renting access to someone else's banking core, it operates its own and connects straight into FedACH, Fedwire, FedNow, and real-time payments.
The pitch is not glamour. It's the opposite of glamour. He wants payroll to be boring and reliable, the kind of system where the engineers responsible can go to sleep peacefully. He wants compliance to stop being a prayer and start being a test you can run - the Bank Secrecy Act and the travel rule encoded into the software, checked like any other line of code.
Increase runs its own banking core and connects directly to Federal Reserve networks, rather than chaining together intermediaries.
Fintechs like Ramp, Check, and Pipe move money programmatically through a full-stack API instead of stitching together legacy systems.
Engineers should understand the regulations and bake them into tests - certainty by construction, not hope after the fact.
In 2025, Buckley acquired a stake exceeding 10% in Twin City Bank, a small community bank in Longview, Washington, about an hour north of Portland. The size of the buy tripped a Federal Reserve disclosure - the regulatory equivalent of a flare going up. It was his third investment in a Washington community bank.
The fintech world reflexively assumed the worst: surely he wanted the bank as a captive partner for Increase. Buckley said no, plainly. Twin City is, and will remain, a community-focused bank. He's been explicit that it should not do sponsor banking - that requires very specific capability and capacity to supervise partners safely. A competitor reportedly hired an agency to plant negative stories about the deal. He kept his answer simple.
His actual heresy is optimism. Where others see community banks as roadkill, he sees an underrated asset: relationships and knowledge that scale poorly for giants and not at all for spreadsheets. Give them technology, he argues, and they can compete on rates, service, and software - the things that matter.
In an industry that often treats regulators as an obstacle to route around, Buckley says the opposite out loud. He credits federal regulators with doing a fantastic job in recent years adding clarity around third-party risk management - the exact corner of the rulebook that has tripped up so many banking-as-a-service arrangements. He is not just complimenting them, though. He has ideas: ways the compliance function might be centralized rather than rebuilt badly at every shop, and possible mechanisms to actually pay for those changes.
It fits the pattern. Buckley is drawn to the unglamorous machinery other people would rather not think about - settlement, reconciliation, supervision, the travel rule - because that machinery is where reliability is won or lost. He is explicit that sponsor banking is not for everyone: it demands very specific capability and capacity to supervise partners safely and soundly, and pretending otherwise is how risk piles up quietly until it doesn't.
Buckley grew up in Limerick, Ireland - the same small city that produced Patrick and John Collison. He studied engineering, picking up degrees from MIT and the University of Michigan, and built a reputation as a serious engineer among his peers. When the Collisons started the company that became Stripe, they reconnected with the hometown engineer, and in 2010 he became employee number one.
At Stripe he was the banking hire. He negotiated, implemented, and scaled the company's agreement with its primary bank partner, Wells Fargo, and helped wire the young company into the financial networks it would eventually reshape. By the time he left, he had earned something rarer than a title: a genuine reputation as an engineer among engineers, the person other builders trust on the hard parts.
He left in 2016, started a family, and traveled - the kind of pause that usually ends a story rather than reloading it. Then, in 2020, he went back to the same problem one layer deeper, founding Increase through Y Combinator's Summer 2020 batch. The throughline is hard to miss: every chapter is the same man finding the floor beneath a floor and deciding to rebuild it properly.
Becomes Stripe's first employee; builds banking infrastructure and lands the Wells Fargo partnership.
Leaves Stripe. Starts a family, takes time to travel.
Founds Increase. Goes through Y Combinator's Summer 2020 batch.
Increase is processing over $100 billion a year with a tiny team.
Buys 10%+ of Twin City Bank, triggering Federal Reserve disclosure - his third Washington community-bank stake.
Buckley's quiet conviction is that better banking comes from specialization. Understanding what a dentist actually needs from a bank is a different problem from serving everyone inside a geographic radius - and the generic, one-size-fits-all approach leaves most of the value on the table. The infrastructure underneath should be neutral and programmable enough that whoever knows the customer best can build exactly the right thing on top.
That is also why he is careful about where the magic stops. He is candid that artificial intelligence can now interpret the semantics of a request - it can read what a human means. But the real bottleneck is downstream: turning that instruction into the precise, correctly-formatted messages the financial networks demand. The intelligence problem is mostly solved. The plumbing problem is the hard part, and the plumbing problem is the business.
Serve a trade you understand deeply rather than a radius you happen to sit inside.
Reading intent is easy now. Converting intent into clean financial-network messages is not.
A tiny team moving eleven figures a year - capital efficiency as a design choice, not an accident.
"Twin City Bank is, and will remain, a community-focused bank."
"Sponsor banking requires very specific capability and capacity to supervise partners safely and soundly."
"We're missing just that last little bit of payment delivery automation."
"The federal regulators have done a fantastic job in the last two years of adding clarity around third-party risk management."
Buckley walks through how Increase connects directly to the Federal Reserve, why specialization beats one-size-fits-all banking, and where AI actually hits a wall in moving money.