Stripe Pioneer Declares War on Payments' Messy Last Mile

A payments engineer who helped build Stripe is betting the next leap forward requires owning a bank, not destroying one.

That's the thesis of Darragh Buckley, Stripe's first employee and now CEO of Increase, which recently launched its own federally chartered institution, Increase Bank. According to an interview with PYMNTS, his experience revealed a paradox: fintech has brilliantly smoothed the front-end of transactions but left the critical "last mile" settlement process as an opaque, unreliable black box. Solving it, he argues, demands an engineering mindset applied to the regulated core of banking itself, bridging a gap neither incumbents nor disruptors have yet closed.

Stripe’s First Architect Declares War on Payments’ Hidden Glitch

For modern software companies, sending a payment is often the easiest part. The real chaos begins in the silent, unpredictable moments that follow. A wire transfer is initiated, an ACH file is submitted, but what happens next is frequently a mystery. Did the receiving bank accept it? Is it stuck in a queue? Has it been flagged? The answer often arrives only when a user support ticket lands, complaining about a missing payment.

Buckley calls this the "messy last mile" of payments, and his move from Stripe to founding Increase was a direct response to the operational friction he witnessed at scale. The fintech industry successfully abstracted away the complexity of initiating a transaction through clean APIs. But it has largely failed to solve for visibility and reliability in the final steps of settlement, the domain of monolithic banking cores and legacy networks. Instead of replacing these rails, Increase's strategy is to make them legible, reliable, and programmable for software-first companies. As Buckley summarized, "Increase’s role is improving that last mile delivery as opposed to burning the house down and trying to build something new."

This isn't just about speed. It's about turning a bank's operations into something software can see, test, and trust. The goal, as we see in our analysis on AI agents shifting loyalty and liability in payments, is a system where autonomous systems can rely on financial infrastructure with zero ambiguity.

How an Engineer and a Bank Examiner Finally Started Speaking the Same Language

Fintech's origin story often pits nimble innovators against slow, obstructive legacy banks. Buckley rejects that narrative. The real conflict, he suggests, is between two different types of precision that are finally aligning.

"The engineer wants to know what the system is doing at every step. The examiner wants to know it’ll do the right thing every time. Those aren’t competing demands."

This conversation represents a seismic shift. Historically, a bank examiner relied on audits, paper trails, and black-box trust in a system's outputs. An engineer needed granular, real-time telemetry to debug and build. Increase's approach is to satisfy both by designing a system where compliance and transparency are engineered from the first line of code. Continuous test transactions monitor connections to the Federal Reserve, Visa, and other networks, proactively checking for issues instead of waiting for customer complaints.

For functions like payroll, "one of the great unsung heroes of the U.S. economy," per Buckley, this means moving from a silent, faith-based process to a confirmed, step-by-step handshake. Increase's system confirms receipt of a payroll file, its submission to the Fed, and the Fed’s acceptance. The business knows where the money is before employees start asking where it isn't. This fusion of engineering and examination creates a bank that is as auditable by software as it is by regulators.

Payment Failures Aren't Bugs, They're a $50 Billion Design Flaw

While the source material doesn't provide a specific dollar figure, the scale of the problem is implied through operational pain points. Failed wires, unreconciled ACH batches, and silent payment drops aren't random glitches. They are the inevitable result of a design where the initiating software and the settling infrastructure speak different languages and operate on different timelines.

The cost manifests in three areas:

• Manual Intervention: A failed wire arrives as an email requiring human review. This is manageable at dozens of transactions but crippling at thousands.
• Lost Productivity: Engineering and operations teams spend cycles diagnosing issues in systems they cannot see or instrument.
• Business Friction: For platforms managing payouts to creators, drivers, or vendors, unreliable settlement erodes trust and complicates cash flow.

Buckley doesn't blame the partner banks for this. "They were built for a different kind of client." Their systems are reliable for their intended use, handling batches for traditional businesses. They were not built to provide the real-time status updates and exception handling that software companies require to build robust financial products. The status of a payment is often as valuable as the settlement itself, because it allows developers to build predictable user experiences around it.

The Legacy Bank, the Fintech Unicorn, and the Missing Bridge in Between

This reveals a tripartite stalemate in the financial ecosystem.

Increase positions itself as that critical bridge. It doesn't seek to replace the Fedwire or ACH networks Buckley praises as "fantastically reliable." Instead, it builds a new operational layer on top of them, a bank whose ledger and processes are as transparent and programmable as a modern API. This means owning the bank is not just about direct network access; it's about removing the final operational layer between the payment network and the software company that depends on it.

The need for such control is becoming a recurring theme, as seen when US regulators ban a bank from lending to its worst clients. Direct control over risk and compliance is increasingly non-negotiable.

Building the Cathedral Before the First Payment Settles

Increase's multi-year path to launching its own bank underscores its foundational thesis. They didn't start by acquiring a charter. They first built their proprietary technology stack, connected directly to payment networks, and developed the operational toolkit. Only after years of working with partner banks and even serving on a community bank board did the final piece, bank ownership, become the obvious conclusion.

This sequence is crucial. It inverts the traditional model where a bank is acquired first, and then technology is awkwardly grafted onto it. Increase built its "cathedral", the complete, engineered system, before needing to process its first customer transaction through its own charter. The charter became a strategic enabler, not the starting point.

Buckley's analogy for the future isn't a blockchain or a rocket ship. It's public transit: "The tunnels don’t change often. What riders feel are better schedules, easier payments and clearer explanations when something goes wrong." The infrastructure remains, but the experience atop it is transformed. This reflects a broader industry maturity, moving from disruptive rhetoric to pragmatic enhancement, much like the evolving focus of networks in the cross-border payments race.

What a Bank You Can't See Means for Every Business That Moves Money

The implications of this engineered "bridge" bank are profound for any business that moves money.

Furthermore, this redefines "community" banking. Buckley argues that in the digital age, community is defined by shared needs, not geography. "The dentist in Bend [Oregon] looks more similar to the dentist in Boston than they do to the restaurant next door from a banking services perspective," he says. An engineered bank can efficiently serve a "community of dentists" with tailored workflows, a vision that points toward vertical-specific embedded finance becoming the norm.

The Inevitable Rise of Banking-as-Code

Increase’s model is a leading indicator of a broader shift: banking is becoming a software discipline. The core competency is shifting from managing balance sheets and branch networks to managing code, data pipelines, and API contracts.

XOOMAR Analysis: Within five years, we predict the quality of a "banking core" will be evaluated less by its age and more by its commit history, test coverage, and API documentation. Niche, vertical-specific "embedded banks" will proliferate, powered by clean-sheet backends that make launching industry-specific financial products as straightforward as integrating a new SDK.

This forces a parallel evolution in regulation. Auditors may increasingly run automated scripts against a bank's API to verify compliance, rather than sifting through paper reports. The conversation between the engineer and the examiner moves from the conference room to the command line.

For the industry, the watchpoint is straightforward: Does this model prove its resilience? The real test for Increase and any follower will be navigating a major operational crisis, a network outage, a significant fraud attempt, with its engineered transparency not only intact but providing a demonstrable advantage in rapid diagnosis and resolution. If it does, the "last mile" of finance will finally be paved with code, not guesswork.

Disclaimer: This XOOMAR analysis is for informational and educational purposes only. It is not financial, investment, legal, tax, or professional advice. It does not provide buy, sell, hold, price-target, portfolio, or personalized recommendations. Verify information independently and consult qualified professionals before making decisions.

Why This Changes Everything

• This directly tackles the hidden operational costs and customer support burdens caused by unreliable payment settlement, a pain point for virtually every digital business.
• It represents a strategic shift from merely abstracting payment complexity with APIs to engineering reliability into the regulated banking core itself.
• Success could force both legacy banks and fintech disruptors to improve their settlement transparency and reliability, raising the baseline for the entire payments industry.