The quiet reinvention of financial infrastructure
The plumbing of global finance is being replaced not in headlines, but underneath them. What I’ve seen across years of investing in this space, and where I believe the next decade of value gets built.
The most important technology transitions are rarely announced. They happen underneath the surface in the plumbing while everyone watches the application layer. The reinvention of financial infrastructure is exactly that kind of transition. It is already well underway, it is compounding, and most of the people it will affect most haven’t noticed it yet.
What the market is actually doing
Global financial infrastructure the rails, ledgers, compliance systems, and settlement networks that move money between people, companies, and countries was built in a different era for a different set of assumptions. SWIFT, the backbone of cross-border payments, connects more than 11,000 financial institutions but routes transactions through chains of correspondent banks that can take days to settle and layers of fees to clear. ACH, the domestic US rail, was designed for batch processing not the real-time, always-on commerce that defines how businesses and consumers operate today. Core banking systems at the largest institutions run on COBOL written in the 1970s, maintained by a shrinking pool of engineers who know how to touch them without breaking them.
The failure mode of this infrastructure is not dramatic. It does not crash. It simply creates friction at every seam friction that gets priced into fees, absorbed into wait times, and worked around through workarounds that create more seams. The system functions. It functions poorly, expensively, and in ways that disproportionately affect the businesses and individuals with the least leverage to demand better.
Sources: SAP Fioneer cross-border payments analysis Nov 2025; GMI Insights embedded finance report Jan 2025; IMARC Group 2025
What changed is not the legacy system. It has not meaningfully changed. What changed is everything around it customer expectations, API infrastructure, regulatory frameworks for open banking, the economics of cloud compute, and critically, the emergence of a generation of founders who grew up using software that worked instantly and could not understand why money should behave differently.
What I saw and what took me time to understand
The clearest signal I got that something fundamental was shifting did not come from a market report. It came from watching a company I had backed an infrastructure layer play, at the time largely uncategorizable by standard fintech taxonomy navigate a period when no one else in the market understood what it was building or why it mattered.
The company was not building a consumer app. It was not building a neobank. It was building the connective tissue between financial institutions and the software businesses that increasingly needed to embed financial functionality without becoming regulated entities themselves. When I first backed it, the category did not yet have a name that investors recognized. The pitch deck had to spend the first five minutes explaining what the problem was before it could explain the solution. Most investors I spoke with at the time passed not because they thought the company was wrong, but because they did not have a mental model for the category it was creating.
That experience taught me more about how infrastructure investment works than any framework I have read since.
What I observed across that investment and across the broader portfolio in the years that followed was a consistent pattern. The companies building the infrastructure layer almost always looked too early, too technical, and too far from consumer validation to generate the kind of excitement that drives competitive processes. They were also almost always the companies that turned out to matter most.
The cumulative version of this pattern crystallized through three repeated observations. First: every portfolio company that tried to build on top of legacy financial infrastructure eventually hit a ceiling imposed by that infrastructure not by their own capabilities. Second: the companies that built or found access to modern infrastructure consistently outgrew their peers on the same metric at 2–3× the rate. Third: the infrastructure providers themselves were almost universally undervalued relative to the application-layer companies they enabled, right up until they were not.
What I learned from being early to the category
Backing an infrastructure play before the market has a name for the category is an uncomfortable position to hold. The early signs of rightness the founder’s clarity, the technical architecture’s elegance, the early customer conversations are all things that are hard to show to an LP or a co-investor who wants to see a comp set and a TAM slide. The thesis lives in your head, assembled from pattern recognition that is difficult to externalize.
What I learned is that infrastructure investment rewards a specific kind of patience that is different from the patience required in application layer investing. In application layer investing, patience means waiting for a market to catch up to a product that is ahead of its time. In infrastructure investing, patience means waiting for the market to realize that the foundation it has been building on is fragile and that someone has already built a better one.
The moment that consistently signals a financial infrastructure investment is about to be re-rated is when a large enterprise customer a bank, an insurer, a payments network publicly announces it is moving a core process onto a new infrastructure provider. Not piloting. Not evaluating. Moving. That announcement is almost always preceded by 18–24 months of quiet adoption that never showed up in press releases.
By the time the announcement happens, the investment opportunity has usually moved to a different point on the risk-return curve. The signal that matters is the quiet adoption period finding it before it becomes legible to the rest of the market.
The second thing I learned is that the most durable infrastructure companies are the ones that make their customers structurally dependent in ways that are genuinely value creating, not just lock-in for its own sake. The distinction matters. Lock-in through switching costs alone produces customers who are resentful and looking for an exit. Lock-in through compounding value — where the infrastructure gets more useful the more data flows through it, the more integrations are built on top of it, the more institutional knowledge accumulates in using it produces customers who become advocates and who help you sell to the next customer. Every infrastructure company I have seen scale to a durable position has that second kind of dependence, not the first.
The investment thesis, as I hold it today
Financial infrastructure is being rebuilt in layers, from the bottom up, and the layer I find most compelling over the next three to five years is embedded finance specifically the Banking as a Service infrastructure that allows non-financial businesses to offer financial products without becoming regulated financial entities themselves.
The thesis has four components that reinforce each other.
The companies with the most direct relationships to end customers e-commerce platforms, gig economy operators, B2B software vendors, logistics networks are not banks. But they are sitting on distribution that banks would pay almost anything to access. Embedded finance inverts the traditional model: instead of banks distributing financial products through their own channels, non-financial platforms distribute financial products through their existing customer relationships, with BaaS infrastructure handling the regulated back end. The platform wins on engagement and monetization. The infrastructure provider wins on volume. The bank partner wins on deposits and regulatory coverage they could not generate themselves.
Consumer embedded finance buy now pay later, embedded insurance, neobank accounts has received most of the attention and capital. The B2B layer is larger, stickier, and earlier in its adoption curve. The B2B embedded payments market is projected to grow from $0.7 trillion to $2.6 trillion by 2030. Enterprise purchasing purchase orders, net-30 terms, manual reconciliation, separate credit relationships is being collapsed into embedded experiences where the buyer gets credit at checkout, the seller gets paid immediately, and the platform earns a spread. The founders building this infrastructure are, in my experience, still raising at multiples that do not reflect the eventual scale of the opportunity.
The regulatory environment for embedded finance and BaaS has been tightening particularly around fintech-bank partnership structures. Several BaaS providers faced regulatory action in 2024 and 2025 for compliance gaps in their partner oversight. The short-term effect is pain for underprepared players. The long-term effect is competitive consolidation around the infrastructure providers that built compliance as a first-order design principle rather than a retrofit. The winners of the BaaS consolidation will be worth significantly more than the pre-consolidation market implied. I am looking for infrastructure companies that treated regulatory complexity as a moat rather than a cost.
The infrastructure providers that are processing the most transaction volume are accumulating proprietary risk models, fraud signals, and behavioral data that cannot be replicated by a new entrant. This is the compounding advantage that makes financial infrastructure investment different from most software categories. The product gets meaningfully better not marginally better with every additional unit of volume. The companies that are 3–5 years into building this flywheel have an advantage that is not visible on a revenue multiple or an ARR chart but is very visible in the conversations I have with their enterprise customers about why they would never switch.
Where I think the market goes from here
Three things will define the next five years of financial infrastructure reinvention, and I am positioning the portfolio around all three.
BaaS consolidation creates category winners
The BaaS market is overcrowded and under-regulated. The players that survive the coming consolidation driven by tightening oversight of fintech-bank partnerships will have built compliance infrastructure as a core competency, not a layer on top. Those survivors will command pricing power and customer loyalty that the current market does not reflect. I am looking for the two or three companies that emerge from this consolidation with dominant positions.
Real-time rails go cross-border
Nearly 80 markets now have domestic real-time payment networks. The next major infrastructure build is connecting them. The G20 has made cross-border payment modernization a formal priority. ISO 20022 standardization is creating the common language that interoperability requires. The infrastructure companies that position themselves at the connection points between domestic real-time networks will be building the next generation of SWIFT but programmable, API-native, and designed for the speed that modern commerce requires.
AI makes compliance a competitive weapon
Compliance has historically been a cost center in financial infrastructure. AI is making it a moat. The infrastructure providers that use AI to make their compliance processes faster, more accurate, and more adaptive to regulatory change will win customers from incumbents whose compliance stack is a manual, expensive liability. This is a large and underappreciated opportunity and one where I am actively looking for founders who understand both the technical and regulatory depth required.
Vertical software eats financial services
The long-term direction of embedded finance is that every vertical software company becomes a financial services company. The construction management platform, the healthcare practice management system, the logistics operator each of these will eventually offer payments, lending, and insurance as embedded products because the data they have about their customers makes them better underwriters than any bank. The infrastructure layer that enables this without requiring every vertical software company to become a regulated entity is the category I am most focused on building exposure to.
The bottom line
Financial infrastructure is not a glamorous category. The companies building it rarely make headlines. Their products are not consumer-facing, their founders are not frequently on stage at the conferences that generate coverage, and their growth metrics are often denominated in transaction volume and API calls rather than in the monthly active user numbers that make for clean narratives.
That is precisely why the opportunity persists. The market systematically undervalues infrastructure because infrastructure is hard to see until it is everywhere and by the time it is everywhere, the investment opportunity has moved. I backed one infrastructure company before the category had a name. I intend to back several more while the same dynamic holds.
The reinvention of financial infrastructure is not coming. It is already happening underneath the surface, in the plumbing, where the most important technology transitions always begin. The founders building it are solving problems that most people in finance have simply accepted as permanent features of how money moves. They are not. And the companies that prove that, at scale, will be among the most valuable built in this decade.