Investment Thesis
Our Thesis
The infrastructure layer of financial services is being rebuilt from the ground up. We back founders who understand the system they intend to replace — not from the outside, but from having worked inside it.
The transformation underway
The financial services industry built its technology infrastructure in layers, and those layers have calcified. Core banking systems designed in the 1970s still process the majority of the world's payments. Credit decisioning still relies on backward-looking scoring models built for paper-based applications. The software architecture that underpins lending, payments, and compliance was designed for a world of batch processing, correspondent banking, and manual review — not for the real-time, data-rich environment that exists today.
What has changed is not the desire to rebuild this infrastructure — that desire has existed for decades. What has changed is that the tools to rebuild it are now mature. Open banking regulation has forced the data portability that makes new payment architectures possible. Cloud-native infrastructure has eliminated the capital requirements that once protected incumbents. And, most significantly, artificial intelligence has arrived at the point where it can genuinely replace the heuristic, rule-based decision-making that sits at the core of financial services — not augment it, replace it.
The window for rebuilding financial infrastructure at the foundational level is open now, in a way it has not been before. Pemberton Ventures was founded in 2019 with the conviction that the next generation of enduring financial services companies would be built at this infrastructure layer. That conviction has strengthened with every passing year. We are early in a very long cycle.
Where we invest
Payments rails & open banking connectivity
PSD2 created the regulatory opening for open banking, but the regulation was the easy part. The commercial infrastructure that sits above it — payment initiation at scale, variable recurring payments that actually work, reconciliation tooling that matches the real messiness of financial data — remains early and underbuilt. We invest in the companies building the plumbing that will make account-to-account payments as frictionless as card payments in practice — not just in principle — and open banking as commercially useful as its advocates have long promised.
The UK is the most advanced open banking market in the world by adoption metrics. That makes it both an interesting laboratory and a genuine commercial opportunity. Companies built here have a path to the EU, and the EU's regulatory trajectory is pulling in the same direction. We are concentrated in UK-headquartered companies with European expansion ambitions.
AI-native credit decisioning
Traditional credit scoring was built for a world of paper applications and monthly bank statements. The FICO score, in its various incarnations, remains the dominant input into lending decisions across the developed world — a scoring methodology designed when the best available data was what you had reported, not what you had actually done. AI allows lenders to underwrite on the full texture of financial behaviour: transaction patterns, cash flow velocity, seasonal rhythms, the gap between what a business reports and how it actually moves money.
The most interesting credit decisioning companies are not building AI overlays on top of existing scoring infrastructure. They are rebuilding the decisioning logic from the ground up — designed around machine learning, with traditional scorecards as one input among many rather than the core logic. This distinction matters enormously for accuracy, for speed, and ultimately for unit economics. We back the rebuilders, not the augmenters.
Embedded finance orchestration
Every software company will eventually offer financial products to its customers — payments, lending, insurance, treasury management — as a native part of their product rather than as a referral to a third party. This shift is structural, not cyclical. The economics are too compelling and the technology is too mature to ignore. What is not yet settled is which infrastructure layer these embedded financial products will be built on.
The embedded finance orchestration layer — the APIs, compliance abstractions, and banking connectivity that let a software company offer financial products without becoming a bank — is early. Most of what exists today is fragile, expensive to integrate, and poorly suited to the compliance requirements of serving financial products at scale. We invest in the companies building the next generation of this infrastructure: developer-first, compliance-aware from day one, and designed to survive the edge cases that current solutions paper over.
What we look for
Financial infrastructure credentials
We back founders who have worked inside the system they want to rebuild. Banking operations, product management at a regulated financial institution, engineering at a payments company — these experiences generate the operational understanding that makes fintech infrastructure hard to replicate. We do not require it, but it is a strong signal when we see it.
Defensible architecture at the infrastructure layer
UI plays and thin wrappers around existing infrastructure are not infrastructure. We look for companies whose defensibility is in the architecture itself — in the proprietary data layer, the hard-won regulatory relationships, the network effects baked into the connectivity fabric. The test: could a well-funded competitor replicate what you have built in twelve months? If yes, it is not infrastructure.
Regulatory fluency from day one
UK and EU financial regulation is not a burden to be managed — it is a competitive advantage for those who genuinely understand it. The best fintech infrastructure companies we have backed treat compliance as a product capability, not a legal checkbox. Regulatory complexity that deters competitors is moat, not overhead. We look for founders who have thought this through before we arrive.
AI embedded in the product core
We distinguish between AI features and AI-native architecture. An AI feature is a language model placed in front of a legacy decisioning system. An AI-native architecture means that the core logic — underwriting, compliance monitoring, fraud detection, cash flow forecasting — is designed around machine learning from the ground up. We back the latter. The former is a roadmap item, not a moat.
Stage and check size
We invest at Pre-Seed and Seed, with selective Series A follow-on into our strongest portfolio companies. Our conviction is highest earliest — when the architecture is being made and the foundational decisions are still live. Later-stage participation is reserved for companies we know well.
We lead rounds or co-invest with conviction. We do not take passive positions in syndicates we have not shaped. When we invest, we are genuinely present — available for the founder conversations that happen at 9pm on a Tuesday, not just the quarterly board meetings.
Our portfolio companies get access to our network of City of London financial services relationships, regulatory contacts, and the operating knowledge of a team that has built and run financial products at scale.
Our funds
Pemberton Fund I
Pemberton Fund I closed in June 2020, establishing the firm's focus on AI-native fintech infrastructure at the early stage. The fund is fully deployed across the first cohort of our portfolio companies, all of which continue to operate and develop.
Pemberton Fund II
Pemberton Fund II closed in November 2023 and is actively deploying into AI-native payments, credit, and embedded finance companies. Fund II reflects the expanded team and a broadened thesis that specifically encompasses AI as infrastructure rather than AI as feature.