Lessons Learned from 10 Years in Embedded Finance with Luke Voiles

In this episode, we’re joined by Luke Voiles, the CEO of Pipe — an embedded finance platform offering credit, cards, and other financial products to a range of B2B2B partners.

Founded as a revenue-based financing business, Pipe pivoted into a platform business under Luke’s directly when he took the reigns in 2023.

Prior to joining Pipe, Luke was the GM of Banking at Square and previously head of QuickBooks Capital, Intuit’s SMB lending offering.

Luke has nearly 10 years of embedded finance experience, especially in lending, and has shaped Pipe’s offering based on his extensive learnings.

In this conversation, Luke and I discuss what makes a good embedded finance offering, the opportunities and challenges in credit, including the economics, cyclicality and risk management considerations, Pipe’s distribution strategy, scaling a fintech platform business internationally and much more.

Will:

It’s great to connect with you. I’ve long been a fan of your work at Pipe, and I’ve followed your journey from distressed debt to Intuit’s QuickBooks Capital, and then to your role as GM of Banking at Square. You’ve had an incredibly accomplished career — and now, you’re leading Pipe.

Just to kick off — super basic question — how do small businesses bank? And how does that potentially differ from what most people assume?

Luke:

A lot of them don’t even have a business bank account. They’re just using their personal account. Here are two good examples.

QuickBooks Self-Employed was created for sole proprietors — single traders — who just needed to file a Schedule C and track business versus personal expenses. They’d usually just link their personal bank account, and then — this was under Alex Chriss, who’s now the CEO of PayPal — his team, including people like John Politi who are also at PayPal now, built this amazing tool. You’d swipe left for business expenses, right for personal. It would help you categorize everything and do your taxes in TurboTax. That’s how a lot of people do it — no business bank account, just one account for everything. And it gets messy.

Here’s another example. One of my favorite QuickBooks insights: their biggest competitor wasn’t another accounting product — it was a shoebox full of receipts. Some small business owners don’t even have their bank data linked in. They’re trying to sort out what qualifies as a business expense at the end of the quarter when their accountant asks, “Hey, do you have any expenses I can use?” And they’re like, “I don’t know. I haven’t been tracking it.” Meanwhile, they’re juggling the business, feeding their kids — they just don’t have time. And that’s the reality for many of these businesses making $50K up to $1M in revenue. You’d be shocked by how chaotic it is behind the scenes.

Another example: at Square, we launched Square Checking. It had tons of users — but the only people who adopted it were the ones who didn’t already have a bank account. So when you onboard for payments, it’s actually a great moment to offer a bank account too. If someone doesn’t already have one, they’ll take it because it’s easy. But if they already have a business checking account? They’re not switching. Timing is everything.

So yeah — it could be a shoebox full of receipts, a simple app, a personal bank account — you name it. It’s just a tough space to navigate.

Will:

So what does the transition look like — from that traditional small business banking model to what you’re building at Pipe? How do you ultimately serve those customers?

Luke:

I don’t think banking should lead. That’s why I believe so strongly in embedded financial services.

What should lead is what our partners are doing — solving real, end-to-end pain points for specific verticals: nail salons, plumbers and electricians, restaurants, retail shops, coffee shops, donut shops. There are so many of these niche verticals now.

The easiest example most people know is Toast — they focused squarely on restaurants and started pulling those customers away from Square. Then Square realized it needed a restaurant-specific product. But the same thing happens again and again. Another vertical pops up, someone builds for it, and suddenly customers are peeling off. Square can’t serve 50 different verticals with one product. It’s horizontal by design.

Same thing with Intuit. Sure, everyone uses it for accounting, but think about where business owners actually live now — in vertical SaaS platforms. They use those to book appointments, manage inventory, take payments, run e-commerce — everything. That’s where the action is. That’s where business owners are getting real time savings and operational leverage.

They don’t go into bank branches anymore to deposit checks or get cash for the next day — it’s all digital now, mostly card-based transactions. So the real problems are being solved by vertical SaaS products, not banks. And that’s exactly why I get excited about embedding financial services there — not as the main feature, but as an add-on that enhances what already works.

Trying to lead with a neobank for small business is just too hard — especially for micro-merchants. Those business owners want time savings, not a fancy bank app. The best way to deliver that is with vertical SaaS built around their specific workflows.

Will:

Makes sense. So, if I try to characterize the business you're building now — it's an embedded finance platform. You started with a lending product, and now you're expanding into card issuance.

You're plugging these financial products into vertical SaaS platforms — like, say, a gym management system that handles class bookings, member onboarding, subscription management, and day-to-day admin. You're integrating financial products into that existing workflow. Is that the right way to think about it?

Luke:

Exactly. And it's all white-labeled.

So, for example, Boulevard offers “Boulevard Capital,” which is powered by Pipe. It’s their brand — the one their customers already trust — and we power the financial infrastructure behind the scenes.

We're working with our partners now to figure out who wants to launch card products. It'll be a full spend management suite — also white-labeled for the partner — and the experience will be seamless. With the flip of a switch, they can turn on additional features.

Down the line, we also want to embed payroll. We’ll build out all the core features: auto-savings for taxes, debit cards, fee-free money movement, card spend, bill pay, payroll. For most of these businesses, that’s where all the spend happens. Once that's in place, then you start to think: “Hey, maybe I could leave Bank of America, because everything I do is already happening right here.”

It starts with pain points — capital, spend management, bill pay — and from there, we layer in additional banking functionality.

And it’s powerful for the platforms. They get to focus on their core customer pain points. They don’t have to take on regulatory or underwriting risk — we handle all of that. And we rev share back 20%, so they pick up 5 extra points of EBITDA without taking on risk. The unit economics just work.

The customers are already in the platform every day. They don’t have to go anywhere else. They see the offers, they manage everything natively, and it becomes the financial operating system within whatever vertical software they use.

For platforms to try to build this themselves — it's incredibly hard. We have a 130-person team now. Building a scalable capital business is one of the hardest things you can do. That’s why we started there. Yes, it drives revenue and helps fund the business, but it's a massive undertaking.

Honestly, unless you’re processing over $100 billion in payments volume, I wouldn’t recommend trying to do this yourself. And even then, you'd still need a 50-person team just to make it work.

Will:

So your capital business — it’s differentiated from traditional lending because it’s based on transactions, right? It’s revenue-based or payments-based lending, instead of a more traditional underwriting model?

Luke:

That’s exactly right. The simplest way to explain it is with a coffee shop.

Let’s say there’s a shop that sells 1,000 coffees a day through one POS terminal. That terminal may only represent 50% of their total revenue — maybe they also have an online store or wholesale — but through that POS, you have perfect visibility into one clear stream of their revenue.

What we’ve done is narrow the underwriting scope to that stream. If we have six months of historical data from that POS, we can very accurately predict the next 12 months of revenue. So if that coffee shop is going to make $1 million, we can advance them $100,000 — and then we collect 10% of their daily card sales until we’re paid back.

That structure gives us seniority from a cash flow perspective, and because we have perfect data, the risk is low and well-defined. We don’t need to look at FICO scores. We don’t need bank statements or credit bureau data. We’re underwriting purely on that transaction history — and it’s incredibly powerful.

Because the risk profile is so strong and narrow, we can originate deals in the morning and sell them to capital markets buyers in the afternoon. It’s balance sheet light and capital efficient — and it’s really the only small business lending product where that kind of model works at scale.

Compare that to what we built at Intuit: we had to understand the full expense side of the business too. That’s much harder. We were able to do it because we had 36 billion accounting transactions and 36 billion bank transactions. But very few players have that level of data access.

Even at Square, we mostly used credit card transaction data. And that’s why the product was so simple — just four clicks. You’re not asking for all this extra information. It’s just: “Here’s the offer. Do you want it?” Approve it, and you’re funded. It feels magical because it’s all automated in the background — the data's already there.

That’s our core product. That’s what we embed. It solves the number one pain point for small businesses — access to working capital. It pays our bills, and it gets us access to real data, which is going to power a lot of what we want to build around AI in the future.

Will:

Embedded finance has been an exciting idea for years now. I think pre-2020, it was largely theoretical. There were certainly people exploring payments integrations, and players like Square — what you guys were doing — were already showing traction. Banking-as-a-service was starting to take shape as a category in fintech.

The promise, or perceived opportunity, was: “We can build great middleware on top of legacy infrastructure. That lets us deliver financial products precisely where and when they should be surfaced.” Think checkout flows on e-commerce sites, onboarding in internet platforms — all these opportunities to insert financial products.

The logic was that we’d see lower customer acquisition costs, higher conversion, and because we could scale horizontally across platforms, we’d build capital-efficient businesses. That was kind of the 2018–2019 thesis.

You've been in this space for a long time now. How would you update that thesis? What have we, as an industry, learned over the past few years?

Luke:

I think a good way to frame it is to look at what didn’t work.

A lot of neobanks — both on the consumer and small business side — thought they could monetize purely on interchange. The model was: issue a debit card, create an FBO wallet, offer some nice features like PFM tools, and make money on that 1% or so per transaction.

But here’s the reality: 1% isn’t enough. Moving money is hard. It’s heavily regulated and complicated. If you want to build a real business, 1% doesn’t get you there.

Look at banks — 80% of their revenue comes from capital and credit risk. You have to take risk to make money. A whole generation of neobanks never figured out credit — and most of them aren’t doing great because of it. Maybe Nubank figured it out. Maybe Revolut to some degree. But most haven’t.

Even something like Chime — early access to paychecks is a nice feature, but it doesn’t solve monetization at scale. And no one has really nailed embedded capital. That’s the part people underestimated: just how critical credit would be.

VCs haven’t loved credit risk historically — too many examples of people getting burned. I was a distressed debt investor for a decade. I bought loans from companies that messed up — structurally or because the market turned on them. There are so many ways to get it wrong that people are scared of it.

But the truth is, if you want to win, you have to be great at credit. You have to execute that part perfectly. Everything else follows.

There was also this belief that wallets could be sticky. “Everyone will use my wallet.” No — the small business owner doesn’t care about your wallet. Even PayPal struggles to get people to leave money in their FBO accounts.

Ask an old-school banker what it takes to convert a small business account. They’ll tell you: $5,000 per customer. That’s what it costs — and even then, you might not retain them.

People assumed customers would flock to their neobank or embedded wallet. They assumed they could monetize with debit interchange or charge SaaS fees for things like bill pay. But now those features are free elsewhere. No micro-merchant is paying $50/month for that.

Some have had success moving upmarket — Bluevine’s a good example. But it’s a slog. In the micro-merchant space, no one’s cracked it yet. The ones who tried didn’t have capital in their roadmap. By the time they tried to bolt it on, it was too late. They messed it up.

That’s what’s finally changing. You have folks like Rex publicly saying: “The money’s not in payments. It’s in credit.” And they’re right. But it’s really hard. You’ve got to be damn good at it to make it work.

Will:

Yeah. Well, to that point—what was it that allowed you to actually build a successful credit product? We’ve talked about how key risk management is. I guess there are two sides to this question.

Let’s start with the first: In terms of building the operations, the risk management, and the underwriting—what were you able to do that others haven’t?

Luke:

It was a really unique setup with Pipe. The company had raised $316 million in equity, and that gave us the opportunity to build a credit business the right way.

Most startups in capital or credit raise $20 million, maybe $30 million, and they’re immediately deploying. But you’re not going to convince a warehouse lender to give you capital to take credit risk that early. You need to prove performance first—and that means you have to deploy equity dollars to start testing.

But then you're faced with a tough tradeoff. Either you grow slowly—deploy a little, see how it performs—or you step on the gas with marketing and risk blowing everything up. It becomes a fight for survival.

That’s why starting a capital or risk-based business is so hard. I’d never do it unless there’s at least $150 million in the bank—enough to give you five years of runway, hire the right team, build everything from scratch, and sign meaningful partnerships. That’s what made Pipe’s situation so unique. There was close to $200 million in the bank when I got here.

So we went all in on building an embeddable Square Capital-style platform—Pipe Capital as a service. We hired veterans with 20 years of experience in every seat. Our chief risk officer had spent 14 years at Amex and seven years at Funding Circle. His head of credit strategy helped build PayPal Working Capital’s early models.

We cherry-picked the best people in the space. Then we bought the data we needed to build the models—hundreds of thousands of small businesses, five years of credit card transaction data. We simulated everything before launching. We did the same thing at Intuit. Most startups skip that and say, “We’ll just deploy and see what happens.” But that leads to messy vintages and years of cleanup.

Because we had the capital, we could be methodical. We modeled our curve, got it dialed in, and now we’re tracking perfectly to it.

So to answer your question—yes, it’s about having the resources, but also about having the experience to know how to deploy those resources properly. Every exec on my team has 20 years of experience in this space. We built the right team, brought in the right data, and just went for it.

That’s what made it work. And honestly, it’s the reason I took the job.

Will:

At Square, you guys were processing payments and also providing the credit. But at Pipe, from what I understand, you're not in the payment flow at all. I think you've done a very elegant job of creating an FBO account structure, where the underlying payments are received and you still have visibility into transaction data—so you’re able to proxy all the same underwriting signals.

Can you talk a bit about the differences between owning the payments flow versus not? And what have you done to account for that difference?

Luke:

Yeah, it's a great question. When you're in the flow of money, like we were at Square, the credit risk is just better—plain and simple.

Being in the flow removes what we call "moral hazard." The business owner doesn’t have to decide whether they’re going to pay you or not. The repayment is automatic. As long as they’re using Square, a portion of each transaction is being held to pay back the advance.

Now, there’s still some evasion. Sometimes a business owner might try to get clever—they’ll shut down their Square account and open a new one under a family member's name at the same location using the same POS terminal. But even then, because you're in the flow, you can see that and reattach the loan to the new user.

At Pipe, we’re not in the payments flow in the same way—but we’ve built an FBO (For Benefit Of) account layer that gives us similar control. We also have the flexibility to start with simpler integrations so our partners can test our product quickly without doing heavy lifting on their end.

For example, we can do a daily back-to-back ACH. We have a fully hosted experience—white-labeled for the partner—where the customer sees the pre-approved offers and goes through the entire process without the partner needing to build anything custom. All we need from the partner is a daily file that shows how much money went into the settlement account.

The next day, we pull 10% of that via ACH. It works—it’s not ideal from a risk standpoint—but it's enough to get started.

Over time, we like to move partners closer to a Square-style setup with either an FBO layer or a true split payment structure. That setup gives us tighter risk controls and more predictability. Before we really scale with a partner, we aim to get as close as we can to that model.

Will:

So if that’s the operational and product flow side of the lending business, let’s talk about the backend — the balance sheet. I assume you’re allocating some portion of equity against each dollar that you lend, at least initially. Over time, I imagine there’s a trajectory where you aim to allocate less or even zero equity per loan, which unlocks greater scale.

Where are you currently in that journey, and what are the levers you’re pulling to improve capital efficiency?

Luke:

Yeah, you’re spot on. The ideal end state is balance sheet light.

That means we can originate a $100,000 advance in the morning — say to a coffee shop — and they agree to pay back $113,000 over the next 12 months. That’s a $113,000 receivable. Then, by the afternoon, we can turn around and sell that whole loan for $106,000. That generates a $6,000 gain on sale, or six points.

That’s the goal — not being a bank, not holding massive amounts of credit risk on our balance sheet. And Pipe is going to get there.

To achieve that, you need completed vintages. We’re talking about 12–18 months of loan performance, and ideally five or six clean, high-performing vintages. So basically two years of track record. If you can prove that, capital markets will lean in — and they’ll take the risk off your hands.

When I was on the other side of the table, in capital markets, I saw this firsthand. You don’t need distressed buyers to take this kind of paper. New York Life Insurance is buying these assets. Pension funds are buying them. If you can get your book rated — Kroll, for example — those institutional investors will take it on directly.

They’re getting 10% IRRs on senior, cash flow–secured positions — which is phenomenal relative to their other fixed income options. Hedge funds like it too, because they can lever that up and get 15%+ returns. The appetite is massive — as long as the asset quality and history are there.

So that’s the North Star. And we’re just a few months away from being able to do forward flow agreements and sell whole loans consistently.

Until then, we have to fund originations ourselves. We’ve got two warehouse facilities with a combined $400 million in capacity. We’re running at high advance rates, which means we don’t need to post much equity per dollar deployed, and we can scale up efficiently. Those facilities alone can support about $1.2 billion in annual origination volume.

By the time we fill that capacity — and we’re close — we’ll be ready to flip to whole loan sales and unlock truly unlimited scale.

But this is why you need serious capital to build a lending business. You can’t start this with $10 or $15 million and expect to make it work. That money gets tied up in your warehouse equity requirements, and then you’re stuck — you can’t originate, can’t raise, can’t scale.

The road is littered with founders who tried to build credit businesses without enough capital or time to prove out their vintages. You need $200 million+ and zero missteps. That’s what it takes.

Will:

In the whole loan sale example, are you taking on any risk at that point? Is there some kind of first-loss coverage, or do you fully offload it?

Luke:

Nope — we fully offload it.

Let’s say we originate a $100,000 advance, and the customer agrees to repay $113,000 over the next 12 months. We sell that $113,000 receivable to a buyer for $106,000. From that point on, all the cash flows go to the buyer.

Now, to be clear, there are carve-outs for things like fraud or first payment defaults. If those happen, they don’t count toward the purchase — we’ll pull those back.

But the fraud risk in our model is extremely low. That’s because we’re working with businesses that have been using their vertical SaaS platform for six months or more. You can’t fake that kind of operational history. It’s not like you can spin up a fake business and fake months of legitimate credit card transactions without triggering chargebacks or red flags. It’s just too hard to spoof that kind of data.

The only real risk we watch closely is account takeover. That’s where a fraudster might try to hijack a payout by changing the destination bank account. But if you're diligent with bank account verification and flagging changes, it’s very manageable.

Occasionally, you might see internal fraud — someone inside the business misdirecting funds. But again, working through embedded, sticky software partners helps mitigate even that. These platforms are deeply integrated with their merchants. It’s very similar to the security posture you get at a company like Square.

Will:

Got it. So by the time you're doing these whole loan sales, are you selling them into a dedicated facility or allocated capacity? Is there some kind of pre-packaging before they go to the secondary market?

Luke:

It’s more like a rolling structure. Think back to what we did at Square — we had, say, 10 buyers, each committing to buy $5 million a month for the next 12 months, as long as the assets met their criteria.

You stagger those commitments so you always have 9 to 12 months of forward capacity lined up. Each month, you’re adding new buyers and refreshing capacity.

Once a buyer takes the loan, they can do whatever they want with it. Some will securitize it, some will tuck it into a closed-end retail fund, some might allocate insurance capital. Doesn’t matter to us.

From Pipe’s perspective, we’ve sold the asset. We still service the advance and handle all customer communications — we’re just passing the cash flows through to whoever bought the loan.

Will:

Got it. So to your point about how you can’t afford to mess anything up—especially around credit risk—how are you thinking about this now, given the current market environment? There’s a lot of uncertainty. It’s not exactly smooth sailing. You’ve got a wide range of customers in different geographies. How are you running your credit business in this environment, and what kind of impact does it have?

Luke:

The structure of our product is a huge advantage in this type of market.

Take a lender like Funding Circle that issues three- to five-year term loans. If you make a five-year loan today, something that happens three years from now can still wipe you out.

Our product is different. It’s either a six-month or a 12-month fully amortizing advance. The weighted average life of the portfolio is only about four and a half months. That means we can turn over our entire portfolio in five months. If we need to change our risk exposure, we can do it really quickly.

Here’s a concrete example. We’re watching consumer discretionary spending closely—because when people start cutting back, they do it in predictable ways. First, they stop buying donuts. Then baked goods. Then dining out. Maybe then coffee. We’ve seen this play out before.

So if we see year-over-year spending dropping in March, we can anticipate it might fall 10% further next month. We adjust our credit policy accordingly.

Instead of assuming a restaurant will generate $1 million in the next 12 months and advancing $100,000, we lower our forecast to $900,000 and advance $90,000. Same structure, shorter exposure, lower risk. We can tune the model every week if we want, and immediately bring the vintages back into line. That’s the power of short duration credit.

On top of that, we manage exposure across verticals and geographies. Nail salons, for instance, tend to be recession-resistant. Same with plumbers and electricians—you’re not going to wait to fix your toilet. So even if discretionary spending drops, that kind of maintenance spending continues.

COVID was a great example. It was devastating for in-person retail, but a boom for e-commerce. So diversification—by industry, by geography—is key. And with our short product cycle, we can constantly rebalance.

This is the only kind of credit exposure I’d feel confident holding in a downturn. If we were sitting on five-year term loans, I’d be panicking right now.

Will:

So those dials you mentioned—can they be adjusted at the sector level, geographic level, that sort of thing? How nuanced is that modeling?

Luke:

We’ve got a world-class machine learning engineering team analyzing the data from every possible angle. If we see pockets of underperformance, we want to understand why—and we proactively dial down exposure in those areas. There are well-known correlations between macro shifts and specific industries. Our seasoned credit team—many of whom are true veterans—can often anticipate those changes and start adjusting ahead of time.

We slice the data every way imaginable: by industry, geography, broader regional trends, term structure, and business size. We don’t currently pull FICO scores, but for a potential upmarket product, we may incorporate that in the future. The short answer is we track and model everything we can. Then we act wherever we see risk—or where we believe risk may be developing. That’s how you stay ahead.

Will:

As you're building Pipe, I imagine there's some level of consideration for how the market values businesses in fintech and financial services—especially in credit.
You touched on capital efficiency and balance sheet usage earlier, and I think that ties into it.
If you look back at how the market has valued companies in this space—like OnDeck, for example, which ended up being a surprisingly poor outcome—it raises the question:
How are you thinking about building a company that's not only operationally strong but also perceived as valuable by the market?

Luke:

The ultimate goal is to have diversified revenue streams.
Yes, capital is a powerful lever—the LTV is high—but that’s just one piece of it.
For context, we recently acquired Glean AI, which was founded by Howard Katzenberg, the former CFO at OnDeck. So we’re very aware of how that story played out.

From our standpoint, capital revenue needs to be balance sheet light.
The math is pretty straightforward: once loans receivable hit 40% of your tangible assets, you're a bank—and you’ll get valued like one, maybe 1–2x book.
That’s not where you want to be if you’re trying to build a high-growth, venture-scale company.
Instead, we’re aiming to stay around 10–15% at most.

Now, that doesn’t mean we won’t hold any assets—we want to keep enough capital to break even even if the market turns. But we’re not building a bank.
We want flexibility, and we want to monetize in multiple ways.

Glean gives us growing SaaS revenue, especially in mid-market AP automation.
We’re also building internal AI agents that we’ll ultimately offer to banks and partners as a SaaS product.
On top of that, we’ve got transactional revenue coming in through card spend management, bill pay, and payroll. These functions generate meaningful float, which we can park with partner banks to get capital discounts.

So Pipe is not a capital business. Capital is a wedge, a monetization layer that pays for everything else and gives us access to rich data.
We're building toward a future where embedded financial services are integrated directly into AI agents and vertical software platforms—and we’re positioning Pipe to play at the center of that.

Will:

So just to clarify—what you’re saying is, when it comes to how the market values your business, you want to avoid being seen as a traditional, capital-intensive lender that’s valued on a price-to-book basis. Instead, you're aiming for a diversified revenue model, where each stream can be evaluated independently and where the business overall is positioned more like a SaaS or tech company. Is that the idea?

Luke:

Exactly. That’s 100% right.

Take AI agents, for example. A lot of startups are saying, “We’re going to build agents for everything.” But any smart VC will tell you: You need focus. You need to start with a single industry and a specific problem. That’s your wedge—and only once you've nailed that can you expand.

The advantage we have at Pipe is we’re already that single industry: financial services. We're building AI agents for our own internal teams—sales, KYC/KYB onboarding, collections, customer success. Each of those roles could eventually be supported by specialized agents.

Now, everyone is building on the same underlying large language models. The differentiator will be in the application layer—how you manage, monitor, and control those agents, especially in highly regulated environments like ours. For example, our sales lead should be able to oversee both human and AI reps, and our compliance team needs full transparency and control.

You can’t have an agent telling a customer, “Take the advance, you don’t need to pay it back.” That would be a massive problem from a regulatory standpoint—FTC, UDAP, you name it.

So where Pipe adds value is by building that app layer—where agents are both effective and compliant. Once we have that tested and working internally, we could turn around and sell it to a bank. Imagine a major bank plugging in our “Sales Agent-as-a-Service,” training it on their own calls and policies, and deploying it alongside human reps. That’s a per-call SaaS model—maybe 10 cents, maybe two dollars a call—but the upside is massive.

There are plenty of startups going after call centers. But because financial services is so complex and compliance-heavy, we believe we’re uniquely positioned to build and sell these kinds of tools. We’re building for ourselves first—but the endgame could be a scalable, enterprise SaaS business selling into banks and financial institutions. And that has very different economics—and valuation potential—than a pure capital business.

Will:

Got it. How do you think about market size? You mentioned the shift over the past few years toward vertical SaaS solutions. From your experience, what portion of GDP—or at least the micro-business economy—is currently being powered by these types of platforms that you can partner with and distribute through? And how much of it still isn't?

Luke:

Honestly, it’s still pretty small today. Most businesses are still using legacy systems—like Heartland for payments or old-school POS terminals sold by ISOs. They might export data into QuickBooks when their accountant tells them to, but they’re not running on true vertical SaaS yet.

That said, the shift is accelerating fast. When we land a new partner, even without adding more partners, we see 50% to 100% growth year-over-year—because our existing partners are growing so quickly. They’re rapidly converting merchants from outdated tools to modern, verticalized platforms that actually solve business pain points.

We’re still early in the transition, which is a huge advantage for Pipe. We get to ride that tailwind. One large partner with 500,000 small business customers could represent $50 million in revenue for us—and if that partner’s growing 50% annually, the compounding is massive. Add a few more partners like that each year, and you’re looking at 5x annual growth.

Every research report I’ve seen points to 25% to 40% annual growth in vertical SaaS adoption. It's still just a fraction of the total market today—but it’s scaling quickly.

Will:

Final question, or maybe a closing topic as we wrap up. You mentioned geographic expansion—excited to hear more about that. But in that context, you’ve essentially repositioned the entire business since taking over. You’ve launched multiple new products, completed an acquisition, and now you’re launching in new countries.

And this isn’t a lightweight SaaS company you’re running—it’s financial infrastructure. That comes with deep integrations, connectivity with bank partners, compliance, and all sorts of complexity. How have you been able to structure the business in a way that allows you to move as fast as you are?

Luke:

Having lived it at Intuit and Square, I’ve seen what slows you down. At Square, we had a ten-year-old Ruby monolith running our banking infrastructure. Launching in a new geo took 18 months because the system wasn’t built for flexibility—basic things like currency support were hard-coded.

So when I got to Pipe, we approached it differently. First, we focused on products that are structurally simpler—like cash advances instead of loans—so we could avoid a lot of the regulatory friction that slows expansion. That dramatically reduces the compliance burden in new markets. You're still responsible for KYC, KYB, AML—but you’re not triggering lending-specific regulation.

Then we picked best-in-class modular tools. We use Alloy for KYC/KYB orchestration. We used Modern Treasury for money movement—same tool we used at Cash App. For servicing, we gave the team the option of tools like Peach, LoanPro, or Canopy. We didn’t reinvent the wheel—we assembled it with great off-the-shelf infrastructure.

That approach made us modular from day one, and that’s been the unlock.

When we expanded to the UK, we launched with GoCardless. They had 80,000 SMB customers and were our first local partner. It only took us six weeks to go live. Why? Because Alloy already worked in the UK. We reused the same decisioning logic and just swapped out the underlying data sources.

Modern Treasury wasn’t optimal in the UK, so we swapped in Airwallex and GoCardless to handle payments. That’s the benefit of building modular systems—you can mix and match what you need for each market without rebuilding the whole stack.

Angela Strange once described FinTech as evolving into “financial services modules.” We’re living that reality. And the beauty is: while the infrastructure is modular, we bring the deep expertise around capital, risk, product design, and customer experience that makes it all work.

For example, card issuance used to be slow. When we launched Square Debit in the UK, even with Marqeta as a global partner, it took 12 months. But now there are modern card issuers who can launch in 40 or 50 geos. Soon, it’ll be 80.

That means Pipe can be a one-stop-shop for large platforms that want to offer capital, card issuance, bill pay, spend management, and other financial tools—across 10 countries at once. Historically, doing that would’ve taken 30 partners, huge engineering lift, and a massive BD team.

But with Pipe’s infrastructure, that same platform can launch in 10 geos, with three products, using one partner—and do it fast. That simply wasn’t possible five years ago.

So yes, financial infrastructure is hard. But the tools are better now, and if you build on them the right way, you can move faster than anyone thought possible.

Will:

Right. So you've replicated the FBO (For Benefit Of) structure in the UK. Did you also need a local partner bank or one of the fintech partners to make that work?

Luke:

Right now, we’re only operating the core capital product in the UK. And that's very intentional—it’s the same playbook we used at Square and now here at Pipe.

You start with the product that’s easiest to launch and delivers the most immediate value: capital. It’s the lowest-lift product from a regulatory and infrastructure standpoint when entering a new market. So that always leads.

Once you have 50,000 to 100,000 small business customers in a given market, that’s when it becomes worth it to layer in the next product.

In our case, that means we’ll roll out card and accounts payable solutions in the U.S. first. After that, the UK will be next, then Canada, followed by Australia, Japan, France, Spain, and Germany—one market at a time.

Once we’ve scaled our existing products across those geos, we have a clear path to move up-market too. We already have a mid-market AP (accounts payable) solution, and card risk is much simpler to manage when you’re underwriting against cash in the account. That opens up the opportunity to eventually go after players like Brex and Ramp—when the time is right.

Will:

Awesome. Luke, I really appreciate all the insight—and this felt like a rapid-fire Q&A in the best way. You’re a fast thinker and a great communicator. Thanks for making the time and sharing your perspective.

Luke:

I hope it was all understandable—or at least something people can slow down to 0.7x speed if I got going too fast.

Will:

Absolutely. Luke, thank you very much for joining us today.

Luke:

This is the stuff I live for—thanks again for having me.