Embedded Lending has had a hot streak over the last couple of months. At least in terms of announcements. In the first half of this year alone, I covered around 15 launches across five verticals: marketplaces, vertical SaaS, POS and payment providers, banks, and even energy and utilities (see here). But in my conversations, one question kept coming back: How real is this? Are the launches the start of something bigger, or just announcements? I didn't have a good answer, but I knew who would: the people reading EFR. So, we put them in a room and asked.
Around 25 showed up on a group call and not a webinar. Everyone visible, hands going up throughout, and two guests anchoring the conversation: Temi Ofong (HSBC) and Nicolas Kipp (Credibur).
Starting with the definition
A few people asked about the definition when they signed up, so that's where we started. Is embedded lending just digital lending with a nicer name? The answer we landed on: lending should be read broadly, not only loans but leases, subscriptions, receivables, invoices, anything that shifts risk and duration. The embedded part is stricter. It’s credit offered inside a non-financial, transactional context, at the moment it's needed. Someone asked whether that means B2B or B2C, and the honest answer is both, with plenty of blurring in between.
Announcements aren't adoption
Then to the real question. Given how many launches I've covered, I'll admit I walked in slightly biased toward "it's happening." But an announcement is the first step, not the destination. The room kept pulling me back to one number: depending on who you ask, e-commerce GMV runs into trillions, and lending penetration into it is still under one percent. The share flowing through these channels is in the single digits. Most merchants still go to their bank. Everyone knows the option exists. Few are using it, yet.
Why isn’t it scaling faster?
Nobody in the room pretended the product is finished. Product-market fit is still a work in progress, more so in B2B, and the journeys aren't good enough. Cost is higher than conventional funding. Capital is constrained in a model still dominated by fintechs. And there's the gap between what the experience platforms want and what regulation allows. The recurring answer was a closer bank-fintech partnership, which neither side thinks is mature.
Building a financial product inside a non-financial brand? Then you are probably comparing providers, and the shortlist question has many layers: who fits your customer base, who has real experience in your segment, whether a specialist beats a generalist, and how much of the stack you want to own yourself. These are exactly the questions I discuss in my office hours: no pitch, no agenda, just an outside view from someone who tracks this market every week.
What a good platform looks like
A thread I didn't expect to land as hard as it did: not all platforms are created equal, and people have surprisingly firm views on what separates the good ones. Some attachment to payments and enough merchants to hit critical mass. Ideally, the ability to operate cross-border, too. But underneath the metrics sat a simpler test. Is lending a must-have for the platform or a nice-to-have? When it's only a nice-to-have, management never does the unglamorous work to get the product right, and you can tell. More than one person in the room had burnt their fingers backing the wrong platform.
Does it grow GMV?
This is the pitch everyone repeats: lending lifts volume on the platform. The GMV-focused platforms believe it, and some cite a two- to three-times multiple. Clean evidence is harder to come by. What stuck with me was the quieter point about churn: a borrower is simply harder to lose. And the bar doesn't have to be Shopify. A platform running a quiet few hundred-million-dollar book that mostly prevents churn isn't a headline, but it can be exactly what sustains the business. One person took it back nearly a century to a manufacturer in a big-ticket consumer market that started offering credit so buyers could afford its product. It didn't just grow that market; it built it. And yet even today, in that same category, only a fraction of buyers take the embedded route. Decades of proven product-market fit, and traditional channels still win most of the volume. That gap is the whole story in miniature.
One lender or many?
Most platforms start with a single lender because it's easiest, then move to several as they learn that different borrowers need different things. One wants fast cash and will pay for it; another has a long track record and expects bank pricing. Done badly, multi-lender is clunky: apply, get rejected, start over. Done well, it's one application with several lenders behind it. The case that stayed with me came from a consumer platform financing an essential service, where customers expect 12 months at 0% and won't pay the spread. The platform subsidises it; the bank still gets paid. It's only a question of who.
Data is the real advantage
If there was one thing the room agreed on, it's that embedded lending was hyped for distribution, but the lasting edge is data. It isn't inherently riskier, just more expensive early, with funding costs falling as the data builds. The platforms that have made it work sit on transaction data nobody else holds or controls the asset itself: stock in the warehouse, the goods they physically move. The caveat is trust. Swapping traditional credit metrics for platform-performance data surprises people in both directions, pleasantly and unpleasantly, and that learning is still underway.
It's also what makes collateral workable and automated, where the platform already holds the goods. But collateral is where things get interesting, because it pulls platforms upmarket. Nobody secures a 20-30k loan; you do it for the larger tickets, and larger tickets mean bank pricing. That means convincing a bank's risk team to see an industry differently, and risk teams move slowly.
The endgame: convergence
The reward for a platform that gets there is offering rates a borrower couldn't get walking into the bank directly, because the idea that banks are always cheaper doesn't hold. Everyone refinances on the same markets in the end. And once a brand has the data, the collateral, and the pricing, the last step is obvious: it stops partnering for a license and goes and gets its own.
What to watch
Towards the end, one participant pointed to two regulatory shifts the room should have on its radar: the revised Consumer Credit Directive and the EU's new AML package. Both will reshape underwriting, and the AML rules will hit the models that currently get by without full KYC.
Thanks everybody for joining! These events work because they put like-minded people in the same room and let the conversation do the rest. I'll keep running them. If you have a topic that fits the format, reach out.
