The lending infrastructure market is in the middle of a structural debate that has significant consequences for how infrastructure providers price their services and which types of capital will fund the growth of embedded lending. The debate is not about whether embedded lending works — the distribution advantage of offering credit within a software context where the lender has superior data is now well-established. The debate is about who bears the credit risk when those loans are made, and what the implications of different risk allocation structures are for the infrastructure layer that sits between the loan origination and the balance sheet that funds it.
The originate-to-distribute model has been the dominant structure for embedded lending since the category scaled meaningfully in 2021 and 2022. Under this structure, the software platform or the embedded lending infrastructure provider originates the loan using its credit decisioning model and data, and immediately sells it — or transfers it — to an institutional balance sheet: a bank, a specialist lending fund, or a structured vehicle. The infrastructure provider earns an origination fee and, sometimes, a servicing fee for ongoing loan management. The credit risk sits with the institutional funder from the point of transfer. The commercial logic is clean: the infrastructure company does not need regulatory capital to fund a loan book, its economics are fee-income rather than net interest margin, and its operational risk profile is more like a SaaS business than a balance sheet lender.
The structural vulnerability of originate-to-distribute becomes apparent in a credit tightening cycle. The institutional funders who purchase originated loans adjust their credit appetite and purchase pricing based on their own portfolio performance and capital requirements — factors that may have nothing to do with the performance of the specific embedded lending programme they are funding. A funder who experienced losses in consumer credit during 2022 and 2023 may reprice or reduce capacity for all origination relationships, including those that have performed well. An embedded lending infrastructure company whose revenue depends on continuous loan transfer is therefore exposed to funder risk in a way that is structurally uncorrelated with its own product quality. This counterparty dependency is the risk that the originate-to-distribute model transfers from the balance sheet to the funding relationship.
The risk-sharing model addresses this by giving the infrastructure provider — or the distribution platform — genuine skin in the credit performance. In a first-loss structure, the platform retains a subordinated tranche of the loan book, absorbing losses up to a defined threshold before the senior funder is impacted. This aligns incentives: the platform's financial interest is directly tied to credit quality, not just origination volume. The infrastructure required to manage a first-loss tranche is meaningfully more complex than a pure originate-to-distribute model: the platform needs loan accounting infrastructure, regulatory capital management, credit portfolio monitoring, and the compliance framework appropriate for a firm with direct credit risk exposure. These requirements raise the bar for infrastructure quality, but they also provide a defensible answer to the question of why the infrastructure company's underwriting quality is worth paying for.
The infrastructure layer that is currently underbuilt is the loan lifecycle management tooling for embedded lending programmes that operate at significant volume across multiple distribution channels. A single embedded lending platform may be originating loans through five or six software platform partners, each with a different customer profile, a different data set, and a different risk tolerance. Managing the credit performance attribution across these programmes — understanding which channel is contributing to portfolio quality and which is degrading it — requires infrastructure that sits above the individual loan servicing system and provides a consolidated view of performance by origination channel, product type, and vintage. This is the type of infrastructure that becomes essential as embedded lending matures from early deployments into established credit programmes, and it is where we expect to see new investment targets emerging over the next eighteen months.