
Embedded lending infrastructure
Turn this teardown into a decision-ready prompt for ChatGPT, Claude, or your agent.
If you only have a few minutes to spare, here’s what investors, operators, and founders should know about R2 (W21).
R2 built the lending department that Latin American marketplaces, payment processors, and delivery apps did not want to build themselves. Its APIs let a platform offer working capital under its own brand while R2 handled underwriting, funding, compliance, servicing, and collections.
The company did not fail. Ant International made a strategic investment in October 2025, earlier investors called it a meaningful exit, and the founders continued running R2 across five countries.[1][2] R2 succeeded because it treated fragmented regulation, scarce credit data, enterprise integrations, and loan capital as one product. That integration created the strategic value Ant wanted and explains why a standalone lender would have been a weaker business.
Roger Larach grew up in Honduras in a family of entrepreneurs, worked in Latin American investment banking, and helped Ualá launch in Mexico. Roger Teran, from El Salvador, began by helping Central American small businesses find growth capital, then moved through consulting, data science, and analytics.[3] Their backgrounds joined capital markets to operating data.
The first plan was to launch in Central America. Platforms showed interest but would not trust an unproven lender with a live pilot. Larach later said, “Primero mostraron interés, pero después, al no tener el track record, las oportunidades no se concretaron.” They turned to Mexico, where Sr. Pago accepted a pilot, and to Ecuador, where Rappi provided a smaller entry point that later expanded into a regional relationship.[4]
COVID-19 made the problem sharper. Banks reduced credit to small businesses just as commerce moved online. The founders wrote that cash transactions fell 35%, forcing merchants to buy and sell digitally and creating transaction histories on the platforms they used.[5] R2 could underwrite from sales behavior when a traditional lender lacked usable statements or collateral.
The founders described the link personally: “Como fundadores Centroamericanos, entendemos de primera mano el impacto positivo que tienen las pequeñas empresas.” Their families had sought business capital in both good and difficult periods.[5] R2 joined YC's Winter 2021 batch and built as a regional company from the start, with distributed technical and credit teams rather than a Mexico-only lender.[3]
R2 sold infrastructure to companies that already had daily relationships with small merchants. A point-of-sale system, marketplace, delivery app, or payment processor integrated R2's APIs and displayed financing inside its own interface. Merchants saw an offer based on their platform activity, accepted without a conventional bank application, received working capital, and repaid in proportion to sales.
R2 handled the hidden stack. It acquired transaction data, produced risk scores and offers, authenticated borrowers, met local compliance requirements, supplied or arranged capital, serviced the financing, reconciled payments, and collected overdue balances. The platform partner kept the customer experience and added financing revenue without taking credit risk.[10]
The product crossed national boundaries but could not ignore them. Each country required local entities, lending or correspondent relationships, disclosures, policies, and capital structures. R2's technology normalized the partner experience while its operating team managed the exceptions. That combination was harder to copy than an API alone.
Read the complete post-mortem, the rebuild playbook, and the exact reasons R2 is still worth studying now.