Nubank (NYSE: NU), which remains Latin America’s largest digital banking / Fintech platform, has shared details of its ambitious artificial intelligence roadmap. During a recent Nu Videocast session, Chief Technology Officer Eric Young and Rohan Ramanath, General Manager of AI Core, explained how the bank’s unique assets are now said to be enabling advancements in lending, customer support, and operational efficiency. Central to this vision is nuFormer, Nubank’s in-house self-supervised foundation model.
Its high accuracy enables the bank to responsibly approve credit for individuals who traditional scoring systems might overlook.
This approach simultaneously aims for broader financial access and supports business expansion.
The model is already operational in the company’s primary credit operations in Brazil and is rolling out to personal loans as well as markets in Mexico and Colombia.
Beyond immediate applications, nuFormer accelerates innovation. Researchers have shortened testing cycles dramatically—from months down to a single day—by creating an automated environment that evaluates cutting-edge techniques from scholarly sources against the bank’s own datasets.
Ramanath highlighted the dual benefits:
“Improved models not only minimize risks but also open doors for more people to access credit in a secure, ethical manner. This creates a powerful combination of social impact and financial returns.”
Nubank’s AI efforts rest on core strengths built over more than ten years.
Its customer base of 135 million users, claiming an 83% engagement rate and strong primary banking relationships, delivers unparalleled insights into spending patterns and financial habits.
The 2024 purchase of Hyperplane further strengthened this by integrating specialists previously at Google, Meta, Apple, and Microsoft.
And, an organization-wide culture embraces AI experimentation. Every team member can use prototyping tools, with leadership tracking adoption and results.
One notable success involved an AI-supported initiative that cut nearly a full year from a complex backend infrastructure overhaul.
Ramanath emphasized:
“Having just one or two is insufficient. Assembling and maintaining all three demands exceptional effort.”
A flagship project is the internal “AI Private Banker,” a suite of tools aimed at empowering everyday users to better organize their money, handle liabilities, and improve borrowing choices.
This brings sophisticated guidance—typically reserved for high-net-worth clients—to a mass audience.
Features have undergone months of trials and now reach over 15 million monthly active users, with a unified offering planned for imminent release.
The effort particularly addresses needs among lower- and middle-income groups, where smarter decisions yield the highest relative benefits.
From a market standpoint, Nubank’s revenue per active customer stood at about US$16 in the first quarter of 2026, below the roughly US$40 average at established banks.
Initiatives in tailored services, expanded product uptake, and the AI advisor are designed to narrow this difference.
Analysts at Morgan Stanley have noted that Nubank is repeating its successful digital disruption approach with AI. That being, acting swiftly, committing fully, and using technology to enhance affordability and user experience while expanding its lead through scale.
Young added:
“We prefer to set new benchmarks rather than follow existing ones. This challenger mindset—reimagining norms and simplifying experiences—is precisely what the current era requires.”
By integrating various datasets, advanced modeling, and inclusive tools, Nubank is positioning itself to adopt AI-driven banking in emerging markets.
