Nubank's AI-Fueled Growth Hits $5B Revenue, But Risks Mount
- $5B Revenue: Nubank achieved $5 billion in revenue for the first time in Q1 2026.
- 135M Customers: The company now serves over 135 million customers.
- 5.0% NPL Ratio: Early-stage credit delinquencies rose to 5.0%, up 89 basis points from the previous quarter.
Experts would likely conclude that Nubank's AI-driven growth strategy is delivering impressive results, but the rise in early-stage credit delinquencies underscores the need for careful risk management as the company expands.
Nubank's AI-Fueled Growth Hits $5B Revenue, But Risks Mount
SÃO PAULO – May 14, 2026 – Digital banking giant Nu Holdings Ltd. (NYSE: NU) showcased a powerful first quarter, smashing the $5 billion revenue barrier for the first time and reporting a net income of $871 million. The fintech powerhouse, now serving over 135 million customers, also announced a pivotal milestone: its Mexican operations have achieved profitability, validating its expansion strategy. Yet, beneath the impressive growth, a rise in early-stage credit delinquencies raises questions about the delicate balance between rapid expansion and sustainable risk management.
The results underscore a strategy heavily reliant on advanced technology. “We are not adding AI to banking, we are rebuilding banking around AI,” founder and CEO David Vélez stated, a sentiment that echoed through the quarter's achievements. This AI-first approach is central to the company’s narrative as it continues its aggressive growth in Latin America and sets its sights on the competitive U.S. market.
Rebuilding Banking from the Algorithm Up
At the heart of Nu's strategy is a profound integration of artificial intelligence that goes far beyond industry norms. The company is leveraging NuFormer, its proprietary suite of foundation models, to fundamentally reshape core banking functions. This technology is already in production, powering credit card decisions in Brazil and Mexico and underwriting unsecured loans in Brazil. The impact is tangible: Nu reports that NuFormer has delivered a threefold performance uplift over traditional machine learning in optimizing credit card limits, allowing for significant credit line increases without a corresponding rise in overall risk appetite.
This AI-driven precision enables real-time, individualized loan approvals. Every personal loan request is now priced and approved based on its predicted net present value in under a second, a capability that has been a key driver of the credit portfolio's 40% year-over-year expansion.
Beyond underwriting, the company is deploying AI to transform customer interaction. Its 'AI Private Banker' functionality—offering financial insights, payment assistance, and credit advice—is already being used by over 15 million customers monthly. Vélez envisions this tool democratizing financial guidance, providing personalized advice to millions of customers who have never had access to such services. This push is also yielding internal efficiencies, with the company reporting a 50% year-over-year increase in engineering throughput and 90% faster testing cycles, freeing up resources to focus on innovation.
The Blueprint for Expansion
Nu's success in its home market of Brazil, where it now counts over 115 million customers and stands as the largest private financial institution, provided the playbook. The question was whether that model could be exported. The first quarter of 2026 provided a definitive answer. In Mexico, Nu reached 15 million customers, became the third-largest financial institution by customer count, and, crucially, achieved break-even profitability.
This inflection point was reached by following the same formula: rapid customer acquisition fueled by a superior digital product, followed by effective monetization. Over four years in Mexico, the customer base has grown sevenfold while average revenue per active customer (ARPAC) has nearly doubled. This success, along with steady growth in Colombia where the customer base is approaching 5 million, has emboldened the company to look north to the United States.
However, the U.S. strategy is markedly different. Described as a “measured, capital-efficient approach,” the expansion is designed to test the waters without risking the core business. Nu has capped its U.S. investment at a level expected to be below 100 basis points of its consolidated efficiency ratio through 2027. After receiving conditional approval for a national bank charter from the U.S. Office of the Comptroller of the Currency (OCC) in January, the company is methodically working toward a full license, a process that could take up to 18 months. The initial focus will be on simplifying access to core products like credit cards and personal loans, potentially leveraging the massive remittance corridor between the U.S. and Latin America as a customer acquisition channel.
Balancing Growth and Risk
While revenue and customer numbers soared, Nu’s asset quality metrics revealed a more complex picture. The company's leading indicator for delinquencies, the 15-90 day Non-Performing Loan (NPL) ratio, rose to 5.0%, an increase of 89 basis points from the previous quarter. Management attributed the majority of the increase to predictable first-quarter seasonality and intentional portfolio growth, including a calculated expansion into higher-risk segments to broaden customer reach.
Company executives stressed that this was a planned move, enabled by the confidence provided by their sophisticated AI risk models. They point to a more encouraging sign in the data: the 90+ day NPL ratio, which represents more entrenched delinquencies, actually declined by 10 basis points to 6.5%, remaining well below its 2024 peak. This suggests that while more new customers are showing early signs of stress—a common occurrence with portfolio expansion—the company has so far been effective at preventing these early issues from escalating into long-term defaults.
This dynamic highlights the central tension in Nu’s business model. The company is betting that its superior data and AI-driven underwriting can profitably extend credit to segments of the population that traditional banks have deemed too risky. The 33% quarter-over-quarter increase in credit loss allowances to $1.79 billion reflects the immediate financial impact of this strategy. The ultimate test will be whether the AI models can consistently outperform the market in pricing this risk, ensuring that the aggressive growth translates into sustainable, long-term profitability.
📝 This article is still being updated
Are you a relevant expert who could contribute your opinion or insights to this article? We'd love to hear from you. We will give you full credit for your contribution.
Contribute Your Expertise →