Santander Boosts AI Performance and Cuts Costs with NVIDIA Partnership

Santander, one of the world’s largest banks, has made significant strides in improving its artificial intelligence capabilities through a collaboration with technology giant NVIDIA. The bank has successfully developed new AI models that enhance conversation quality while reducing costs by more than 50% in experimental customer-service use cases.

As part of this effort, Santander’s AI Lab has been working closely with NVIDIA to explore and test emerging AI technologies before broader deployment across the bank. This lab is responsible for evaluating AI models and agents for banking applications, helping identify which model is best suited for each task and deploy it in the most efficient way.

Santander has been shifting its focus from large general-purpose models towards smaller, specialized models designed for specific banking tasks. In selected experimental use cases, this approach has led to a significant reduction in serving costs, from around €5 per interaction to approximately €0.03, while maintaining strong performance. This achievement is attributed to the bank’s adoption of NVIDIA’s AI infrastructure and tools.

The Santander AI Lab has developed Model Arena, which evaluates open-weight models against banking tasks and identifies the right model for each workload. Additionally, the lab is specializing models using NVIDIA NeMo Customizer, a tool that enables companies to quickly turn structured data into predictive insights needed for agentic services. This approach has delivered strong performance on targeted banking tasks in customer-service comparisons.

The bank’s collaboration with NVIDIA has also led to the development of specialized intelligence for core banking workloads on NVIDIA Kumo, a family of open-weight foundation models for structured data. This innovation allows companies to quickly turn structured data into predictive insights needed for agentic services. The Lab is industrializing inference within Santander’s own architecture using NVIDIA Dynamo-Triton and developing Model Arena to design, simulate, evaluate, and supervise banking agents.

As part of the next phase of their collaboration, Santander and NVIDIA have established a three-year framework covering areas such as foundation models, AI agents, fraud and anti-money laundering capabilities, real-time decisioning, and advanced simulation technologies. This partnership aims to establish a rigorous experimentation path through which emerging AI technologies are optimized and stress-tested against banking requirements before successful capabilities scale across Santander.