Santander NVIDIA AI Collaboration Advances with Three-Year Banking Framework

On October 19, 2026, Banco Santander and NVIDIA announced a three-year framework agreement to expand their AI collaboration across banking applications. The deal builds on existing work at the Santander AI Lab and covers foundation models, AI agents, fraud detection, real-time decisioning, and digital twin simulation technologies.

Technical Tools and Measured Cost Reductions

The Santander AI Lab has created evaluation systems that test open-weight models on specific banking tasks. One tool, called Model Arena, compares performance across different models. Another uses NVIDIA NeMo Customizer to specialize models for individual workloads, while NVIDIA Kumo handles structured data and NVIDIA Dynamo-Triton manages inference integration.

Experimental results show serving costs dropping from roughly five euros per interaction to three euro cents while holding performance levels steady. In a customer-service pilot, conversation quality rose and operating costs fell by more than half. These figures come from controlled tests conducted inside the Santander AI Lab before wider rollout.

The approach moves away from large general-purpose models toward smaller, task-specific ones. This shift supports tighter governance and lower compute demands when models run at global scale across Santander operations.

Strategic Targets Through 2028

Santander expects the AI program to deliver more than one billion euros in combined revenue gains and cost savings between 2026 and 2028. The bank already recorded 35 million euros in AI-related benefits during the first quarter of 2026 and projects over 200 million euros for the full year. These results form part of a larger target to exceed 20 billion euros in profit by 2028.

Chief AI Officer Iñaki Bernal noted that the combination of NVIDIA infrastructure and Santander data scale enables controlled deployment of AI agents. NVIDIA Global Banking Lead Aser Blanco highlighted the metric of cost per intelligence output as a central measure for future work. The framework also supports continued testing of emerging technologies inside the Santander AI Lab before bank-wide use.

The agreement adds to Santander’s existing partnerships with other AI providers and focuses on increasing control, safety, and efficiency as specialized models scale across core banking functions worldwide.