At the GTC Berlin conference, several financial industry participants are presenting their latest artificial intelligence implementations. According to an NVIDIA blog post, banks, payment companies, and financial data providers are currently demonstrating what production-scale intelligence looks like in real-world environments. Rather than relying exclusively on managed third-party platforms, these organizations are running these systems directly on their own infrastructure. In addition, the presented enterprise workloads are built on NVIDIA accelerated computing and powered by open weight models.
Among the organizations detailed at the event, fintech firm Revolut is executing a broad infrastructure transformation. The company is actively rebuilding its technology stack around a single foundation model designated as PRAGMA. Revolut's foundation model architecture is powered by NVIDIA accelerated computing to handle its computing requirements. Furthermore, the company's new technological approach involves using the NVIDIA cuDF data library within its stack.