Leading European Financial Services Organizations Embrace Open Models to Build Proprietary Intelligence
European financial firms are adopting open-weight AI models to build proprietary intelligence on NVIDIA’s accelerated computing platform, reducing costs and enhancing proprietary data control.
At GTC Berlin, major European financial institutions demonstrated production-scale AI systems built on NVIDIA’s accelerated computing infrastructure. Revolut is deploying its single foundation model, PRAGMA, using NVIDIA cuDF and Dynamo-Triton software on Blackwell Ultra GPUs to serve 80 million customers, with engineering capacity to scale threefold. Pavel Nesterov, Revolut’s head of AI, emphasized the model’s ability to operate at banking scale across domains, bridging the gap between research and production systems. The collaboration with NVIDIA underscores the feasibility of foundation models in high-stakes financial environments.
Starling Bank became the first U.K. bank to train an agentic model on sovereign cloud infrastructure, developing its ‘Starling Intelligence’ layer using Civo’s U.K. cloud and proprietary transaction data. Lloyds Banking Group introduced ATUM, a transaction foundation model converting billions of banking transactions into reusable intelligence, accelerated by NVIDIA RTX PRO 6000 Blackwell GPUs. SBS launched a sovereign intelligence platform enabling banks to run generative and agentic models on-premises, ensuring data remains within institutional boundaries.
Neobank bunq’s agentic assistant Finn now operates on open-weight models, achieving performance comparable to leading closed models at 60% of the cost per answer when deployed on Nebius with NVIDIA accelerated computing. NatWest is developing technology to recognize diverse British accents, dialects, and financial terminology using NVIDIA’s speech models, including Nemotron 3.5 ASR. Modulr is constructing specialized models for distinct payment operations based on the NVIDIA Nemotron-3-Ultra-550B open model, targeting efficiency in fragmented workflows.
Santander is an early adopter of NVIDIA’s Kumo family of open-weight foundation models for structured financial data, deploying them to build specialized intelligence for core banking workloads. LSEG is collaborating with NVIDIA to accelerate analysis of historical tick data up to 10x faster using the Polars GPU engine and NVIDIA cuDF on RTX PRO 6000 Blackwell GPUs. RavenPack built a dataset generation pipeline on Bigdata.com, transforming plain-language research hypotheses into labeled, backtest-ready datasets using a fine-tuned NVIDIA Nemotron model for market trend analysis.