Revolut scales pragma for 80 million customers as European banks adopt NVIDIA open models
At GTC Berlin, a range of European banks, payment firms and data providers demonstrated production‑scale AI built on NVIDIA accelerated computing and open‑weight models. Revolut is rebuilding its stack around a single foundation model called PRAGMA, running on NVIDIA Blackwell Ultra GPUs with the cuDF toolkit and Dynamo‑Triton software. The system already serves an initial 80 million customers…
Key points
- Revolut runs PRAGMA on NVIDIA Blackwell Ultra GPUs, serving an initial 80 million customers and engineered to scale 3×.
- bunq’s Finn assistant matches closed‑model quality at 60 % of the cost per answer using NVIDIA open weight models.
Starling Bank trained an agentic model on U.K. sovereign cloud infrastructure, creating a "Starling Intelligence" layer on Civo’s cloud. Lloyds Banking Group released ATUM, a transaction foundation model that ingests billions of banking transactions, while bunq’s Finn assistant runs on open models at about 60 % of the cost per answer. NatWest is developing speech technology with NVIDIA’s Nemotron 3.5 ASR, and Modulr and Santander are building specialized models on Nemotron‑3‑Ultra‑550B and the Kumo family respectively.
These deployments show that open‑weight models combined with NVIDIA GPUs can give European financial institutions proprietary intelligence while controlling costs and latency, narrowing the gap between research breakthroughs and production banking systems.
Leading European Financial Services Organizations Embrace Open Models to Build Proprietary Intelligence
NVIDIA Blog · 22 October 2026
Loading the full article…
This text was published by NVIDIA Blog and written by Jason Renders. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Enterprise & Industry
All →- Nvidia powers European retailers with open models, cutting inference cost by 90% · 1 src
- EBU partners with NVIDIA to expand sovereign AI framework for European broadcasters · 3 src
- ServiceNow and Nvidia release Arc AI agent with security-focused guardrails · 2 src
- Opinion: former Microsoft HR VP outlines three ways to use AI at work · 1 src
- Cloudflare pairs OpenAI Daybreak with network data to patch software flaws · 1 src
Comments
via GitHub Discussions