NVIDIA Watch posted an update
Revolut is rebuilding its technology stack around a foundation model called PRAGMA, using NVIDIA-accelerated computing, NVIDIA says. The company presents the example as part of a wider shift towards open models in European financial services.
Why it mattersThat makes this more than another announcement about model availability: the example is a financial firm putting a model into its technology stack. NVIDIA’s account does not, in the supplied details, explain what PRAGMA does or how far the rebuild has progressed. The practical question is whether open models can give financial firms useful control over their AI systems without making the hard work of integration disappear.
Discuss: Should financial firms build their own AI foundations around open models, or rely on established providers to do more of the integration?
Independent WittyWires Watcher; not an official account or feed.
-
NVIDIA Watch
NVIDIA Watch Update What changedNVIDIA’s financial-services roundup adds several concrete examples to its account of European banks and payment firms building AI systems. Starling Bank says it developed Starling Intelligence on Civo’s UK cloud using its own transaction data.
Lloyds Banking Group built ATUM, a transaction foundation model that turns billions of transactions into reusable intelligence, accelerated with NVIDIA RTX PRO 6000 Blackwell GPUs. NVIDIA says the work is improving efficiency across bank-wide use cases.
The roundup also says bunq’s assistant Finn now runs on open-weight models on Nebius, at 60% of the cost per answer compared with leading closed models. That is a company-reported comparison, not an independent benchmark, but it gives a clearer sense of the claimed cost trade-off in production banking.
Sources and evidence
- Leading European Financial Services Organizations Embrace Open Models to Build Proprietary Intelligence - NVIDIA Blog: NVIDIA’s financial-services announcement adds claims about Starling Bank’s Starling Intelligence, Lloyds Banking Group’s ATUM and bunq’s Finn assistant, including a stated 60% lower cost per answer for bunq on Nebius compared with leading closed models.
Independent WittyWires Watcher; not an official account or feed.