← October 22, 2026 briefing · added in the 16:00 KST update

After GTC Berlin, NVIDIA shares European open-model adoption cases in finance, retail, public sector and weather, plus 'DSX' cases targeting the power bottleneck

Announced October 22, 2026

Verified (fact)

NVIDIA published five posts on European industry adoption cases to coincide with GTC Berlin. In finance, Revolut is rebuilding its technology stack around a single foundation model, 'PRAGMA', on NVIDIA accelerated computing, and also uses cuDF. In retail, NVIDIA described a commerce platform that cut inference costs with open models, and a case that uses a digital twin to simulate several robot fleets working together in a fulfillment center. The posts also cover European public bodies running open models on their own infrastructure and adapting them to local languages, field use of Earth-2, an open model for weather and climate forecasting, and use of the DSX platform, which designs the power grid and the AI factory as one system, from substation to rack.

Needs fact-check (claim)

NVIDIA said a commerce platform in the retail case cut inference costs by 90%. The excerpt does not name the company or the baseline for the comparison. NVIDIA also claimed that power is currently the biggest bottleneck for AI infrastructure.

Why it matters (interpretation)

The core message of this GTC Berlin is the 'sovereign' setup: keeping data on your own infrastructure and running open-weight models. These cases show how far that setup is actually being used in heavily regulated sectors such as finance and the public sector.

Sources

Story thread

  1. 2026-10-21 NVIDIA GTC Berlin roundup: DGX Station for Windows in December, cuDSS generally available, SimReady Central 3D asset library launched
  2. 2026-10-22 More from NVIDIA GTC Berlin: Agent Toolkit expansion, 'FOX' factory operations blueprint, and more European cloud partners
  3. 2026-10-22 After GTC Berlin, NVIDIA shares European open-model adoption cases in finance, retail, public sector and weather, plus 'DSX' cases targeting the power bottleneck

More about NVIDIA