NVIDIA Cloud Partners Expand European AI Infrastructure with Vera Rubin and OpenShell Security

NVIDIA Cloud Partners are expanding regional compute capacity for model training and inference, establishing secure boundaries for AI agents, and supporting agentic and physical AI applications across Europe, the Nordics, Africa, and Central Asia.
Nscale demonstrated that NVIDIA DSX MaxLPS enabled 37% more GPUs and nearly 50% higher aggregate inference throughput within the same provisioned power budget using NVIDIA GB300 NVL72 systems.
The NVIDIA OpenShell Reference Architecture, announced at GTC Berlin, provides full-stack control for AI agents using NVIDIA Vera Rubin NVL72 for inference, NVIDIA Vera CPUs for agent harnesses, and NVIDIA BlueField-4 DPUs running NVIDIA Sentry for in-silicon enforcement.
Mistral Large 4, the company's latest frontier multimodal model, was trained on 3,800 NVIDIA Grace Blackwell GPUs in European data centers, with NVIDIA Vera Rubin infrastructure planned for delivery in 2027.

Infrastructure and Model Expansion

Crusoe is developing Spark, liquid-cooled modular AI factories for the NVIDIA Vera Rubin platform, to serve low-latency AI inference for clients including Base44, Gradium, LTX, and Wonderful.
Nebius adopted NVIDIA Groq 3 LPX with Spectrum-X Ethernet networking for production inference and offers a Physical AI Workbench for robotics companies like Rhoda AI, RoboForce, and Sereact.
Fireworks and Era4 are partnering to bring open models to enterprises in the U.K. and Europe using NVIDIA Blackwell Ultra GPUs at Era4’s Chesterfield facility.

The OpenShell Reference Architecture is supported by partners including Aleria, Baseten, Cassava, CoreWeave, Crusoe, DigitalOcean, Firebird, Fireworks, Firmus, G42, GMI Cloud, IREN, Lambda, Lightning AI, Nebius, Nscale, Radiant, Verda, and Vultr.
The architecture ensures agents operate within defined boundaries for tool use and system access, preventing actions beyond what organizations permit.