Hello OpenNebula community 
We’re excited to announce that, following our expanded integration with NVIDIA, we’ve published a series of new blog posts to help organizations build secure, sovereign, and production-scale AI factories.
Here’s what’s new:
From Bare Metal to AI Factory: End-to-End Multi-Tenant Automation
Ready to deploy fully automated, production-ready AI Factories? The integration of OpenNebula with NVIDIA NCX Infra Controller provides a unified path from bare-metal GPU servers to fully instantiated AI Factory environments.
The result?
- Predictable provisioning times
- Lower operational overhead
- Improved GPU utilization
- Strong tenant isolation
- Automation across the entire infrastructure stack
Virtualizing NVIDIA GB200 NVL4: Zero Performance Overhead
Traditionally, AI infrastructure forces a choice: bare-metal GPUs for performance or virtualized GPUs for flexibility—with performance loss. OpenNebula removes that compromise, and here’s how:
- PCI passthrough for near bare-metal performance
- Full isolation and multi-tenant control
- Fine-grained MIG allocation
- Efficient GPU sharing without sacrificing speed
Read more: Virtualizing NVIDIA GB200 NVL4: Zero Performance Overhead with Full Isolation
Securing North–South Traffic in AI Factories
Scaling AI, telco, or edge infrastructure introduces networking and security challenges. The integration of OpenNebula and NVIDIA BlueField DPUs offloads networking and security to hardware, enabling:
- Lower latency
- Higher throughput
- Strict tenant isolation
- Unified, controlled operations
Read more: Securing North–South Traffic in AI Factories with OpenNebula and NVIDIA BlueField DPUs
We’d love to hear your thoughts on these new NVIDIA integrations. Which features are most valuable to you, and what else would you like to see in future updates? Let us know in the comments!
Best regards,
OpenNebula Team 