From Chips to Skills, NVIDIA and Korea Expand Sovereign AI Collaboration

The partnership integrates Korean NPU and XPU makers into NVIDIA NVLink Fusion and deploys NVIDIA Dynamo software to domestic model builders.
Key points
- NVIDIA will collaborate with SK Telecom and Rebellions to develop sovereign infrastructure using NVLink Fusion.
- Three consortia and four Korean foundation model developers will use NVIDIA Dynamo software to scale AI inference.
- The program includes NVIDIA training, certifications, hackathons, and seminars to increase GPU utilization.
Why it matters
Korean AI chipmakers and model developers gain direct access to NVLink Fusion interconnects and Dynamo optimization tooling.
Sources
- From Chips to Skills, NVIDIA and Korea Expand Sovereign AI Collaboration (NVIDIA AI, Nov 9, 2026, 5:03 AM)