NVIDIA, SK Telecom and Rebellions plan to link Korean R200 chips with NVLink
Reported by NVIDIA, Unite.AI, Blockchain News and 2 more outlets
NVIDIA, SK Telecom and Rebellions announced a plan to connect Rebellions’ R200 neural processing unit to NVIDIA’s NVLink Fusion. NVIDIA GPUs would process prompts, while R200 chips generate tokens; deployment is expected in 2028, Rebellions says.
Key points
- The plan assigns prompt processing to NVIDIA GPUs and token generation to Rebellions R200 chips
- Rebellions expects R200 systems to be ready for deployment in 2028
- SK Telecom plans to use the systems in its AI data center with its models and public services
- R200 memory capacity and detailed technical specifications were not announced
Sources
- NVIDIARead ↗From Chips to Skills, NVIDIA and Korea Expand Sovereign AI Collaboration
- Unite.AIRead ↗NVIDIA Deepens Korean Sovereign AI Work on Chips and Rack Systems
- Blockchain NewsRead ↗NVIDIA and Korea Deepen Sovereign AI Collaboration Across Ecosystem
- Business InsiderRead ↗NVIDIA, Rebellions and SK Telecom Announce Collaboration to Advance Sovereign AI Infrastructure With NVIDIA NVLink Fusion
- NVIDIARead ↗NVIDIA Announces New Planned Investments in Korean Startups at the MSIT-NVIDIA AI Startup Partnership Forum
- Tech TimesRead ↗NVIDIA, Rebellions Announce NVLink Fusion Plan: GPUs Process Prompts, Korean Chips Generate Replies
Details
Tech Times reports that the companies announced the collaboration on Nov. 9, Seoul time, with South Korea’s Ministry of Science and ICT supporting the effort. Rebellions will build rack-scale systems on NVIDIA’s MGX architecture, with R200 chips connected over NVLink Fusion.
The announcement does not specify R200’s memory capacity or how the attention-state data would move between the GPU and NPU. It also does not say whether the systems will use NVIDIA’s Dynamo serving framework. Rebellions’ 2028 deployment timeline is an expectation reported by Tech Times, not an independently confirmed result.
SK Telecom plans to deploy the systems in its AI data center and connect them to its own models, including A.X K2, and public services under Korea’s AI for All project. NVIDIA separately says its broader collaboration with the ministry includes training programs, hackathons, work with three selected consortia and four foundation-model builders, and efforts spanning infrastructure, models, agents and security.
Written from coverage as of
Images
- NVIDIA
- Tech Times
- Blockchain News
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh