Back to feed
Artificial intelligencePrimary source

NVIDIA, Rebellions and SK Telecom to link R200 chips with NVLink Fusion

NVIDIA, Rebellions and SK Telecom announced a collaboration to develop South Korea’s sovereign AI infrastructure. Rebellions will connect its R200 neural processors to NVIDIA’s platform, while SK Telecom plans to deploy the systems in its AI data center; deployment is expected in 2028.

Preferred on Google
Text size

Image: nvidianews.nvidia.com · source articleEditorial excerpt for news reporting

NVIDIA, Rebellions and SK Telecom announced a collaboration to advance South Korea’s sovereign AI infrastructure. Rebellions will use NVLink Fusion to connect its Korean-designed Rebel R200 neural processing units to NVIDIA’s open infrastructure platform.

SK Telecom plans to deploy the systems in its AI data center alongside proprietary models and public-facing services, including those developed under the government’s “AI for All” project. Rebellions’ R200 systems are expected to be ready for deployment in 2028.

R200 chips enter rack-scale development

Rebellions is developing the R200 for low-latency, energy-efficient inference, including mixture-of-experts models with multitrillion parameters. The NPUs will connect within racks through NVIDIA NVLink scale-up networking, while NVIDIA’s MGX rack architecture and reference designs will support system development.

The companies also plan a disaggregated inference approach: NVIDIA GPUs would handle prefill and Rebellions NPUs decode. The participants say the planned design is intended to improve inference throughput and power efficiency.

Government backs Korean AI infrastructure

South Korea’s Ministry of Science and ICT supports the collaboration. The project is intended to link Korean semiconductor development with NVIDIA infrastructure and SK Telecom’s operational capabilities to deliver AI services tailored to Korea’s language, industries and priorities.

South Korean Deputy Prime Minister and Minister of Science and ICT Bae Kyung-hoon said the government would support the project, including cooperation with NVIDIA on physical AI and GPU supply. SK Telecom plans to connect its data centers, proprietary models including A.X K2, and public-facing services as part of its full-stack sovereign AI strategy.

What we know

  • Rebellions will connect Rebel R200 neural processors to NVIDIA’s platform through NVLink Fusion.
  • SK Telecom plans to deploy the systems in its AI data center with proprietary models and services.
  • R200 systems are expected to be ready for deployment in 2028.
  • The planned design assigns prefill to NVIDIA GPUs and decode to Rebellions NPUs.
  • South Korea’s Ministry of Science and ICT supports the collaboration.

Trust: Primary source

  • The story cites a primary statement. It establishes what its author said, but does not guarantee every claim is true.
  • Sources: NVIDIA Newsroom.
Related news
View sources1

COMMUNITY

Discussion

—

Loading comments…