Nvidia, Rebellions and SK Telecom team up on Korean AI infrastructure
The three companies will link Rebellions' Rebel R200 NPU to Nvidia platforms via NVLink Fusion, with R200 systems expected to be deployable from 2028.
Nvidia, Rebellions and SK Telecom announced a collaboration to advance South Korea's sovereign AI infrastructure, with support from the country's Ministry of Science and ICT (MSIT). The announcement was dated November 9, 2026.
Under the plan, the three companies will connect Nvidia's open infrastructure platform, Rebellions' Korean-designed AI processor and SK Telecom's AI data center, model and service capabilities using NVLink Fusion.
Rebellions intends to connect its Rebel R200 NPU to Nvidia platforms through NVLink Fusion and to develop a rack-scale system based on the NVIDIA MGX rack architecture. SK Telecom plans to deploy those systems in its own AI data center, drawing on its operating experience and its model and service capabilities.
The companies also plan to develop a disaggregated inference approach that assigns prefill to Nvidia GPUs and decode to Rebellions NPUs, aimed at improving inference throughput and power efficiency.
Rebellions' R200 systems are expected to be deployable from 2028, according to the announcement.
Quick answers
What will the Nvidia, Rebellions and SK Telecom collaboration involve?
The companies plan to link Rebellions' Rebel R200 NPU to Nvidia platforms via NVLink Fusion and build rack-scale systems based on the NVIDIA MGX rack architecture for SK Telecom's AI data center.
When will Rebellions' R200 systems be available?
Systems based on the R200 are expected to be deployable from 2028, according to the announcement.
Is the South Korean government involved?
Yes. The collaboration is being pursued with support from South Korea's Ministry of Science and ICT (MSIT).