Rebellions plans to connect R200 chips to NVIDIA’s platform

The collaboration was announced by NVIDIA, Rebellions and SK Telecom and is described as an effort to advance Korea’s sovereign AI infrastructure.[1] Under the agreement, Rebellions will connect its Rebel R200 NPUs to NVIDIA’s platform through NVLink Fusion.[1] Rebellions also plans to use NVIDIA’s MGX rack architecture to develop rack-scale systems.[1] NVIDIA describes NVLink Fusion as a way for chipmakers to combine specialized silicon with NVIDIA’s scale-up networking and infrastructure ecosystem.[1] The announcement is dated November 8, 2026, while an SK Telecom quote places NVIDIA and Rebellions at NVIDIA AI Day Seoul at COEX on November 9.[1][2]

The companies plan a disaggregated-inference design that assigns different stages of AI response generation to different processor architectures.[1] In the proposed arrangement, NVIDIA GPUs would handle prefill, while Rebellions NPUs would handle decode.[1] The companies say this division of work aims to improve throughput and power efficiency.[1] Rebellions is developing the R200 for low-latency, energy-efficient inference using frontier models.[1]

SK Telecom plans to deploy the systems, including for proprietary AI models such as A.X K2 and public-facing services, including those under Korea’s AI for All project.[1] The company says it will test the integrated systems in its AI data center and assess performance, stability and potential for commercial service use.[2] Rebellions’ R200 systems are expected to be ready for deployment in 2028.[1]

NVIDIA says the MGX ecosystem gives Rebellions access to an established supply chain of original design manufacturers and component suppliers.[1] It also says MGX can let GPU and NPU systems share data-center design elements, including rack footprint, power delivery and cooling.[1] South Korea’s Ministry of Science and ICT supports the collaboration.[1] Separately, the science ministry said it plans to support expanding cooperation with NVIDIA to other domestic NPU companies.[3] The announcement frames NVLink Fusion as a route for Korean-developed NPUs to enter the computing infrastructure market alongside NVIDIA.[1]