NVIDIA, Rebellions and SK Telecom Announce Sovereign AI Infrastructure Collaboration
The South Korean NPU designer will adopt NVIDIA NVLink Fusion for rack-scale inference, with SK Telecom planning deployment in its AI data center around 2028.
NVIDIA, South Korean NPU designer Rebellions, and SK Telecom announced a collaboration to advance sovereign AI infrastructure, unveiled at NVIDIA AI Day in Seoul on 8 November 2026. Under the arrangement, Rebellions will adopt NVIDIA's NVLink Fusion architecture to connect its next-generation R200 NPUs to NVIDIA rack platforms for rack-scale inference.
SK Telecom plans to deploy the combined system in its AI data center with a target around 2028, drawing on its experience operating AI data centers and its proprietary models. At the same forum, held alongside South Korea's ministry of science, NVIDIA also announced investments in Korean startups including Point2, RLWRLD, Upstage, and Motif.
The announcement extends a trend of telecommunications operators taking on the role of national AI data center operators, a pattern also visible among operators in Southeast Asia. The 2028 deployment target and any performance claims originate from the vendors' joint press release and have not yet been independently verified.