Rebellions announced on the 23rd that it has successfully run SK Telecom's proprietary artificial intelligence (AI) foundation model, "A.X K1," on a neural processing unit (NPU) server.
The demonstration ran the 500-billion-parameter (500B) K1 model on Rebellions' NPU server, the "REBEL Server," and included a demonstration of an AI agent service. Rebellions said it had proven infrastructure capabilities comparable to global high-performance graphics processing unit (GPU) servers using only a single-server configuration.
In the AI agent service demonstration, the company tested map-related services and successfully processed multiple simultaneous requests. Even when several users accessed a single service at the same time and entered different commands, the K1 model displayed results instantly on the map without bottlenecks.
Park Sung-hyun, CEO of Rebellions, said, "This demonstration is a practical case of sovereign AI, showing that large-scale AI services can be operated using only the combination of a domestic hyperscale model and a domestic NPU."
