KT Cloud has launched an "NPU Server" product applying Rebellions' next-generation neural processing units (NPUs) to its dedicated cloud for public institutions, the company said Tuesday. It marks the first such offering among domestic cloud providers.
KT Cloud commercialized a virtual machine-based NPU infrastructure within its public-sector data center and became the first NPU-as-a-Service (NPUaaS) provider in Korea to obtain Cloud Security Assurance Program (CSAP) certification. The move aims to provide an AI infrastructure environment that meets public-sector security guidelines, in line with the government's policy of expanding the use of domestically produced artificial intelligence (AI) semiconductors.
Operators participating in public-sector AI transformation (AX) projects can simultaneously satisfy security compliance and policy bonus-point requirements through the NPU Server product. Various AI solution companies can also leverage the secured NPU infrastructure to provide AI agent services such as civil complaint counseling, administrative task support, and document search and analysis to public institutions and local governments.
KT Cloud's NPUaaS is based on "ATOM Plus," a data center NPU developed by Rebellions, a Korean AI semiconductor startup. NPUs, which are chips optimized for inference, are considered an alternative for improving the operational efficiency of AI infrastructure while reducing cost burdens.
In April, KT Cloud also launched "AI Nexus," an integrated AI platform available to both public and private customers. AI Nexus combines the existing training-dedicated infrastructure "AI Train" and inference-dedicated infrastructure "AI Serv," offering the entire pipeline from AI model training to inference and deployment on a single platform.
"We launched the public-sector dedicated NPU product in line with the government's stance on expanding the use of domestic AI semiconductors," KT Cloud CEO Kim Bong-kyun said. "As an optimal AI infrastructure platform provider, we will support AX innovation at public institutions and enterprises and contribute to strengthening the competitiveness of the national AI industry."
