Connecting Domestic NPU to NVIDIA GPU…SK Telecom and Rebellion Join Forces - 머니투데이

The three companies signed the memorandum of understanding at Nvidia AI Day Seoul 2026, held at COEX in Gangnam-gu, Seoul on November 9. The agreement establishes a framework for building a heterogeneous AI computing system that links Nvidia's GPUs with Rebellions' domestically developed neural processing units (NPUs) at SK Telecom's AI data centers.

NVLink Fusion Technology

The partnership leverages Nvidia's NVLink Fusion platform, which enables high-speed data transfers between different types of chips. This technology allows Rebellions' NPUs to operate alongside Nvidia GPUs, creating a more efficient computing system for AI workloads.

The partnership is seen as opening a path for Korean-made NPUs to enter Nvidia's AI computing ecosystem.

Strategic Importance

As demand for AI inference grows, heterogeneous computing systems combining GPUs and NPUs are gaining attention. GPUs excel at large-scale parallel processing, while NPUs are designed for specific AI tasks, potentially improving data center efficiency.

The agreement follows an earlier initiative to integrate Korean NPUs with AMD technology. By expanding cooperation with Nvidia, the South Korean government aims to create more opportunities for domestic AI chip makers in global markets.

Startup Collaboration

The event focused on potential technology partnerships and commercialization opportunities.