TLDR
- NVIDIA stock plunges 5.09% to $177.14 despite a major South Korean AI deal.
- Rebellions will connect its R200 AI processors through NVIDIA’s NVLink Fusion.
- SK Telecom plans to deploy NVIDIA-linked systems in its Korean AI data centers.
- South Korea backs the partnership to strengthen domestic AI computing systems.
- NVIDIA and Rebellions plan to split inference tasks for greater energy efficiency.
NVIDIA Corporation announced a major partnership with Rebellions and SK Telecom to expand South Korea’s sovereign AI infrastructure. The collaboration combines NVIDIA’s computing platform with Korean semiconductor technology and advanced data center services. NVIDIA (NVDA) shares plunged 5.09% to $177.14, losing $9.50 despite the announcement.
NVIDIA Stock Drops 5% Despite South Korea AI Expansion
NVIDIA shares faced heavy selling pressure, pushing the stock lower during the trading session. The decline followed a retreat from intraday highs near $180, reflecting persistent bearish momentum. Meanwhile, the company’s latest agreement highlights its continued expansion into national AI infrastructure projects.
The partnership brings together NVIDIA’s infrastructure technology, Rebellions’ custom processors, and SK Telecom’s data center operations. South Korea’s Ministry of Science and ICT supports the initiative through its sovereign AI development strategy. The companies aim to build computing systems that support Korean language models and domestic industries.
South Korea has increased its focus on developing domestic AI technology and semiconductor capabilities. Consequently, the collaboration connects local chip development with NVIDIA’s established computing infrastructure and networking systems. The agreement also supports Korea’s broader efforts to strengthen domestic computing capabilities and expand AI services.
NVIDIA NVLink Fusion Connects Rebellions’ Advanced AI Processors
Rebellions plans to integrate its next-generation Rebel R200 neural processing units with NVIDIA’s NVLink Fusion platform. The technology supports high-speed communication between specialized processors and NVIDIA’s existing infrastructure. Furthermore, the integration aims to improve processing efficiency and reduce delays during complex AI workloads.
The R200 processors target large language models and demanding inference applications requiring substantial computing resources. Rebellions will also use NVIDIA’s MGX rack architecture to develop systems for large-scale deployment. This standardized design provides access to established hardware suppliers and supports faster system development.
Additionally, the companies plan to divide AI inference tasks between different processor architectures. NVIDIA GPUs will handle initial processing, while Rebellions NPUs will manage subsequent response generation. This approach aims to increase throughput and improve energy efficiency across demanding workloads.
SK Telecom Expands Sovereign AI Infrastructure With NVIDIA
SK Telecom plans to deploy the integrated computing systems across its AI data center infrastructure. The company will combine these systems with proprietary foundation models, including its A.X K2 model. Moreover, the deployment will support public-facing services under South Korea’s AI for All initiative.
The MGX architecture allows different processors to share standardized rack designs, power systems, and cooling infrastructure. As a result, SK Telecom can coordinate hardware deployment with its existing data center operations. The company also plans to connect domestic models with scalable computing resources.
Meanwhile, South Korea’s government intends to strengthen cooperation with NVIDIA across additional technology projects. These efforts include physical AI development and broader access to GPU computing resources. The partnership expands NVIDIA’s role in Korea’s AI infrastructure despite the stock’s latest decline.
