- MSIT signs an MoU with global AI semiconductor company AMD
- Comprehensive cooperation will be pursued to develop and demonstrate heterogeneous AI computing infrastructure; conduct joint research on “AI for Science” with the National AI for Science Research Center (NAIS); establish an “AMD AI Center of Excellence” in Korea; and foster AI talent and conduct joint research on physical AI
【Related National Policy Agenda】 22. Securing Leadership in AI Technologies and Talent

​

The Ministry of Science and ICT (MSIT; Deputy Prime Minister and Minister Bae Kyunghoon) announced that it signed a Memorandum of Understanding (MoU) with AMD Chair and CEO Lisa Su on cooperation in the AI semiconductor ecosystem at AMD Advancing AI 2026, held in San Francisco, United States, at 3:00 p.m. on Thursday, July 23 (U.S. Pacific Time). The MoU aims to strengthen cooperation in AI semiconductors and computing.

​

AMD is a global semiconductor company with a portfolio spanning central processing units (CPUs) and graphics processing units (GPUs). The company recorded annual revenue of $34.6 billion in 2025. Driven in particular by growing demand for server CPUs and GPUs, its data center business generated $16.6 billion in revenue, cementing AMD’s position as a major player in the AI computing market.

​

AMD is working to expand the AI computing ecosystem through its open-source software platform ROCm* and open standards. The company is also actively helping build an open AI infrastructure ecosystem, including as a founding member of UALink**, an open standard for high-speed interconnection between AI accelerators, alongside a number of global companies such as Google, Apple, and Microsoft.

* ROCm (Radeon Open Compute): AMD’s open-source-based software platform for AI and high-performance computing, designed to support GPU-based AI development and application implementation.
* UALink (Ultra Accelerator Link): An open interconnect standard for high-speed data transfer and connectivity between AI accelerators, enabling efficient connections among different types of computing devices within large-scale AI systems.

​

The recent global AI semiconductor market is rapidly shifting away from computing architectures centered on a single GPU toward heterogeneous computing architectures* that combine different types of computing devices—including CPUs, GPUs, and NPUs—to suit different service requirements and improve cost efficiency. This shift is being driven by rising inference demand as generative AI and agentic AI become more widespread. It reflects a broader change in AI infrastructure competition toward using the optimal computing resources for each workload in order to reduce the cost of delivering AI services while optimizing performance and power efficiency.

* Heterogeneous computing architecture: An infrastructure technology that combines and optimally configures different types of AI semiconductors, such as CPUs, GPUs, and NPUs, based on computational characteristics, rather than relying solely on GPUs, to reduce AI service delivery costs and improve power efficiency.

​

The MoU was signed in response to these changes in the global AI computing architecture. It aims to build heterogeneous AI computing infrastructure that integrates AMD CPUs and GPUs with Korean AI semiconductors (NPUs) that have strengths in inference, while expanding global collaboration for Korea's AI semiconductor ecosystem.

​

First, the two sides will work to develop and demonstrate heterogeneous AI computing infrastructure integrating AMD CPUs and GPUs with Korean NPUs. Through this effort, an integrated computing environment will be created that enables diverse computing resources to be efficiently combined and utilized according to the characteristics of each AI service, while an open AI computing ecosystem will be established in which Korean companies across related sectors—including not only AI semiconductor companies but also those specializing in memory, DPUs, CXL, networking, servers, and orchestration software—can participate together. The collaboration will also serve as a basis for developing practical reference models to facilitate the integration of Korean AI semiconductors into the global AI computing supply chain.

​

The two sides will also pursue cooperation on AI for Science research, centered on the National AI for Science Research Center (NAIS). To support AI research in advanced science and technology fields as well as ambitious research and development (R&D), AMD plans to provide high-performance computing resources, including CPUs and GPUs, along with AI software and technical experts.

​

In particular, AMD plans to establish an AI Center of Excellence in Korea. The AI Center of Excellence will serve as a cooperation hub supporting technological cooperation among Korean companies, universities, and research institutions, AI semiconductor software development, computing technology validation, and joint research based on AMD’s AI computing resources and software capabilities. Through this, AMD’s global network and Korea’s industry-academia-research capabilities will be connected to create various cooperation opportunities based on an open AI ecosystem.

​

The two sides will also cooperate in fostering outstanding AI talent. AMD plans to support education and research programs based on AI computing platforms and computing resources at major universities and graduate schools in Korea, while expanding opportunities for Korean students, professors, and researchers to experience the latest AI technologies and global development environments. They also plan to promote joint research and technological cooperation in next-generation AI, such as physical AI, and thereby expanding the foundation for industry-academia-research cooperation to secure future AI technological competitiveness.

​

The cooperation between the MSIT and AMD gained momentum following AMD CEO Lisa Su’s visit to Korea in March, and the two sides have since specified areas of cooperation and implementation plans through close consultations. President Lee Jae Myung’s visit to the United States provided opportunities for concretizing practical cooperation projects in areas such as AI computing infrastructure, AI research, and talent development, leading to the signing the MoU. In particular, the signing ceremony was held at the invitation of Lisa Su, providing another opportunity to reaffirm the two sides’ commitment to strategic cooperation in addressing changes in the global AI ecosystem.

​

Deputy Prime Minister Bae Kyunghoon said, “Global AI competition is changing toward heterogeneous computing that combines various computing resources such as CPUs, GPUs, and NPUs, and toward open ecosystems. To enable Korea to advance as one of the world’s top three AI powers, it is important to secure an independent yet open foundation for AI computing infrastructure. Through cooperation with AMD, we will be able to enhance the competitiveness of Korea’s AI semiconductor ecosystem and create a new cooperation model that addresses changes in global AI computing infrastructure.”

​

Lisa Su, President and CEO of AMD, said, “Through cooperation with the Ministry of Science and ICT of the Republic of Korea, AMD is bringing together its capabilities across high-performance computing technologies, open software, and global ecosystem partnerships to support Korea to build its future AI infrastructure. We will accelerate Korea’s AI innovation and economic growth simultaneously through close technological cooperation, AI talent development, and the establishment of an open AI ecosystem foundation.”

​

The two sides will create a joint working group for the practical implementation of this agreement and will continue discussing detailed cooperation agenda and implementation plans through quarterly working-level meetings.

​

For further information, please contact the Public Relations Division (Phone: +82-44-202-4034, E-mail: msitmedia@korea.kr) of the Ministry of Science and ICT.

​

Please refer to the attached PDF.