AMD CEO visits Korea for AI partnerships

Following Nvidia’s lead in the global AI semiconductor market, AMD Chief Executive Officer Lisa Su visited South Korea to deepen partnerships with Samsung and Naver. Her trip comes as a series of high-profile visits by global AI leaders, including Nvidia CEO Jensen Huang and OpenAI CEO Sam Altman, underscores Korea’s rising role as a key hub in the intensifying global AI race.

According to industry sources on March 18, Su visited Samsung Electronics’ Pyeongtaek campus and signed a memorandum of understanding with senior executives, including Vice Chairman and Device Solutions Division Head Jeon Young-hyun, to expand cooperation in next-generation AI memory and computing technologies.

Under the agreement, Samsung will serve as the primary supplier of high-bandwidth memory HBM4 for AMD’s next-generation accelerator, the Instinct MI455X. The sixth-generation HBM4 is also expected to power Nvidia’s newly unveiled AI accelerator, Vera Rubin. The two companies are also understood to have discussed a broader partnership extending beyond memory supply to include advanced foundry production and packaging services.

Jeon said the agreement would expand the scope of cooperation between the two companies, emphasizing that Samsung has the full range of capabilities needed to support AMD’s AI roadmap, from HBM4 to cutting-edge foundry and packaging technologies. Su welcomed the partnership, saying she was pleased to combine Samsung’s leadership in advanced memory with AMD’s platform. She is scheduled to meet Samsung Electronics Chairman Lee Jae-yong for a dinner at Seungjiwon on March 18 and to meet DX Division Head Noh Tae-moon on March 19.

Su’s visit comes amid intensifying competition with Nvidia CEO Huang. Around the same time, Nvidia hosted its annual GTC 2026 developer conference in the United States and announced that Samsung would manufacture its inference-focused AI chips. Huang visited Korea in October, meeting Lee and Hyundai Motor Group Chairman Chung Eui-sun to reinforce ties with Korean companies.

Analysts say the wave of visits by global AI executives reflects the growing urgency of securing a stable supply of HBM, a critical component for AI accelerators. With Korean chipmakers offering end-to-end capabilities from design to production, and with memory supply tightening, their strategic importance is expected to grow further.

Beyond hardware, AMD is also moving to expand its software ecosystem through partnerships with Korean firms, including Naver and Upstage. Earlier in the day, Su visited Naver’s second headquarters, “1784,” in Seongnam and signed an MOU with CEO Choi Soo-yeon and other executives to cooperate on AI ecosystem expansion and next-generation infrastructure. The partnership aims to optimize high-performance GPU environments for Naver’s large language model, HyperCLOVA X, while strengthening the infrastructure needed for stable AI deployment.

Choi said the company would broaden the use of AMD’s platform across Naver Cloud and its AI services, while continuing cooperation on next-generation technology stacks and services. Su said she looks forward to building an open AI ecosystem together that researchers and developers around the world can trust.


이동훈 dhlee@donga.com