Lisa Su, the dynamic CEO of AMD, made headlines with her much-anticipated first official visit to South Korea since taking the helm in 2014. Over a whirlwind two-day schedule beginning March 18, 2026, Su’s agenda was packed with high-stakes meetings, ambitious partnership signings, and strategic discussions designed to reshape the AI and semiconductor landscape in the region. Her trip, as reported by multiple Korean outlets including Chosun Biz, Hankyung, and The Korea Economic Daily, signals AMD’s intent to challenge Nvidia’s dominance and deepen ties with Korea’s tech giants.
Su’s first stop was Naver’s iconic ‘1784’ headquarters in Seongnam, Gyeonggi Province. Arriving at 11:12 a.m., she was greeted by Naver CEO Choi Soo-yeon. The two leaders wasted no time, diving into talks about expanding cooperation between their companies. The highlight of the morning was the signing of a memorandum of understanding (MOU) for AI ecosystem expansion and next-generation infrastructure cooperation, as confirmed by Kyunghyang Shinmun. The MOU focuses on building a high-performance GPU computing environment optimized for Naver’s large language model, HyperCLOVA X, and aims to foster a flexible, diverse AI infrastructure ecosystem.
“We will discuss more today,” Su told reporters, hinting at the depth of the ongoing negotiations. She emphasized her excitement about visiting Korea and the potential for deepening collaboration, stating, “How can we become even closer? There’s a lot to discuss with Naver and also with Samsung Electronics’ foundry.”
The MOU is significant for both companies. Naver, which recently completed the country’s largest AI computing cluster using 4,000 units of Nvidia’s next-generation B200 GPU, is seeking to diversify its suppliers to reduce dependency. According to The Elec, Naver’s plans to expand its Sejong data center—boosting capacity to 135 MW by 2029—will require robust and flexible hardware solutions. By partnering with AMD, Naver gains access to EPYC CPUs and Instinct GPUs, including the current MI350 series and upcoming MI400 and MI450 models. This move, as industry insiders told Yonhap Infomax, is a strategic effort to stabilize GPU supply and ensure the company’s AI ambitions aren’t tethered to a single vendor.
For AMD, the partnership with Naver is a strategic coup. The Korean tech giant is one of the largest domestic consumers of GPUs, making it a coveted customer as AMD seeks to expand its footprint in the AI data center market. With Nvidia holding about 90% of the global data center AI GPU market share, AMD’s recent wins—like onboarding OpenAI and Meta as clients—are part of a broader campaign to challenge Nvidia’s dominance. As one industry observer put it to Hankyung, “This is a sales drive to catch up with Nvidia, the number one in the AI semiconductor market.”
The scope of the AMD-Naver partnership is broad. Beyond infrastructure, the two companies plan to provide AI computing resources to academic researchers and collaborate on joint projects, fostering a more vibrant AI research ecosystem in Korea. As Naver’s CEO Choi Soo-yeon put it, “The cooperation with AMD is a meaningful milestone for securing technological diversity and enhancing AI infrastructure competitiveness.” Su, for her part, called Naver “the optimal partner to implement AMD’s next-generation AI GPU technology.”
After a productive morning at Naver, Su’s schedule shifted to Samsung Electronics’ Pyeongtaek campus. There, she met with senior executives including Jeon Young-hyun, Vice Chairman of the DS division, and Han Jin-man, head of the foundry business. The focus: expanding collaboration beyond memory to include foundry services and securing a stable supply of high-bandwidth memory (HBM4), a critical component for AMD’s forthcoming Instinct MI450 GPU accelerators.
Samsung, already a supplier of general DRAM and cutting-edge HBM3E to AMD, has begun mass production of 6th generation HBM4, which boasts a data transfer speed of up to 13 Gbps—an industry-leading figure. Discussions, as reported by Seoul Economic Daily, centered on supplying HBM4 for AMD’s next-generation accelerators and potentially manufacturing AMD’s chips using Samsung’s advanced 2nm SF2P process. Such a move could see Samsung’s foundry business further bolstered, especially as the company also produces AI chips for Tesla and Nvidia.
The day concluded with a high-profile dinner attended by Samsung Electronics Chairman Lee Jae-yong at Seungjiwon in Hannam-dong, Seoul—a venue steeped in corporate symbolism. This meeting signaled the importance both companies attach to their burgeoning alliance, which now spans memory, foundry, and AI solutions for everything from PCs to premium mobile devices.
Su’s Korean itinerary didn’t end there. On March 19, she was slated to meet Noh Tae-moon, head of Samsung’s DX division, to explore further cooperation on AI PCs and mobile products. She also planned a visit to Upstage, a homegrown AI startup recognized for its advanced large language model SOLAR, which ranks highly on Hugging Face’s global leaderboard. Upstage, which currently uses Nvidia GPUs, is seen by AMD as a potential customer for its EPYC CPUs and Instinct GPUs. Su’s visit was expected to include a pitch for AMD’s data center products, further signaling her intent to expand AMD’s influence among Korea’s most promising AI companies.
Industry watchers noted that Su’s visit marks a pivotal moment for AMD in Asia. Not only is the company seeking to chip away at Nvidia’s market lead, but it’s also forging multi-faceted alliances that span hardware, software, and research. The timing is critical: Naver is preparing to commercialize AI at scale in 2026, and Samsung is ramping up AI chip production for global clients. By positioning AMD as a flexible, innovative partner, Su is betting on Korea’s tech ecosystem to help drive the next wave of AI infrastructure—and, perhaps, to rewrite the balance of power in the global semiconductor industry.
As Su wrapped up her meetings and prepared for possible onward travel to neighboring countries like Japan, the message was clear: AMD is no longer content to play second fiddle. With strategic partnerships, bold investments, and an eye on the future, Lisa Su’s Korean visit may well be remembered as a turning point in the AI hardware race.