AMD pursues a broader strategic partnership with Samsung to secure advanced memory (HBM) supply, leveraging Samsung's interest in foundry partnerships and memory integration.
AMD Chief Executive Lisa Su is visiting South Korea to secure additional supplies of HBM memory for AI accelerators, nearly seven months after AMD and Samsung signed a memorandum of understanding covering memory supply and collaboration. According to reports, Samsung is seeking to expand the relationship beyond memory into foundry chip production orders.
"We are going to talk about our broader partnership," Lisa Su said when asked by reporters about her planned discussions with Samsung executives. "We have many partners at Samsung."
Before meeting with Samsung's leadership, Su and Jack Huynh, senior vice president and general manager of AMD's Computing and Graphics Group, attended an AMD Gamer Day event in South Korea. There, they characterized local memory makers as "important strategic partners," underscoring the significance of data center-grade DDR5 and HBM memory to AMD's AI and data center businesses.
"Samsung Electronics and SK hynix are key supply-chain partners for AMD," Su said. "AI demand is extremely high, so we are planning our long-term supply chain with our partners. It is important to plan three to five years ahead with our partners."
The non-binding MOU signed in March covers four areas over a three-year period: supply of next-generation memory products including HBM4 and advanced DDR5; co-development and optimization of memory solutions for AI and high-performance computing systems; exploration of foundry manufacturing opportunities for future AMD platforms; and co-optimization of advanced packaging and related semiconductor technologies.
While Samsung has become a principal supplier of HBM4 for AMD's Instinct MI400-series accelerators, expanding collaboration beyond memory is considerably more complicated. It requires designing chips for Samsung Foundry's process technologies and carries risks associated with ramping production at a new contract manufacturer. AMD already uses Samsung Foundry and UMC to produce certain integrated circuits in the form of programmable logic devices, but moving more advanced products to Samsung carries substantial risk. Samsung, meanwhile, badly needs major customers like AMD, and AMD's dependence on HBM supply gives Samsung considerable leverage.
AMD acknowledged these risks in its most recent Annual Report: "If we transition the production of some of our products to new manufacturers, we may experience delayed product introductions, lower yields or poorer performance of our products."
AMD has begun shipments of its Instinct MI455 accelerators with 432GB of HBM4 memory onboard, as well as rack-scale Helios AI systems. "We have started shipping our next-generation MI455 and Helios systems," Su told journalists. "These are all systems that use HBM4."
The company appears focused on securing more memory supply, but the broader question remains whether AMD will accept the risks of adopting Samsung Foundry for more sophisticated products—beyond niche programmable chips—in pursuit of greater HBM supply and higher-margin AI accelerator sales.