South Korean tech giant Samsung has teamed up with American chipmaker AMD to develop a range of new artificial intelligence‑based memory and computing solutions.
Top leadership from both companies recently gathered at Samsung’s advanced chip manufacturing facility in Pyeongtaek, Korea, to sign a memorandum of understanding and formally cement their collaboration.
Both Young Hyun Jun, vice chairman and chief executive officer of Samsung Electronics, and Dr Lisa Su, chair and chief executive of AMD, were in attendance to announce their firms’ new AI partnership.
Under the agreement, Samsung will supply AMD with its HBM4 chip to be used in a range of AMD AI products, including the AMD AI accelerator, the AMD Instinct MI455X GPU and the DRAM powering the 6th Gen AMD EPYC CPUs.
Paired with its Instinct GPUs, AMD EPYC CPUs and AMD Helios rack‑scale architecture, these various AI innovations from AMD will form the basis of computing systems powered by AI technology.
The news comes as AMD this week unveiled its vision for a new computing category called agent computers. The chipmaker expects personal computers, which are currently controlled by humans using a mouse and keyboard, to be eventually replaced by systems run almost entirely by AI agents.
Dr Lisa Su said: “Powering the next generation of AI infrastructure requires deep collaboration across the industry.
“We are thrilled to expand our work with Samsung, bringing together their leadership in advanced memory with our Instinct GPUs, EPYC CPUs and rack‑scale platforms. Integration across the full computing stack, from silicon to system to rack, is essential to accelerating AI innovation that translates into real‑world impact at scale.”
Samsung’s HBM4 will be a key ingredient in AMD’s ambitions to transform traditional computing with AI. Although Samsung is not the only firm to create an HBM4 chip – a technology designed to accelerate data transfer speeds for AI and other areas of advanced computing – it is widely seen as a commercial leader in the area, with its HBM4 currently in mass production.
The firm produces its HBM4 chip using a 6th‑generation 10‑nanometre‑class DRAM process. By equipping it with a 4nm logic base die, Samsung has made it possible for the chip to deliver record 13 gigabits‑per‑second processing speeds and 3.3 terabytes‑per‑second bandwidth.
According to Samsung and AMD, their work together will result in the memory and power needed to sustain the rapid rise of AI technology and the increasing workload of data centres responsible for facilitating this technological shift.
Their partnership is likely to be a long‑term one as well. Both firms have committed to exploring ways in which they can work together on the development of future AMD products by leveraging Samsung’s foundry expertise.
Young Hyun Jun said: “Samsung and AMD share a commitment to advancing AI computing, and this agreement reflects the growing scope of our collaboration.
“From industry‑leading HBM4 and next‑generation memory architectures to cutting‑edge foundry and advanced packaging, Samsung is uniquely positioned to deliver unrivalled turnkey capabilities that support AMD’s evolving AI roadmap.”
