Samsung And AMD Expand Collaboration On Next-Generation AI Memory Solutions

Samsung Electronics and AMD have signed a memorandum of understanding (MOU) to expand their strategic collaboration on AI memory and computing technologies, with the deal covering HBM4 supply for AMD’s upcoming GPU accelerators and DDR5 memory for its next-generation server processors.

The signing ceremony took place at Samsung’s chip manufacturing complex in Pyeongtaek, Korea, attended by AMD Chair and CEO Dr. Lisa Su and Samsung Electronics Vice Chairman and CEO, Young Hyun Jun.

HBM4 For AMD’s Next AI Accelerator

Under the agreement, Samsung will serve as a primary HBM4 supplier for the next-generation AMD Instinct MI455X GPU. The deal also covers advanced DRAM solutions for 6th-Gen AMD EPYC CPUs, codenamed “Venice.”

These technologies will underpin next-generation AI systems combining AMD Instinct GPUs, EPYC CPUs and rack-scale architectures such as the AMD Helios platform.

Samsung’s HBM4 is built on its 6th-generation 10-nanometre-class DRAM process (1c) and a 4nm logic base die. The memory features processing speeds of up to 13 gigabits-per-second and maximum 3.3 terabytes-per-second bandwidth.

Samsung was the first to enter mass production with HBM4.

The MI455X GPU is expected to serve as a building block for the AMD Helios rack-scale architecture, which is designed to deliver the performance and scalability required for large-scale AI infrastructure.

Two Decades Of Partnership

The two companies have collaborated for nearly two decades across graphics, mobile and computing technologies. Samsung currently serves as the primary HBM3E partner to AMD, powering the AMD Instinct MI350X and MI355X AI accelerators.

Young Hyun Jun pointed to the breadth of Samsung’s capabilities in supporting the expanded relationship.

“Samsung and AMD share a commitment to advancing AI computing, and this agreement reflects the growing scope of our collaboration,” Jun stated. “From HBM4 and next-generation memory architectures to foundry and advanced packaging, Samsung is positioned to deliver turnkey capabilities that support AMD’s evolving AI roadmap.”

Dr. Su emphasised the importance of cross-industry collaboration in building out AI infrastructure.

“Powering the next generation of AI infrastructure requires deep collaboration across the industry,” Dr. Su remarked. “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.”

Memory Bandwidth A Growing Bottleneck

As AI models grow larger and infrastructure scales out, memory bandwidth and power efficiency are becoming increasingly important to overall system performance. The collaboration between Samsung and AMD is aimed at delivering more optimised memory solutions for AI training and inference workloads across data centre systems.

As part of the deal, the two companies will also work together on high-performance DDR5 memory optimised for the 6th-Gen EPYC CPUs. The goal is to deliver DDR5 solutions tailored for systems built on the Helios rack-scale architecture.

Foundry Discussions Also On The Table

Beyond memory, Samsung and AMD will also discuss opportunities for a foundry partnership, under which Samsung would provide manufacturing services for next-generation AMD products. No specific details on the scope or timeline of any foundry arrangement were disclosed.

The expanded collaboration comes during a broader visit by Dr. Su to Korea, where AMD has also announced partnerships with Naver Cloud and AI lab Upstage as part of a push to support sovereign AI infrastructure in the country.

Last Updated on March 19, 2026 by Nick Ross