Samsung to Begin Commercial Shipments of HBM4 Memory in February, Beats Competitors

• February 9, 2026

Samsung Electronics is set to commence commercial shipments of its latest high-bandwidth memory, HBM4, as early as mid-February. This move positions Samsung as the first company to bring HBM4 technology to the commercial market, outpacing other memory manufacturers in the industry.

Partnership with Nvidia to Power Next-Generation AI Platforms

The initial batch of HBM4 chips will primarily be provided to Nvidia, which intends to integrate this advanced memory in its upcoming AI-focused computing platform codenamed Vera Rubin. This platform targets the development of cutting-edge AI supercomputers, emphasizing high performance and efficiency.

Industry observers anticipate that Nvidia will officially unveil graphics accelerators utilizing this new HBM4 memory technology at the GPU Technology Conference (GTC) scheduled for March 2026. These accelerators are expected to enhance computational capabilities significantly for AI workloads, leveraging the increased bandwidth and efficiency offered by HBM4.

Samsung’s achievement in advancing HBM4 to commercial availability marks a critical milestone within the semiconductor sector, particularly for applications requiring massive data throughput and energy-efficient memory solutions. These developments underscore the ongoing race among component manufacturers to supply next-generation memory to leading AI and high-performance computing firms.

The broader impact of Samsung’s advancement is likely to influence the competitive dynamics around memory technology, especially with Nvidia’s adoption signaling strong demand and validation for HBM4 in sophisticated AI infrastructures.

Details regarding pricing, shipment volumes, and further partnerships related to HBM4 have not been disclosed at this time. Attention will now turn to the upcoming GTC 2026 event for additional announcements and technical disclosures from Nvidia regarding the deployment of HBM4-enabled accelerators.

Samsung will start delivering HBM4 memory chips commercially in February, supplying Nvidia for its next-gen AI supercomputing platform.