Samsung Electronics’ fourth-generation high bandwidth memory (HBM3) chips have received Nvidia’s approval for use in its processors for the first time. However, this approval is limited to the less advanced H20 graphics processing unit (GPU) developed for the Chinese market under U.S. export controls.
It remains unclear whether Nvidia will use Samsung’s HBM3 chips in other AI processors or if further testing is required. Samsung is still undergoing testing to meet Nvidia’s standards for its fifth-generation HBM3E chips.
High bandwidth memory, a DRAM standard introduced in 2013, is essential for GPUs in processing large data sets for AI applications. Nvidia’s approval of Samsung’s HBM3 chips comes amid high demand for advanced GPUs driven by the AI boom.
With only three major HBM manufacturers—SK Hynix, Micron, and Samsung—Nvidia is eager for Samsung to meet its standards to diversify its supplier base. As SK Hynix plans to shift more production to HBM3E, Nvidia’s need for HBM3 will likely increase.
Samsung is expected to start supplying HBM3 for Nvidia’s H20 processor as early as August. The H20 is a China-specific GPU with capped computing power to comply with U.S. sanctions. Despite a slow start, sales of the H20 are now growing rapidly.