Via techweez.com
The semiconductor giant's HBM4 chips are already shipping to AMD, with NVIDIA partnerships in the pipeline and a 12-layer HBM4E variant pushing the boundaries of what AI hardware can do.
Samsung Electronics has crossed the $1 billion sales mark for its next-generation AI memory chips, a milestone that took roughly four months from the start of mass production. The company’s high-bandwidth memory (HBM) 4 chips, which began shipping on February 12, 2026, claim performance improvements exceeding 40% over previous standards.
The HBM4 rollout and the AMD deal #
Mass production kicked off in February, and by March 18, the company had locked down a memorandum of understanding with AMD to supply HBM4 chips for the Instinct MI455X GPU and EPYC processors.
The partnership targets bandwidths of up to 3.3TB/s, achieved through a 10nm-class manufacturing process.
Samsung and NVIDIA held discussions at the GTC 2026 event focused on HBM4E integration and future HBM5 architecture.
HBM4E: the 12-layer monster #
By May 2026, Samsung began shipping samples of its 12-layer HBM4E, which the company describes as an industry first. The chips reach speeds of up to 16Gbps per pin, a specification designed to maximize power efficiency for AI workloads.
What this means for the AI hardware supply chain #
The $1 billion sales figure, reached by June 23, 2026, tells a clear story about demand. Samsung’s aggressive timeline, from mass production in February to billion-dollar revenue by June, mirrors the urgency across the entire AI supply chain.
SK Hynix has been the dominant HBM supplier to NVIDIA in recent generations, and Micron has been pushing its own advanced memory solutions. Samsung’s HBM4 launch, and particularly the HBM4E samples, represent a serious attempt to reclaim market share in what has become the most strategically important segment of the memory industry.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our