Samsung's 12-layer HBM4E chip with 48GB capacity began shipping to customers. The move pressures SK Hynix and Micron as Nvidia and Google await next-gen memory.
Samsung Electronics shares surged as much as 6.51% after the company said it has begun shipping samples of its 12-layer HBM4E memory chip, a next-generation high-bandwidth memory component designed for AI accelerators. The move directly challenges SK Hynix and Micron in the premium AI memory segment and gives Nvidia and Google a new supplier option for upcoming processors.
The HBM4E chip is the first 12-layer stack in the industry, offering 48GB capacity – a more than 30% increase over the prior generation – with data transfer speeds up to 16 Gbps. Samsung said the chip also delivers improved energy efficiency and thermal performance. The company plans to expand the lineup to include 8-layer 32GB and 16-layer 64GB configurations depending on customer demand.
Samsung began shipping HBM4 chips to customers in February and is now stepping up with HBM4E samples. “Through our advanced manufacturing capabilities and preemptive infrastructure investments, we will continue to drive the growth of the global AI memory market,” said Sang Joon Hwang, executive vice president and head of memory development. The comment underscores Samsung's intent to close the gap with SK Hynix, which has led the HBM market in recent quarters.
The immediate read-through is for SK Hynix and Micron, which have dominated HBM supply to Nvidia. Samsung's HBM4E sample shipments raise the possibility of a three-way race for next-generation AI memory contracts. Nvidia's upcoming Rubin architecture and Google's Ironwood TPU are expected to require HBM4-class memory, making Samsung's entry a potential shift in supplier dynamics. AlphaScala data shows NVDA currently trading at $214.25, up 0.78% on today's session, with a Moderate Alpha Score of 71. The stock's modest move suggests investors are still weighing the timeline to volume production.
Samsung's stated plan to offer multiple layer configurations (8H, 12H, 16H) gives it flexibility to serve different tiers of AI workloads. Mass production timelines remain the key uncertainty. SK Hynix currently leads in HBM4 qualification with Nvidia, and Samsung will need to deliver volume before its samples translate into revenue.
The next decision point is Samsung's volume production ramp for HBM4E and its qualification status with Nvidia. If Samsung achieves parity with SK Hynix on yield and delivery schedules, the AI memory sector could see a rebalancing of pricing and market share. For investors tracking the space, the watchlist items are Samsung's future announcements on customer wins and the HBM4E production timeline.
For more context on Nvidia's position, see the NVDA stock page and NVIDIA profile.
Prepared with AlphaScala research tooling and grounded in primary market data: live prices, fundamentals, SEC filings, hedge-fund holdings, and insider activity. Each story is checked against AlphaScala publishing rules before release. Educational coverage, not personalized advice.