NVIDIA considers reducing Rubin Ultra GPU configuration to alleviate HBM supply shortage
- Nvidia is considering adjusting the design of its next-generation Rubin Ultra GPU.
- The adjustment involves reducing high-bandwidth memory (HBM) configurations.
- This decision is in response to the tight supply of advanced HBM chips.
- Nvidia has tested at least three different versions of the Rubin Ultra GPU in recent weeks.
- Some versions feature lower memory capacities than previously announced plans.
According to The Information, Nvidia is considering adjusting the design of its next-generation Rubin Ultra GPU by reducing high-bandwidth memory (HBM) configurations to address the tight supply of advanced HBM chips.
People familiar with the matter disclosed that over the past few weeks, Nvidia has tested at least three different versions of the Rubin Ultra GPU, some of which feature memory capacities lower than the company's previously announced plans.
Nvidia is considering launching a lower-memory version mainly because the market may not be able to obtain sufficient quantities of advanced HBM chips.
HBM is a key component of AI accelerators, delivering higher data transfer speeds and serving as a critical foundation for training and inferencing large AI models. As demand for AI computing power surges, HBM supply has become one of the major factors limiting the production capacity of high-end AI chips.
NVIDIA考虑降低Rubin Ultra GPU配置以缓解HBM供应短缺
根据《信息》报道，Nvidia正在考虑通过减少高带宽内存（HBM）配置来调整其下一代Rubin Ultra GPU的设计，以应对先进HBM芯片的供应紧张。
知情人士透露，Nvidia在最近几周测试了至少三种不同版本的Rubin Ultra GPU，其中一些版本的内存容量低于公司之前公布的计划。
Nvidia考虑推出低内存版本，主要是因为市场可能无法获得足够数量的先进HBM芯片。
HBM是AI加速器的关键组件，提供更高的数据传输速度，并作为训练和推理大型AI模型的关键基础。随着对AI计算能力的需求激增，HBM供应已成为限制高端AI芯片生产能力的主要因素之一。