Huawei is reportedly developing a memory for AI-specific workloads that is claimed to replace HBM with effective capacities and efficiency.
Huawei's AI SSD Might Come With No Capacity Limitations, Offering a Viable Alternative to HBM
When you look at China's AI hardware advancements, it seems that the region is held back by a lack of 'capable' high-bandwidth memory solutions. Chinese firms like CXMT have made notable developments; however, there still seems to be a gap with the memory solutions the West has come up with. Now, according to Chinese media sources, Huawei is reportedly developing a dedicated AI memory solution, which is claimed to be a solid-state drive but tuned for datacenters.
While the specifics of the solution haven't been revealed yet, Huawei's advancement is claimed to break the domestic industry's reliance on HBM, given that the supply of the memory solution has seen massive restrictions driven by geopolitical factors. The report states that Huawei's AI SSD won't have capacity barriers and will significantly boost computational capabilities. However, the implementation still hasn't been shared yet, so we should take these claims with a grain of salt for now.
In the realm of AI memory storage, Huawei has also unveiled a UCM (Unified Cache Manager) software suite that speeds up LLM training across HBM, standard DRAM, and SSDs. This allows the Chinese firm to expand memory usage for AI-related workloads, without introducing capable hardware, which is a potential workaround to the limitations faced with HBM. This shows that the firm has been trying to find solutions with the restrictions on the company, and it seems like an AI SSD might be the firm's next target.
Huawei has expanded its AI compute arsenal at a rapid pacewhen it comes to inferencing or training hardware. While the company is still behind the likes of NVIDIA, the efforts being made are indeed narrowing the gap.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
