据Digitimes报道，全球领先的图形处理单元（GPU）制造商英伟达的H100 AI GPU供应紧张情况出现明显缓解。台积电台湾地区总经理Terence Liao透露，H100 GPU的交货等待时间已从之前的3-4个月大幅缩短至目前的2-3个月，即8-12周。这一积极变化表明，英伟达与台积电之间的生产合作可能已取得显著进展。
服务器代工厂商也证实了这一情况，他们表示，与2023年初几乎无法获取英伟达H100 GPU相比，当前的供应瓶颈正在逐步解除。这一进展对依赖H100进行高性能计算和人工智能应用的数据中心和企业来说，无疑是一个重大利好。英伟达的H100 GPU因其在机器学习、深度学习和大数据处理领域的卓越性能，一直备受业界瞩目。
随着交货时间的缩短，市场预期英伟达的客户将能够更快地获取到这款高性能GPU，从而加速其在AI领域的研发和部署。然而，尽管供应状况有所改善，是否能完全满足市场需求仍有待观察。英伟达和台积电将继续合作，以确保产能的稳定提升，以满足全球不断增长的AI计算需求。
英语如下：
News Title: “NVIDIA H100 AI GPU Shortage Eases: Delivery Times Cut to 2-3 Months”
Keywords: NVIDIA H100, GPU shortage, shortened delivery times
News Content:
Title: NVIDIA H100 AI GPU Shortage Significantly Improves, Delivery Times Reduced to 2-3 Months
According to Digitimes, the global leading graphics processing unit (GPU) manufacturer NVIDIA has seen a notable easing of its H100 AI GPU supply crunch. TSMC’s Taiwan General Manager, Terence Liao, disclosed that the wait time for H100 GPU deliveries has dropped from the previous 3-4 months to the current 2-3 months, or 8-12 weeks. This positive development suggests that NVIDIA’s production collaboration with TSMC may have made significant progress.
Server original design manufacturers (ODMs) have also confirmed this improvement, stating that compared to the almost unavailability of NVIDIA H100 GPUs at the start of 2023, the current supply bottleneck is gradually easing. This progress is a major boost for data centers and enterprises relying on the H100 for high-performance computing and AI applications. NVIDIA’s H100 GPU, renowned for its exceptional performance in machine learning, deep learning, and big data processing, has consistently garnered industry attention.
With shortened delivery times, the market expects NVIDIA’s customers to gain access to this high-performance GPU more swiftly, thereby accelerating their AI research and deployment. However, it remains to be seen whether the improved supply can fully meet the market demand. NVIDIA and TSMC will continue their collaboration to ensure stable capacity increases to cater to the globe’s growing AI computing requirements.
【来源】https://www.ithome.com/0/761/155.htm
Views: 1