Nvidia's (NVDA.US) new Blackwell chip is delayed again, and overheating issues have been reported.
1min read According to insiders, Nvidia's (NVDA.US) new Blackwell AI chips have serious overheating issues in high-capacity server racks. These problems have led to design adjustments and project delays, causing major customers such as Google, Meta and Microsoft to worry about whether they can deploy Blackwell servers as planned.
The insiders said that the Blackwell GPU, designed for AI and high-performance computing (HPC), has overheating issues in servers with 72 processors, which can consume up to 120 kilowatts of power per rack. The overheating forced Nvidia to make multiple changes to the rack design, not only limiting GPU performance but also potentially damaging hardware. This has caused customers to worry that these technical issues will delay the deployment of processors in data centers.
To address this challenge, Nvidia has asked suppliers to adjust the rack design and worked with partners to optimize the cooling system. Although this engineering improvement is a regular step in large-scale technology releases, it has further delayed product delivery.
Nvidia spokesperson responded to the matter by saying that the company is working closely with cloud service providers and that design adjustments are a normal part of the R&D process. Nvidia hopes to ensure that the final product meets performance and reliability expectations through this collaboration while addressing technical bottlenecks.
The revised Blackwell GPU entered mass production in late October and is expected to be shipped as early as the end of January. The delay affects the research and development plans and product releases of tech giants such as Google, Meta, and Microsoft, which rely on Nvidia GPUs to train their most powerful AI models.
Nvidia showcased the Blackwell chip in March this year and said it would ship in the second quarter.