NVIDIA Corporation's (NVDA.US) new generation Blackwell chip delayed delivery again, reported to have overheating issues.
18/11/2024
GMT Eight
According to informed sources, the new generation Blackwell AI chip from NVIDIA Corporation (NVDA.US) has serious overheating issues in high-capacity server racks. These issues have led to design adjustments and project delays, causing major customers such as Alphabet Inc. Class C, Meta, and Microsoft Corporation to be concerned about whether they can deploy Blackwell servers as planned.
Informed sources revealed that the Blackwell GPU is designed specifically for AI and high-performance computing (HPC), but is experiencing overheating problems in servers with 72 processors, where each rack can consume up to 120 kilowatts of power. The overheating issues have forced NVIDIA Corporation to make multiple modifications to rack designs, limiting GPU performance and potentially causing hardware damage. Customers are therefore worried that these technical issues will delay the deployment of processors in data centers.
To address this issue, NVIDIA Corporation has requested suppliers to adjust rack designs and is working with partners to optimize the cooling system. Although such engineering improvements are a routine step in large-scale technology releases, they have further delayed product delivery.
A spokesperson for NVIDIA Corporation responded to the matter by stating that the company is closely collaborating with cloud service providers, and design adjustments are part of the normal R&D process. NVIDIA Corporation hopes that through this collaboration, the final product will meet expectations in terms of performance and reliability, while also accelerating the resolution of technical bottlenecks.
The revised Blackwell GPU only entered mass production at the end of October this year, with the earliest shipment expected by the end of January next year. Tech giants such as Alphabet Inc. Class C, Meta, and Microsoft Corporation rely on NVIDIA Corporation GPUs to train their most powerful AI models, so the delay will inevitably have an impact on these customers' research and product release plans.
In March of this year, NVIDIA Corporation showcased the Blackwell chip, stating at the time that shipments would begin in the second quarter.