Nvidia's (NVDA, Financials) latest Blackwell AI chips, launched earlier this year, are experiencing overheating issues in server configurations, according to a report by The Information.
As they are ready to develop data centers depending on the technology, these problems might affect big cloud service consumers as Meta Platforms (META, Financials), Alphabet's Google (GOOGL, Financials), Microsoft (MSFT, Financials).
The overheating issue arises from chips placed in server racks meant to support up to 72 units. To solve these problems, Nvidia apparently has requested vendors to repeatedly rebuild the racks. Citing Nvidia staff, suppliers, and consumers, the research revealed that overheating issues still exist.
A spokesperson for Nvidia stated, "Nvidia is working with leading cloud service providers as an integral part of our engineering team and process. The engineering iterations are normal and expected."
Originally revealed in March, the Blackwell chips were supposed to arrive in the second quarter. Delays, however, pushed delayed delivery, therefore affecting client deployment schedules. The chips combine two silicon components into one unit meant to run 30 times quicker than past versions in jobs like chatbot answer creation.
With its products essential to key technological platforms, this story tracks Nvidia's attempts to establish its leadership in artificial intelligence and cloud computing. The overheating problem, however, might cause questions over whether the rising needs of data-intensive AI projects can be satisfied.
Nvidia has not revealed which vendors help to remedy the design problems or when the overheating concerns could be fixed. The delays could influence the larger AI infrastructure plans of important customers.
