(Reuters) – Nvidia’s new Blackwell AI chips, which have already confronted delays, have run into issues with their accompanying servers overheating, inflicting some clients to fret they will not have sufficient time to get the brand new knowledge facilities up and operating, Information reported Sunday. .
Blackwell graphics processing models overheat when plugged collectively in customized server racks designed by the corporate, the report mentioned, citing sources conversant in the problem.
The AI chipmaker has repeatedly requested its suppliers to vary rack designs to handle overheating issues, in accordance with Nvidia workers who’ve labored on the problem, in addition to clients and suppliers with information of it, it mentioned. within the report with out naming the producer. suppliers.
Nvidia didn’t instantly reply to a request for remark exterior of regular enterprise hours.
Nvidia unveiled the Blackwell chips in March and beforehand mentioned they might ship within the second quarter earlier than encountering delays, which may have an effect on clients together with Meta Platforms, Alphabet’s Google and Microsoft.
Nvidia’s Blackwell chip takes two squares of silicon the dimensions of the corporate’s earlier providing and combines them right into a single part that is 30 occasions sooner at duties like offering responses from chatbots.
(Reporting by Gursimran Kaur in Bangalore; Editing by Bill Berkrot)
