
As Nvidia has long remained in hot waters for failing in the timely delivery of its latest Blackwell AI chips, the chipmaker giant's woes have deepened with the chips' servers overheating.
Coming as fuel to the fire, the overheating issue in the latest Nvidia chips left potential buyers to worry about the lack of time it could take to have new, fully functional data centres in place.
Citing sources, Reuters reported that the Blackwell graphics processing units overheat when connected to server racks designed to hold up to 72 chips.
Read more: AI in education — Experts discuss pros, cons at 40th SPELT International Conference 2024
The company has sought from its suppliers the reconfiguration of chip racks to do away with the overheating problem, the international news outlet quoted Nvidia employees working on the issue, as well as customers and suppliers familiar with the matter.
"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," said an Nvidia spokesperson.
The tech giant lifted the veil off its Blackwell AI chips back in March while claiming the new chips will be shipped in Q2 of 2024, and won't face any delays distressing potential customers such as Meta Platforms, Alphabet's Google and Microsoft.