The server rack Nvidia designed for Blackwell —which can connect up to 72 GPUs at a time—is reportedly causing the overheating issue. Nvidia has repeatedly redesigned the racks, which could result in GPU server shipments being delayed and new Google, Microsoft, or Meta data centers may not be able to open on schedule.

Nvidia is back in the spotlight this week as it prepares to report earnings. Sales of its Blackwell chips are likely to be the key variable as the market gauges how long the chip maker will dominate the artificial-intelligence hardware market.

Nvidia's next-generation Blackwell processors are facing significant challenges with overheating when installed in high-capacity server racks, reports The Information. These issues have reportedly led to design changes and delays and raised concerns among customers like Google,

The new Blackwell graphics processing units are said to overheat when connected together in customized racks designed to accommodate 72 units, which are intended for building powerful AI systems.

Claims that Blackwell GPUs designed for AI tasks and HPC are overheating come from sources who spoke to The Information.

NVIDIA's (NASDAQ:NVDA) new Blackwell AI chip is said to be having issues with overheating when connected in customized server racks, The Information reported Sunday afternoon. The company is working with suppliers to change

Nvidia is urgently addressing overheating problems in its new Blackwell AI chips when integrated into customized server racks, causing anxiety among major clients like Microsoft and Meta. The company is collaborating

Nvidia Corp. has in recent months asked suppliers to change the design of the server racks for its new Blackwell graphics processing unit due to an overheating problem, leading to worries about delays,