"NVIDIA AI Chip 'Blackwell' Server Overheating Problem"..."Normal Engineering Process"

Nvidia's next-generation AI chip 'Blackwell' reportedly experienced server overheating issues ahead of its launch.

Citing sources, U.S. IT media Deformation reported that Nvidia's Blackwell was overheated when it was connected to a custom-designed server rack.

"NVIDIA has repeatedly asked suppliers to change the design of server racks to fix the problem," Deformation quoted an Nvidia employee as saying.

Nvidia responded, "We continue to work with Nvidia's engineering team and cloud service providers," adding, "It is normal and expected to repeat engineering."

Nvidia, which first unveiled Blackwell in March, originally planned to launch it in the second quarter of this year but planned to mass-produce it in the fourth quarter of this year after defects were found in its own production.

If the problem of overheating the Blackwell server turns out to be true, it will take longer for customers such as Meta, Microsoft, and Google to apply the product to their data centers.