Some of Nvidia's biggest customers have been told that prices for servers containing its artificial intelligence chips will rise by more than 15% in many cases as memory costs soar, Bloomberg reported Saturday, citing sources familiar with the situation.

The price hikes, scheduled to take effect on systems shipped out early next year, will affect servers with Nvidia's flagship Grace Blackwell and Vera Rubin chips. The precise amount of the increase will vary based on the chip's generation and memory configuration, per the report.

Server manufacturers under contract for major data center operators including Microsoft, Alphabet's Google, and Oracle recently notified customers of the increases, the report states. Nvidia representatives did not respond to Bloomberg's requests for comment.

Reuters noted it could not independently verify the report and did not receive a response from Nvidia outside regular business hours.

Memory shortage hits AI servers

Rising demand in AI chips has given three companies, Samsung Electronics, SK Hynix, and Micron Technology, greater power over pricing, as those firms produce the dynamic random access memory (DRAM) components that Nvidia's chips rely on to supply data to their processors.

However, the Bloomberg report did not say whether or not Nvidia intends to increase the prices of its standalone chips; the report concerned only complete servers, whose prices also incorporate memory and other components.

The Latent reported last month that Samsung expects the current memory shortage crisis to worsen in 2027 and persist into 2028. New fabrication plants take nearly four years to begin producing chips, Samsung claimed, which limits the industry's ability to add memory supply quickly.

Nvidia's largest cloud customers, including Amazon, Microsoft, Google, and Meta, are developing their own AI chips but remain dependent on Nvidia systems for their data center expansions, Bloomberg reported. Their internal chips also require memory from Samsung, SK Hynix, or Micron.

Higher server prices will likely add to the cost of AI data center projects already facing construction delays, labor shortages, tight capital markets, and intensifying local opposition. Nvidia has also moved deeper into infrastructure development; the company took a minority stake in Cloverleaf Infrastructure on Friday as part of a partnership to develop AI factory sites.

Nvidia is scheduled to report its Q2 financial results on Wednesday, Aug. 26, according to the company's investor relations site.