Nvidia Price Increase: AI Servers Up More Than 15%

Nvidia price hike on Vera Rubin and Grace Blackwell servers

Bloomberg reported on 22 August 2026 that some of Nvidia's biggest customers have been told that prices of servers containing its AI chips will rise more than 15% in many cases. The report ties the increase to soaring memory chip costs.

The increases will take effect on systems shipped in early 2027 and will affect systems including those with the flagship Vera Rubin and Grace Blackwell chips. The sources add that the size of each increase will depend on the generation of Nvidia chips and the memory configurations.

This is a report, not an Nvidia announcement. Bloomberg's sources are people familiar with the process, who asked not to be identified because the communications have not yet been made public.

According to them, companies that build servers under contract for large data center operators such as Microsoft, Google and Oracle have recently notified their customers of the forthcoming increases. Bloomberg adds that Nvidia representatives did not respond to requests for comment.

DRAM prices and the memory shortage behind the hike

DRAM, short for dynamic random access memory, is the working memory an AI processor is paired with. Bloomberg says Nvidia's accelerators are central to computers that create and run AI software, and that how well they perform depends on how much DRAM comes with them.

Three companies make most of that memory: Samsung Electronics, SK Hynix and Micron Technology. Bloomberg says they account for most of the world's DRAM production and have been raising output, but have not yet caught up with the surge in demand for AI infrastructure.

Bloomberg says the industry's most dominant company has been unable to keep prices steady or absorb growing costs, which shows how much bargaining power memory chip makers have. It adds that Apple and Qualcomm have recently said chip shortages forced them to charge more for their products.

Amazon, Microsoft, Google and Meta are each pursuing in-house chip programs, according to Bloomberg, and each still depends on purchases from Nvidia for its data center build-out. Bloomberg adds that how far they can move toward greater independence will also depend on their access to supply from Samsung, SK Hynix and Micron.

Nvidia Q2 earnings: gross margin and Q3 guidance

Nvidia published its second-quarter results on 26 August 2026, four days after the Bloomberg report. Gross margin is the share of revenue left after the direct cost of what was sold, and guidance is the forecast a company gives for its next quarter.

For the quarter ended 26 July 2026, revenue was $96.2 billion, up 18% from the previous quarter and up 106% from a year ago. Data Center revenue was $89.0 billion, up 117% from a year ago. GAAP and non-GAAP gross margins were both 75.0%.

For the third quarter, Nvidia expects revenue of $108.0 billion, plus or minus 2%. It expects GAAP and non-GAAP gross margins of 74.0%, plus or minus 50 basis points, which is half a percentage point either way. The outlook assumes no Data Center compute revenue from China.

The release also says Vera Rubin is in full production. It does not mention the reported price notices, and neither source says that one caused the other.

To keep a report like Bloomberg's for later, use the URL to Markdown tool, which converts a web page into Markdown from its address. Paste the URL and copy the result.