Nvidia to raise AI server prices by more than 15% as memory costs climb
Nvidia is preparing to increase prices on many AI servers that use its chips by over 15% after memory component costs surged, a move that will be reflected in its upcoming earnings report.
Nvidia Corp. is set to lift prices on a range of AI servers that incorporate its Grace Blackwell and next-generation Vera Rubin processors by more than 15%, according to Bloomberg. The hikes, which will take effect when shipments begin early next year, vary by chip generation and memory configuration and have already been communicated by contract manufacturers serving major customers such as Microsoft, Alphabet and Oracle.
The price adjustments stem from sharply higher costs for high-bandwidth memory and conventional DRAM, where suppliers including Micron, Samsung and SK Hynix have struggled to keep pace with AI-driven demand. Nvidia’s ability to pass these costs downstream suggests that its customers remain willing to pay premium prices for scarce computing capacity.
In the most recent quarter, Nvidia reported revenue of $81.6 billion, an 85% year-over-year increase, with data-center sales climbing 92% to $75.2 billion and a non-GAAP gross margin of 75%. The company projects fiscal second-quarter revenue of roughly $91 billion and expects to maintain a similar gross-margin level.
Analysts will watch the company’s earnings report, scheduled for after the market close on Aug. 26, for guidance on margins, demand for Vera Rubin chips, supply conditions and any commentary on ongoing memory price inflation. If Nvidia can sustain its margins while shifting higher component costs to buyers, the price hikes would underscore its pricing power in the AI market.