Nvidia Prepares Price Hikes Amid Soaring Memory Costs

Nvidia is preparing to raise prices on its AI servers by more than 15% due to soaring memory costs and relentless data-center demand. The move highlights the company's pricing power ahead of its upcoming earnings report.

According to Bloomberg, some of Nvidia's largest customers have been informed that systems containing Grace Blackwell and next-generation Vera Rubin chips will become more expensive when shipments begin early next year. The price increases will vary depending on the chip generation and memory configuration.

The pressure point is AI servers' massive need for high-bandwidth memory, which conventional DRAM suppliers such as Micron MU, Samsung, and SK Hynix have struggled to expand capacity quickly enough to meet demand. This creates an unusual setup for Nvidia: while higher memory prices raise the cost of delivering complete AI systems, the company's ability to push those increases downstream indicates customers remain willing to pay heavily for scarce computing capacity.

Nvidia generated $81.6 billion in revenue last quarter, a 85% year-over-year increase, with Data Center revenue surging 92% to $75.2 billion and non-GAAP gross margin standing at 75%. The company guided fiscal second-quarter revenue to roughly $91 billion and another 75% non-GAAP gross margin.

Investors will closely watch Nvidia's earnings report on August 26, focusing on gross-margin guidance, Vera Rubin demand, supply availability, and any commentary on memory inflation. If Nvidia can preserve margins while passing higher costs to customers, the price increases would reinforce its pricing-power thesis.