Some of Nvidia's largest customers have been informed that prices on servers containing the company's AI chips will rise more than 15% in many cases, according to Bloomberg1,2,3.
The price increases will apply to systems shipped starting early next year and will affect systems built around Nvidia's flagship Vera Rubin and Grace Blackwell chips, according to people familiar with the communications. Memory chip costs are among the factors driving the increases, with the report citing soaring memory prices as a contributor.
The notifications went to some of Nvidia's biggest customers, though the sources who described the communications asked not to be identified because the pricing discussions have not been made public.
ANALYSIS A 15%-plus increase on high-end AI server systems represents a material cost escalation for the data-center buyers that purchase these platforms at scale.
The cited role of soaring memory chip costs suggests the price increases are not solely a function of Nvidia's own chip pricing but reflect broader component-level pressure feeding through the server supply chain.
Timing the increases to early 2027 shipments aligns with the expected ramp of Vera Rubin-based systems, meaning customers placing orders for next-generation hardware will face both a platform transition and higher unit economics simultaneously.