Some of Nvidia’s biggest customers have been notified that prices for servers containing the company’s artificial intelligence chips are set to rise by more than 15% in many cases, as surging memory costs add pressure to the global AI infrastructure buildout.
The increases are expected to apply to systems shipped from early 2027 and will affect configurations using Nvidia’s flagship Vera Rubin and Grace Blackwell chips, according to Bloomberg, citing people familiar with the communications. The exact increase is expected to vary depending on the generation of Nvidia processors and the memory configuration used in individual systems.
Server manufacturers that build systems for major data centre operators, including Microsoft, Alphabet-owned Google and Oracle, have recently informed customers about the upcoming increases. Nvidia had not publicly commented on the reported price changes at the time of publication. Reuters said it could not independently verify the Bloomberg report.
The higher prices come as demand for memory components continues to rise alongside global investment in AI computing infrastructure. Nvidia’s AI accelerators work alongside dynamic random access memory, or DRAM, to process the large amounts of data required for training and running artificial intelligence models.
Samsung Electronics, SK Hynix and Micron Technology account for most global production of these memory chips. While manufacturers have been expanding production, supply has struggled to keep pace with demand from AI data centres and other computing applications. This has pushed component prices higher and increased the influence of memory suppliers across the technology hardware market.
The development could add another cost consideration for technology companies investing heavily in AI infrastructure. Nvidia’s accelerators remain central to many of the data centres being built to support generative AI models, enterprise AI applications and increasingly compute-intensive workloads.
Nvidia has maintained significant pricing power as demand for its AI processors continues to exceed available supply. The company has been able to charge tens of thousands of dollars for individual accelerator chips, while its gross margin stands at around 75%, according to Bloomberg.
At the same time, major cloud companies are developing their own processors to reduce their dependence on external suppliers. Amazon, Microsoft, Google and Meta have all invested in custom silicon programmes, although they continue to purchase Nvidia hardware for large-scale data centre deployments.
The reported price increases come amid broader challenges surrounding the rapid expansion of AI infrastructure. Data centre projects are already navigating capital requirements, construction delays, labour constraints and local resistance in some markets.
Nvidia is scheduled to report its fiscal second-quarter results on August 26, providing investors with another update on demand for AI infrastructure and the company’s position within the global AI computing market.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.