Nvidia AI Server Prices May Increase Over 15% Due to Memory Cost Surge

On August 22, Bloomberg reported that several large clients of Nvidia received notices indicating potential price increases of over 15% for AI server systems. The primary factor behind this expected rise is a significant increase in memory costs, which affects the overall price of systems that include accelerators, memory, networking, CPUs, and cooling and power infrastructure. These new prices may apply to systems planned for delivery in early 2027, such as Nvidia's Vera Rubin and Grace Blackwell models. Nvidia has not officially confirmed these reports, and the information is sourced from Bloomberg and Reuters. Such a price increase could substantially raise capital expenditures for hyperscalers and large-scale AI projects, reflecting the growing costs of AI infrastructure components.