Archive · 24 Aug 2026
Nvidia Raises AI Server Prices Over 15% Amid Memory Costs
Nvidia makes important computer chips used to build and run artificial-intelligence software.
Companies that assemble servers for large technology firms have told customers that prices will rise.
In many cases, the increase will be more than 15 percent.
The higher prices are expected to affect machines shipped early next year.
A major reason is that memory chips, which work alongside Nvidia’s processors, have become much more expensive.
Samsung Electronics, SK Hynix and Micron make most of these memory chips.
Big companies are trying to make more of their own AI chips, but they still rely on Nvidia for many data-center projects.
The higher costs could make some AI construction projects more difficult or slower.
Nvidia customers have been notified that server prices will rise by more than 15% in many cases.
The increases will affect systems shipped early next year, including those using Vera Rubin and Grace Blackwell chips.
Rising DRAM costs are driving the increases, with Samsung Electronics, SK Hynix and Micron dominating supply.
Major customers such as Microsoft, Google, Oracle and Meta remain dependent on Nvidia despite developing their own chips.
Higher prices could complicate AI data-center construction, which already faces delays, labor shortages, tighter financing and community resistance.
- Who
- Nvidia and the companies that build servers for major data-center operators, including Microsoft, Google and Oracle.
- What
- Server prices for systems containing Nvidia AI chips are set to rise by more than 15% in many cases.
- When
- The increases are expected to apply to systems shipped early next year.
- Where
- The report concerns the global AI infrastructure and semiconductor industry; no specific location for the price notices was given.
- Why
- Soaring memory-chip costs are increasing the expense of servers that pair Nvidia processors with DRAM.
- Expected increase
- More than 15% in many cases, with the exact amount depending on Nvidia chip generation and memory configuration.
- Effective timing
- Systems shipped early next year.
- Affected systems
- Servers using Nvidia’s Vera Rubin and Grace Blackwell chips, among others.
- Main cost pressure
- Rising prices for dynamic random-access memory, or DRAM.
- Leading memory suppliers
- Samsung Electronics, SK Hynix and Micron Technology account for most global DRAM production.
- Nvidia gross margin
- The article says Nvidia has a gross margin of 75%.
- Nvidia response
- Nvidia representatives did not respond to requests for comment.