In the case of gaming graphics cards only, Nvidia products are more or less equipped with publicly available information on suggested retail prices and are not updated over time. Nvidia’s server solutions in this regard remain in the shadows, with only releases Bloomberg Just so you know, starting in the new year, the cost of a single AI accelerator will rise by more than 15%.
Image source: NVIDIA
Bloomberg quoted sources familiar with Nvidia’s plans as saying that server systems including the Vera Rubin and Grace Blackwell series of computing accelerators will begin shipping early next year, and prices will increase by a certain amount. Price increases will not be uniform; it will depend on the generation and accelerator memory configuration of the GPU itself. In fact, rising memory prices are the main factor driving up the price of Nvidia’s computing accelerators. The company’s customers have been notified accordingly.
Essentially, this confirms the inability of large electronics manufacturers such as Apple and Nvidia to endlessly resist rising memory chip prices. Although not always public, they are forced to admit that they will pass on rising costs to the production costs of their end customers. At the same time, Nvidia obviously will not operate at a loss, and its profit margin has recently reached 75%. This is lower than memory manufacturers, which is a bit natural, but still high by industry standards. There have also been recent reports that retail prices for Nvidia gaming solutions have also increased globally. When it comes to servers, rising prices for Nvidia components could complicate the implementation of many already costly data center projects.
If you find an error, select it with your mouse and press CTRL+ENTER.