Nvidia Alerts Customers Of Over 15% AI Server Price Increase

Nvidia customers are being notified of price increases of more than 15% on some servers containing the chipmaker’s artificial intelligence (AI) processors, as surging memory costs add to the expense of building AI infrastructure.

The increases are expected to take effect on systems shipped early next year and will affect servers using Nvidia’s latest AI platforms, including its Vera Rubin and Grace Blackwell chips, as per a report by Bloomberg (BBG).

The extent of the increase will vary depending on the Nvidia chip generation and the amount of memory configured in each system.

The pricing changes come ahead of Nvidia’s fiscal second-quarter earnings, due next week. The results are closely watched by investors and the wider technology industry as spending on AI infrastructure continues to accelerate.

Memory Costs Drive Server Prices

Contract manufacturers that build servers for major data centre operators, including Microsoft, Alphabet’s Google and Oracle, have recently informed customers about the planned increases, as per BBG.

The development highlights the growing pricing power of memory manufacturers Samsung Electronics, SK Hynix and Micron Technology as demand for AI infrastructure continues to outpace supply.

The three companies account for most global DRAM production, and prices for the commodity-like components have risen sharply amid persistent demand, the BBG report said.

Nvidia’s AI accelerators rely heavily on DRAM to deliver the performance required for training and running AI models. Although memory producers have been expanding capacity, supply has yet to catch up with demand.

Nvidia Faces Rising AI Infra Costs

Nvidia remains one of the semiconductor industry’s most profitable companies, with its gross margin at about 75%, as per Bloomberg.

Its AI processors can cost tens of thousands of dollars each, while supply constraints at Taiwan Semiconductor Manufacturing Co., which manufactures the chips, continue to limit the ability to meet demand.

Nvidia has also increased prices for some gaming-focused graphics cards, technology publication Tom’s Hardware reported earlier this month, as per the BBG report.

The latest increases could put further pressure on the cost of large AI data centre projects. These developments are already facing challenges including project delays, labour shortages, tighter financing conditions and resistance from local communities.

Major technology companies such as Amazon, Microsoft, Google and Meta are reportedly developing their own AI chips to reduce dependence on Nvidia.

However, they continue to rely on Nvidia processors for data centre expansion and will remain exposed to memory supply constraints.