Nvidia customers notified of AI-related price rises above 15%

Some of Nvidia's biggest clients have been informed that prices for servers housing its artificial intelligence chips will increase by over 15 percent in numerous instances, owing to soaring memory chip costs, according to sources familiar with the matter. These price hikes will take effect for systems shipped early next year, affecting systems featuring Nvidia's flagship Vera Rubin and Grace Blackwell chips, among others.

The extent of the price increases hinges on the specific Nvidia chip generation and memory configurations, as revealed by the sources. Companies manufacturing servers on contract for major data center operators like Microsoft, Alphabet's Google, and Oracle have notified their customers about the impending price hikes, the sources said.

Nvidia representatives declined to comment on the matter. The inability of Nvidia, the industry's dominant player, to maintain stable prices or offset rising costs highlights the formidable leverage wielded by memory chip manufacturers such as Samsung Electronics, SK Hynix, and Micron Technology, amidst a surge in demand for AI infrastructure.

Tech giants like Apple and Qualcomm have also had to raise prices for their products due to ongoing chip shortages. Nvidia's AI accelerator processors serve as the core of computers that generate and operate AI software, with their performance contingent on the amount of dynamic random access memory (DRAM) they are paired with. While Samsung, SK Hynix, and Micron Technology have ramped up production, they have yet to catch up with the surging demand, resulting in a massive price surge for these essential components and conferring significant influence upon their manufacturers.

Nvidia, one of the most profitable semiconductor companies, can charge tens of thousands of dollars per chip due to supply constraints from contract manufacturer Taiwan Semiconductor Manufacturing. The company boasts a gross margin of 75 percent, which stems from its origins in PC gaming chips that sold for hundreds of dollars, with AI accelerators' prices being driven up by insatiable demand and a lack of viable alternatives to Nvidia's offerings.

Nvidia has also increased prices for its gaming-oriented PC graphics cards, as reported by industry news outlet Tom's Hardware earlier this month. The impact of these price increases on Nvidia's customers and potential opportunities for competitors will hinge on whether they can secure sufficient memory supplies. Major customers such as Amazon, Microsoft, Google, and Meta are pursuing their own in-house chip programs but remain reliant on Nvidia for their data center buildouts.

Their capacity to progress independently will also depend on their access to memory supplies from Samsung, SK Hynix, and Micron Technology. The price hikes could further complicate the industry's ambitious AI data center buildout efforts, which are already facing challenges such as project delays, labor shortages, tightening capital markets, and community resistance to new developments.

Nvidia is scheduled to release its second-quarter earnings next week, a highly anticipated update for the technology industry and investors who have invested heavily in AI infrastructure in the belief that it will revolutionize the economy.