Nvidia customers told AI server prices to rise more than 15% on soaring memory costs

Some of Nvidia's largest customers have been told that prices of servers containing its AI chips will rise by more than 15% in many cases, with memory chip costs soaring, as reported by Bloomberg News on Saturday, citing people familiar with the process. The price hikes will go into effect on systems shipped early next year and will impact systems including those with the flagship Vera Rubin and Grace Blackwell chips. The increases will depend on Nvidia's chip generation and memory configurations, sources told Bloomberg.

Companies that build servers under contract for large data center operators such as Microsoft, Google, and Oracle have recently informed their customers of the upcoming increases, according to the same report. Reuters, which covered the Bloomberg story, noted it could not immediately verify the report, and Nvidia did not immediately respond to a request for comment.

The news comes as Nvidia is set to report second-quarter results on August 26. The company has become a proxy for the broader AI ecosystem spanning chip makers and companies financing the rapid expansion of data center capacity.

Memory costs squeeze the AI supply chain

The driver behind the price hikes is a sharp rise in DRAM prices, the type of memory paired with Nvidia's accelerator processors. Samsung, SK Hynix, and Micron together produce most of the world's DRAM, and despite ramping up output, they have not caught up with surging AI-driven demand. This supply-demand imbalance has given memory makers unprecedented leverage, forcing even the industry's most dominant company to pass on the rising costs. Apple and Qualcomm have also raised their own prices in recent months due to chip shortages, as noted in coverage of the Bloomberg report.

The increase will not be uniform. As reported by The News International, Blackwell-based NVL72 GB200 racks have been priced around $2.8 million to $3.4 million, while early enquiries for Vera Rubin-based NVL72 have reached $5 million to $7 million. A 15% rise would add hundreds of thousands of dollars to the price of a single system.

Nvidia is not alone in facing margin pressure. Major technology companies including Apple and Microsoft have failed to hold their product price lines because of the chronic chip shortage, as The Economic Times noted in its coverage of the price increases.

Indian IT stocks rally on 'anti-AI' bets

Meanwhile, Indian IT stocks have rallied on the news. The Nifty IT Index edged up three out of the last four trading days, with Infosys and HCLTech gaining more than 1% each on Monday. The KOSPI, South Korea's tech-heavy benchmark, fell 3% as investors sold off AI-linked shares.

HSBC said India can serve as an "anti-AI" diversifier as sharp swings in technology-exposed markets encourage foreign investors to broaden their portfolios. Hexaware was the biggest gainer among IT stocks on Monday after the company revealed that more than half of its revenue is already 'AI-infused', leading CLSA to project a 37% rise in its stock in the next 12 months. CLSA has also downgraded TCS, Infosys, and Tech Mahindra to 'Hold', and Wipro and Mphasis to 'Underperform', as per The Economic Times.

The memory chip shortage is not expected to ease soon. New semiconductor fabs take years to build, and new production is unlikely to come on stream until late 2028. The chip shortage could extend until the end of the decade, according to some estimates. Rising server costs will increase AI model training and deployment expenses, and these costs are likely to be passed on to client enterprises. AI companies are also facing energy and compute constraints, as well as community resistance to data centre buildouts.

Outlook

Nvidia's price increases could slow the expected decline in the cost of building and running AI systems. The company has not publicly confirmed the reported price increases, and the exact cost pass-through and impact on spending plans remain questions. Nvidia is also raising prices for its gaming-oriented PC graphics cards, according to a report from Tom's Hardware, and has already raised GPU prices this year, as The Verge noted. The company's production is constrained by its contract manufacturer's ability to meet demand.

As Nvidia's largest customers, including Amazon, Microsoft, Google, and Meta, continue to develop their own processors, the price hikes add another layer of complexity to the AI data center build-out. Nvidia is set to report its second-quarter results on August 26, which will be closely watched for guidance on pricing and supply.