Nvidia, the world's most valuable company by market capitalization and the largest artificial intelligence (AI) chip developer, signaled a price hike for AI servers.
On the 22nd (local time), Bloomberg News, citing sources, reported that Nvidia recently told major clients that prices for its AI chip systems would rise by more than 15%. The price increase is set to apply to units shipped from early next year. Flagship products "Vera Rubin" and "Grace Blackwell" are also included in the increase.
As demand has surged amid the recent AI boom, memory semiconductor prices are also trending higher. Apple and Qualcomm have already raised product prices, citing a chip shortage. This time, Nvidia has joined the price hike bandwagon.
This is due to a shortage of memory semiconductors. The accelerator processors Nvidia sells are like the heart essential for running AI software, and for these processors to deliver full performance, pairing with DRAM memory semiconductors is indispensable.
At this point, most of the world's memory semiconductor output is handled by Samsung Electronics(005930) and SK hynix(000660), and Micron, three corporations in total. Bloomberg News explained that, in connection with this AI server price increase, this clearly shows how much leverage these memory chipmakers have at the negotiating table.
Currently, Nvidia's major customers, including Amazon, Microsoft (MS), Google, and Meta, are pursuing in-house semiconductor development. But even so, they cannot avoid relying on supply from memory chip manufacturers.
Nvidia's AI server price hike is expected to present a new obstacle to building large-scale AI data centers. Bloomberg News said the increase will add to the burden, as data center construction is already struggling with project delays, labor shortages, tight capital markets, and community opposition to development.