The American technology corporation Nvidia is creating a new division that will develop custom microprocessors, including for artificial intelligence (AI) technologies. The agency reports this Reuters with reference to knowledgeable sources. Nvidia, as the agency notes, is seeking to gain a share of the rapidly growing market for custom microprocessors for AI and protect itself from the ever-increasing number of companies creating alternatives to its products.
Nvidia now controls about 80% of the market for high-performance AI chips. Thanks to this, its market value tripled last year, and since the beginning of this year it has soared by another 40% to reach $1.73 trillion. Nvidia’s clients include OpenAI, Microsoft, Alphabet, Meta Platforms (recognized as extremist in the Russian Federation and banned), and they are rushing to intercept the dwindling supply of Nvidia chips in order to successfully compete in the field of AI.
Nvidia’s H100 and A100 chips serve as general-purpose AI processors for many of these large customers. But technology companies began to develop their own customized chips for specific needs. This helps reduce energy consumption and can potentially reduce costs. And Nvidia, sources tell Reuters, is trying to “play a role” in helping these companies develop such specialized chips. Previously, such orders went to Broadcom and Marvell Technology, which dominate the market for custom chips for data centers. Needham analysts estimate that the custom chip market will be about $30 billion in 2023—that’s 5% of annual global sales of all microprocessors.
After Reuters published information about Nvidia’s plans, the company’s shares on the NASDAQ stock exchange added more than 3%. And Marvell Technology shares fall 2.3%.