Arista's CEO Points to a Shift in AI Chip Preference, Sending Ripples Through the Market
A notable divergence in the stock prices of Nvidia and Advanced Micro Devices this week points to a subtle but significant shift in the high-stakes AI hardware sector. The catalyst was commentary...
A notable divergence in the stock prices of Nvidia and Advanced Micro Devices this week points to a subtle but significant shift in the high-stakes AI hardware sector. The catalyst was commentary from Arista Networks CEO Jayshree Ullal, who indicated a portion of her company's deployments are now favoring AMD's accelerators.
Speaking on Arista's earnings call, Ullal noted that while Nvidia was nearly universal a year ago, AMD is now the preferred choice in roughly a quarter of new deployments. Following her remarks, Nvidia shares slipped nearly 3%, while AMD's stock edged up about 1%.
The movement underscores a market beginning to test Nvidia's long-held dominance in AI chips. Since the AI boom accelerated in late 2022, Nvidia has commanded an estimated 90% of this critical market. AMD, with a market value of $335 billion, represents its most direct challenger.
Arista, a key player in networking for AI data centers, partners with AMD to build specialized AI clusters. However, Ullal's observation also reflects business reality. Nvidia has been developing its own networking technology, Spectrum-X, reducing its reliance on partners like Arista. Last fall, announcements that Meta and Oracle would adopt Nvidia's networking contributed to a sharp, temporary drop in Arista's stock.
"It helps to be designed into Nvidia's systems, and Arista is being designed out, at least to some extent," noted Ben Bajarin of Creative Strategies.
While the stock shifts were modest, Ullal's statement is a clear signal: the architecture of AI infrastructure is evolving, and with it, the balance of power among its suppliers. Nvidia declined to comment. AMD did not immediately respond to a request for comment.
Source: CNBC
Ready to Modernize Your Business?
Get your AI automation roadmap in minutes, not months.
Analyze Your Workflows →