Nvidia CEO Signals Major Push Into CPU Market to Challenge Intel, AMD

Nvidia's CEO Jensen Huang is positioning the company to compete directly with Intel and AMD in the central processing unit market. The graphics chip giant believes CPUs are making a comeback as artificial intelligence shifts from training to deployment phases.

Graphics chip powerhouse Nvidia is setting its sights on a direct confrontation with longtime processor rivals Intel and Advanced Micro Devices, as CEO Jensen Huang signals a major expansion into the central processing unit market.

While Nvidia built its massive wealth through specialized graphics chips that drive artificial intelligence systems, Huang is now expressing growing enthusiasm for the more versatile CPU technology that has long been dominated by Intel and AMD.

For many years, central processing units served as the primary computing engine in computers. However, Huang frequently points out how the computing landscape has dramatically shifted, with his company’s specialized chips now handling tasks that CPUs once managed almost exclusively.

The tide appears to be turning once again. As artificial intelligence companies transition from developing their systems to actually putting them to work, CPUs are experiencing a resurgence – and Nvidia wants a significant piece of that action.

“We love CPUs as well as GPUs,” Huang told financial analysts during Wednesday’s fourth-quarter earnings discussion.

The executive emphasized that Nvidia stands ready for the CPU’s return to prominence, confident that the company’s data center processors, which debuted in 2023, will surpass competing products.

During January’s Consumer Electronics Show in Las Vegas, Huang made an even bolder prediction, suggesting that high-performance Nvidia CPUs in data centers would see explosive growth. He said he wouldn’t be shocked “if Nvidia becomes one of the largest CPU makers in the world.”

The fundamental difference between these two chip types has defined their roles for decades. Central processing units function as versatile workhorses, capable of managing virtually any computational challenge software developers present, though at moderate speeds due to their broad capabilities.

Graphics processing units take a different approach, focusing on simpler mathematical operations but executing thousands of these calculations simultaneously. This parallel processing power made them ideal for rendering video game graphics and, more recently, for the matrix calculations that power artificial intelligence systems.

The computing landscape is evolving as AI companies deploy autonomous “agents” capable of independently handling complex tasks like programming, document analysis, and research compilation. According to Creative Strategies analyst Ben Bajarin, this type of work “is happening more and more, and sometimes primarily, on the CPU.”

Nvidia’s current top-tier AI server, the NVL72, incorporates 36 CPUs alongside 72 graphics processors. Bajarin anticipates this could shift to equal numbers for agent-based computing, or potentially eliminate graphics chips entirely for certain applications.

Highlighting its processor ambitions, Nvidia recently struck a deal with Meta Platforms, where the social media giant will purchase substantial quantities of Grace and Vera CPU chips as standalone products. This marks a departure from Nvidia’s typical server configurations that pair each CPU with multiple graphics processors.

However, Meta hasn’t abandoned its existing CPU suppliers – the company is simply diversifying its sources. AMD subsequently announced its own significant Meta agreement, continuing a relationship spanning several years.

During the analyst call, Huang explained Nvidia’s distinctive CPU philosophy, contrasting it with the modular approaches favored by Intel and AMD. He described how Nvidia’s processors excel at handling sequential simple tasks while maintaining excellent access to substantial memory resources.

“It is designed to be focused on very high data processing capabilities,” Huang explained. “And the reason for that is because most of the computing problems that we’re interested in are data driven – artificial intelligence being one.”

HotTech Vision and Analysis principal analyst Dave Altavilla believes Nvidia aims to demonstrate that Intel’s traditional CPU dominance “is no longer the assumed default foundation of modern compute infrastructure. Instead, it becomes just one architectural option among several.”

Huang indicated that additional CPU details would be revealed at the company’s upcoming annual developer conference in Silicon Valley next month.