MW Nvidia's Jensen Huang wants to be king of a brand-new empire
By Britney Nguyen
The chip maker is paving its way into a potential $200 billion market for central processing units
Nvidia, led by CEO Jensen Huang, has its sights set on a $200 billion total addressable market for CPUs.
Nvidia's graphics processing units powered its rise to becoming the world's most valuable company, and now the chip maker has its sights set on dominating another booming area of the chip market.
The market for central processing units is one "we have never addressed before," Nvidia Chief Financial Officer Colette Kress said on the company's earnings call on Wednesday evening. Nvidia's (NVDA) first standalone CPU for artificial intelligence, Vera, was introduced in March and opens up a "brand new" $200 billion total addressable market for the company, Kress said on the call, adding that hyperscaler customers and AI system makers are deploying the CPUs.
With visibility into nearly $20 billion in CPU revenue this year, Kress said Vera is "setting us up to become the world-leading CPU supplier."
See more: Arm's stock could rise another 45% as the 'renaissance of CPUs' takes hold, analyst says
UBS analyst Timothy Arcuri said that figure is "a much larger number this year than we would have expected." It also tops the $16 billion in CPU revenue Arcuri is modeling for Advanced Micro Devices $(AMD)$, which has a more than 50% share of the cloud segment, he said in a Thursday note.
As a standalone chip, Vera can either be attached to a non-Nvidia AI chip or be deployed in a system that includes CPUs but also memory components, Arcuri noted. Because expensive memory chips are also part of the latter package, he noted that the $20 billion estimate might not be all it seems on the surface. He sees Vera making up about 30% of the value of the system and is modeling that Nvidia's CPU revenue will be more in the range of $5 billion to $7 billion this year.
Bernstein analyst Stacy Rasgon said the $20 billion is likely due to Nvidia selling its CPUs on a rack level instead of individually. Still, "the company seems set to potentially become the CPU king with a possible business as big or even bigger than their more traditional competitors" in the CPU market, he said in a Thursday note.
While much of the ongoing AI race has depended on Nvidia's GPUs for training, the current agentic wave and shift to inference, or the process of running AI models, has renewed intense demand for server CPUs.
That has sparked a resurgence for embattled chip maker Intel $(INTC)$, while chip designer Arm Holdings $(ARM)$ announced its first-ever in-house chip earlier this year in the form of a CPU to support inference and agentic AI workloads.
Morgan Stanley analyst Joseph Moore said the $20 billion target would put Nvidia "right at the cusp of market leadership," though he is modeling slightly higher revenues for Intel's data-center business, which would also count some networking offerings. Like the other analysts, Moore said that Nvidia's number likely includes CPUs being delivered on GPU cards, "which counts but is a bit easier of a sale."
Still, that shouldn't be "incrementally bad" for competing chips built on Intel's x86 architecture, Moore said in a Thursday note. Vera is based on the competing Arm chip architecture.
In Arcuri's view, Nvidia's venture into CPUs will impact already tightening capacity and is therefore "negative for the aspirations of" competitors such as Arm and Qualcomm $(QCOM)$.
Don't miss: Arm has a ton of chip demand. There's just one problem.
Moore said Nvidia's ability to source what it needs should put it in a strong position despite supply constraints. Arcuri, too, noted that Arm could be particularly negatively impacted, given that it has less-established relationships with suppliers of chip wafers and memory.
-Britney Nguyen
This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
May 21, 2026 12:20 ET (16:20 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments