Nvidia eyes nearly $20 billion in CPU sales as AI agents push demand beyond GPUs

Nvidia is increasingly being framed as a potential force in CPUs, not just the graphics processing units (GPUs) that power most of today’s artificial intelligence workloads. In a recent analysis published by Yahoo Finance, the central question is whether an AI agent, not just a model, needs a CPU to operate effectively, and the answer offered there points to a meaningful opportunity for Nvidia in the server chip market.

The report argues that AI agent functionality can require more than GPU acceleration. While GPUs handle the heavy lifting for machine learning computations, CPUs also manage general-purpose tasks such as coordinating system activity and running parts of the software stack. In that framing, an agent is not purely a GPU workload, which creates a pathway for CPU demand to scale alongside the AI boom.

Yahoo Finance also highlighted a specific business expectation: Nvidia “expects almost $20 billion in CPU sales this year.” That figure, as presented in the market commentary, suggests Nvidia sees CPU revenue as a major growth pillar rather than a side business.

The same analysis connects that expectation to competitive dynamics in the broader PC and server processor ecosystem. AMD and Intel have historically dominated CPU sales, and any shift in AI-driven infrastructure spending could reshape the customer mix for data center hardware. Nvidia’s move would be significant because the company is best known for GPUs, and most buyers still associate it primarily with AI accelerators.

Even without new, detailed disclosures in the article itself, the implication is that Nvidia wants a bigger role in the parts of compute systems that are not strictly GPU compute. That is a notable strategic adjustment because it changes how customers might plan their “full stack” infrastructure for training and inference, including orchestration between CPUs, GPUs, memory subsystems, and networking.

In the technology sector, AI spending has been heavily concentrated in accelerators, but buyers are also building out the surrounding platform. If AI agents require more CPU-driven integration work, the market could reward vendors that can supply both accelerators and the general compute processors that help run the end-to-end system.

What remains unclear from the limited information available here is how Nvidia expects to meet this CPU revenue target in terms of product mix, shipment timing, and customer adoption. The market commentary does not provide granular guidance on contracts, product specifications, or which customers are driving CPU growth.

Investors and industry watchers will likely focus on whether Nvidia’s CPU opportunity translates into sustained orders and margin profile, and whether the industry’s definition of “AI agents” consistently leads to higher CPU content per deployed system. The next key indicates to watch are Nvidia’s own disclosures around CPU-related revenue and any additional detail that clarifies why AI agent workloads increase CPU demand, not just GPU utilization.