Nvidia Corporation (NVDA) is shedding light on its anticipated revenue from CPU sales for the current fiscal year, with significant expectations in place. During an interview with Yahoo Finance on Thursday, Chief Financial Officer Colette Kress indicated that the company is targeting $20 billion in revenue from the sale of both standalone CPUs and those integrated within Nvidia’s newly introduced Grace Blackwell and Vera Rubin superchips.
Kress highlighted the opportunity to sell standalone CPUs independently, stating, “We now have an opportunity to also have what we call a standalone CPU, and selling that.” She mentioned that there is already a demand from customers for these CPUs in conjunction with the Vera Rubin and Grace Blackwell chips, and the company believes that this combination could result in $20 billion in CPU revenues this fiscal year.
This reinforcement of expectations aligns with comments made by CEO Jensen Huang during the company’s recent earnings call, where he noted that standalone CPU sales might reach $20 billion. Huang also emphasised the larger potential of the CPU market, valued at a staggering $200 billion for Nvidia.
According to Kress, the customer base for CPU-only servers will likely mirror those currently purchasing Nvidia’s superchip-based systems that rely on graphic processing units (GPUs). “We see this continuing within our full systems and a very big market now for just the standalone as well,” she elaborated.
Nvidia’s success has largely been built on its GPUs, which are pivotal in training and executing Artificial Intelligence (AI) models. Historically, the company has produced its own CPUs to complement its GPUs within AI server configurations. For instance, Nvidia’s GB300 chips incorporate one Grace CPU and two Blackwell Ultra GPUs, while the Vera Rubin superchips consist of Vera CPUs and Rubin GPUs.
A notable shift is occurring as Nvidia now intends to market its CPUs as standalone components in their own servers, primarily driven by the growing demand for AI agents. These AI agents function as semi-autonomous or fully autonomous digital assistants, capable of performing tasks such as managing emails, browsing the internet, and sorting files. In this setup, the GPU manages the AI model, while the CPU executes these tasks efficiently.
This development is redefining dynamics in the semiconductor industry and elevates the CPU, which previously resided in the shadow of GPUs. Back in March, Nvidia announced plans to commence shipments of its first CPU-only Grace servers to Meta, with Vera-only servers slated for deployment by 2027.
The changing landscape in AI technology and computing architectures is likely to create a fertile environment for continued growth in the CPU market, further enhancing Nvidia’s position within an increasingly competitive industry. As the company continues to innovate, the reliance on both CPUs and GPUs in tandem is expected to drive substantial revenue and opportunity in the years to come.
