NVIDIA, having established dominance in the global GPU data center market, is now setting its sights on becoming a leading global CPU supplier. This strategic shift was highlighted during the company's Q1 earnings call on Thursday.
Chief Financial Officer Colette Kress announced the company's entry into the CPU space, describing it as a total addressable market worth $200 billion, a segment NVIDIA had not previously targeted. Kress stated that the company currently expects its standalone CPU revenue to approach $20 billion this year, with all major hyperscalers and system manufacturers collaborating with NVIDIA to deploy these chips.
CEO Jensen Huang emphasized the broader industry transformation, stating, "The world is rebuilding the computing architecture for agentic AI and robotic physical AI. NVIDIA is at the center of these shifts."
While Graphics Processing Units (GPUs), with their capability to train and run AI models, have been central to NVIDIA's success, the rise of AI agents is making Central Processing Units (CPUs) an increasingly critical component of AI data centers. AI agents can be thought of as semi-autonomous or fully autonomous digital assistants capable of tasks like organizing emails, browsing the web, and managing files. While GPUs power the AI models themselves, CPUs are essential for executing the operations these agents perform on behalf of users.
NVIDIA has offered its own CPUs for years, primarily pairing them with GPUs in AI servers. For instance, the company's GB300 chip integrates one Grace CPU with two Blackwell Ultra GPUs. Similarly, its next-generation Vera Rubin superchip combines a Vera CPU with a Rubin GPU.
In March, NVIDIA announced it would ship its first CPU-only Grace servers to Meta Platforms, Inc., with plans to deploy Vera-only servers by 2027. Huang noted, "We are going to need a lot more CPUs. Vera is designed to be the agentic AI CPU."
The surging market demand for CPUs has particularly benefited Intel, whose stock has soared 222% year-to-date. However, the company still faces resource constraints, with demand significantly outstripping supply.
NVIDIA's competitor, AMD, also reported increased CPU demand in its latest earnings, attributing $5.8 billion in revenue—a 57% year-over-year increase—to strong performance from its EPYC CPUs and Instinct GPUs.
Comments