Nvidia is preparing to unveil new details about its agentic-optimized CPUs at its annual GTC AI conference, marking a significant pivot in the company's chip strategy as CPUs become increasingly central to AI workflows [1]. According to Dion Harris, Nvidia's head of AI infrastructure, CPUs are 'becoming the bottleneck' in growing out AI and agentic workflow capabilities, presenting an 'exciting opportunity' for the company [1]. The upcoming GTC conference, which kicks off on Monday, is expected to showcase a CPU-only rack, highlighting Nvidia's renewed focus on central processing units [1].
Nvidia's CPU strategy shifted in February with the announcement that standalone processors are now deployed in Meta data centers, following a multiyear deal that included the first large-scale deployment of Grace CPUs, with plans to deploy the next-generation Vera CPUs in 2027 [1]. Thousands of standalone Nvidia CPUs are also powering supercomputers at the Texas Advanced Computing Center and Los Alamos National Lab [1]. The chip giant's first data center CPU, Grace, was launched in 2021, and Vera is now in production, typically deployed alongside Nvidia's Hopper, Blackwell, or Rubin GPUs in rack-scale systems [1].
Exploding demand for GPUs has propelled Nvidia to a $4.4 trillion market cap, making it the most valuable publicly traded company in the world [1]. In the latest quarter, Nvidia generated over $62 billion in data center revenue, up 75% year-over-year [1]. Bank of America forecasts the CPU market could more than double from $27 billion in 2025 to $60 billion by 2030, underscoring the growing importance of CPUs in AI infrastructure [1].
The resurgence of CPUs is driven by a fundamental shift in compute needs as AI adoption moves from simple chatbots to complex, task-oriented agentic applications. While GPUs excel at parallel processing for AI model training and inference, CPUs are increasingly required for orchestrating large-scale agentic workflows, moving vast amounts of data and coordinating multiple AI agents [1]. Nvidia CEO Jensen Huang emphasized on the recent earnings call that agentic systems are spawning teams of agents, increasing the demand for general compute power [1].
CONCLUSION
Nvidia's strategic pivot toward AI-optimized CPUs, highlighted by upcoming announcements at GTC, signals a major shift in the semiconductor landscape as CPUs become critical for next-generation AI workflows. With strong market growth forecasts and significant deployments in data centers and supercomputing, Nvidia is positioned to capitalize on the evolving demands of agentic AI. The market response is likely to be positive, given Nvidia's leadership and the expanding role of CPUs in AI infrastructure.