What has changed?
NVDA is not only selling CPUs now—they stated that they have commitments for some $20 billion worth of sales this year.
According to the company, this sets NVDA up to become “the world leading CPU-supplier.” Over the long term, they view CPUs as a $200 billion market opportunity.
As we explained last week in NVIDIA Delivers Durable—and Broadening—Growth, NVDA has strong technological motivations to move in this direction. The CPUs they are designing are not the CPUs of yesterday but are specifically designed as AI-enabled complements to their own suite of GPUs.
On the earnings call, CEO Jensen Huang described the new Vera CPU architecture as a major pillar of the next generation of AI infrastructure.
To be clear, NVDA is not trying to compete head-on with the likes of Intel (INTC) in the legacy PC market. Instead, the company is building highly specialized processors designed specifically to work alongside their own GPUs inside AI factories.
AI agents are driving the shift.
AI is rapidly evolving from basic chatbot interactions into autonomous “agentic AI” systems that are capable of reasoning, coding, planning, and interacting with software in real time.
While GPUs are still doing most of the heavy lifting, CPUs need to keep up.
CPUs are playing an increasingly critical role in many areas, including orchestrating data movement, handling memory operations, and running the surrounding software infrastructure.
As Jensen explained on the call, agentic AI systems, like Anthropic’s Claude Code or OpenAI’s Codex, are essentially a “harness” for an AI model that runs on a CPU. With the number of AI agents potentially scaling into the billions, this creates a vast need for CPUs that perform this function efficiently.