Jensen Huang has built a reputation as one of the tech industry’s most aggressive corporate promoters, but Nvidia’s consistent financial performance has helped validate many of his ambitious claims.

During Wednesday’s earnings call, following another record-breaking quarter in which Nvidia generated $81.6 billion in revenue and forecast $91 billion for the next period, Huang said that Nvidia’s new Vera CPU has opened up a previously untapped $200 billion total addressable market.

The claims come at a time when investors are increasingly questioning whether Nvidia can maintain its dominance as competition intensifies in AI hardware.

While Nvidia dominates GPUs, CPUs have long been controlled by companies such as Intel and AMD. At the same time, cloud providers including Amazon Web Services are investing heavily in their own AI processors.

AWS recently announced a major agreement with Meta involving millions of internally developed AI CPUs, and Amazon CEO Andy Jassy has openly argued that AWS can compete with Nvidia on AI chips.

What Vera actually is

Nvidia launched the Vera CPU in March, which it describes as the world’s first processor purpose-built for agentic AI and reinforcement learning.

The company said Vera delivers twice the efficiency and 50% faster performance than conventional rack-scale CPUs, targeting workloads tied to AI reasoning, orchestration, coding agents and large-scale inference.

According to the company, leading cloud providers including Alibaba Cloud, ByteDance, Meta and Oracle Cloud Infrastructure are collaborating to deploy Vera, while manufacturing partners include Dell Technologies, HPE, Lenovo, Supermicro, ASUS, Foxconn and others.

Huang said that while GPUs perform the computational “thinking” behind AI models, CPUs are critical for operational workloads performed by autonomous agents. He predicted that the future would include billions of AI agents, each requiring computing systems similar to today’s PCs.

Unlike traditional cloud CPUs focused on maximizing multi-core application performance, Vera is optimized for rapid token processing.

Huang said Nvidia has already sold $20 billion worth of standalone Vera CPUs this year and believes the product positions Nvidia at the center of the transition toward agentic and robotic AI computing.