355
Audio & Video Production344
Automation & Workflow224
Software Development250
Marketing & Growth192
AI Infrastructure & MLOps173
Writing & Content Creation203
Data & Analytics140
Design & Creative169
Customer Support130
Photography & Imaging156
Sales & Outreach125
Voice & Speech135
Operations & Admin87
Education & Learning131
Jensen Huang says Nvidia’s Vera CPU is built for AI agents and could create a new $200B market, as the company looks beyond its best known AI chips.
In short: Nvidia CEO Jensen Huang says a new Nvidia CPU called Vera could open a $200 billion market by powering AI agents.
Nvidia CEO Jensen Huang told investors that he sees a new growth area for the company, selling CPUs for what he calls “AI agents.” AI agents are AI systems that can take actions for you, like booking meetings, moving files, or running steps in a business process (like a helpful assistant that can actually do the work, not just answer questions).
Huang connected this idea to Nvidia’s new CPU product, called Vera, which Nvidia introduced in March. A CPU is the main “general purpose” chip in a computer. It is like the manager that handles many everyday tasks, while a GPU, which Nvidia is best known for, is like a specialist that is very fast at certain kinds of heavy math.
Huang said that GPUs handle much of the “thinking” behind modern AI models, but that agents mostly run on CPUs when they carry out tasks. He also said Vera is designed to process “tokens” quickly, meaning it can handle the small chunks of text and instructions that AI systems work with, similar to how a cashier scans items one by one.
Huang said Nvidia has already sold $20 billion worth of standalone Vera CPUs this year. He also said “every major hyperscaler,” meaning the biggest cloud computing companies, and system makers are partnering with Nvidia to deploy Vera.
If more software starts using AI agents, demand could grow for the kind of chips that run them, not just the chips that train AI models. That could shape which companies control the costs and speed of future AI services.
Source: TechCrunch AI