Jensen Huang - founder and CEO of Nvidia continues to attract attention when announcing that the company has found a completely new market worth about 200 billion USD thanks to the Vera AI CPU line.
Speaking at a recent business results announcement meeting, Huang emphasized that Vera could become Nvidia's next growth engine in the context that the company is dominating the GPU market serving artificial intelligence.
The announcement was made after Nvidia continued to record impressive business results with quarterly revenue reaching 81.6 billion USD and forecasts of about 91 billion USD for the next quarter. This shows that global AI infrastructure demand is still increasing strongly.
For many years, Nvidia has been known as the "GPU king", providing graphics processors for AI data centers. However, the CPU market was previously mainly controlled by Intel and AMD.
The AI boom has caused the competition to expand to the CPU, which is the central processor playing a role in operating many tasks of the AI system. Large technology corporations such as Amazon, Google or Microsoft are all developing their own chips to reduce dependence on Nvidia.
Last month, Amazon Web Services (AWS) announced that it had signed a large contract with Meta to provide millions of self-developed AI CPUs. Amazon CEO Andy Jassy even affirmed that AWS can create AI chips capable of competing with Nvidia.
Under that pressure, Nvidia introduced the Vera CPU in March as part of its strategy to expand to the agent AI market.
According to Mr. Jensen Huang, Vera is the first CPU specifically designed for AI agents, which are systems capable of performing tasks on their own instead of humans.
Unlike traditional CPUs that focus on optimizing the ability to run multiple applications at the same time, Vera is designed to process AI tokens at the highest possible speed. This is an important factor for AI agents, which need to respond and perform tasks continuously.
Mr. Huang believes that while GPUs take on the "thinking" part of the AI model, CPUs will play a role in operating AI agents in reality.
He predicts that billions of AI agents will appear in the future acting as digital employees, each agent may need a separate computing system similar to today's personal computers.
Currently, there are about one billion human users in the world. But in the future there will be billions of AI agents and all need tools to operate," Mr. Huang said.
According to Nvidia CEO, the company has sold about 20 billion USD of independent Vera CPUs this year and is only in the early stages of the market.
Nvidia also said that major cloud service providers and many system manufacturers are cooperating to deploy Vera.
This reflects that the AI infrastructure construction race is moving into a new phase, where CPUs and GPUs are increasingly closely combined.
However, Nvidia still faces great competitive pressure. Global technology companies are investing heavily in self-developed AI chips to reduce costs and avoid dependence on a single supplier.
However, with the current growth momentum and central position in the global AI ecosystem, Nvidia is still showing ambition to expand its influence beyond the traditional GPU field.
