TechCrunch reported that during the latest earnings call, NVIDIA CEO Jensen Huang described the Vera CPU as a key product for the company’s next phase, directly linking it to the demand for agent AI. The article suggests that this statement goes beyond product promotion—it’s also a response to Wall Street’s ongoing concerns about NVIDIA’s growth ceiling and competitive pressures in the CPU market.
Vera is being expanded into new markets.
Jensen Huang stated that Vera is a CPU designed for agent AI, defining it as a new market that NVIDIA had not truly addressed before. He noted on the earnings call that this direction represents a potential market size of approximately $200 billion, with major hyperscale cloud providers and system vendors collaborating with NVIDIA on deployment.
According to him, the "thinking" process of AI models relies primarily on GPUs, while agents consume more CPU resources when performing tasks. If the number of agents continues to grow in the future, demand for CPUs will rise accordingly.
The market is concerned about intensifying CPU competition.
The article points out that although NVIDIA dominates the GPU market, CPUs have long been a stronger domain for Intel and AMD. Recently, one of the concerns about NVIDIA has been that cloud providers and other chip companies are accelerating their development of proprietary AI chips in an effort to undermine its advantage.
The article mentions that Amazon Web Services recently highlighted progress on its in-house AI CPU and announced a major deal with Meta. AWS CEO Andy Jassy has also repeatedly stated that AWS aims to build stronger self-reliance in both GPU and CPU technologies. These moves have increased external interest in whether NVIDIA can extend its GPU leadership into the CPU market.
Jensen Huang signals sales momentum
Huang Renxun further stated on the earnings call that Vera generated approximately $20 billion in sales this year as an independent CPU product. He used this to illustrate that the business is no longer just conceptual but has already begun generating revenue.
The article argues that Jensen Huang’s core insight is that the future global landscape will not only include billions of human users but also billions of agents. These agents will similarly require tools and computing endpoints, driving increased demand for underlying compute power and, consequently, more CPUs. Whether Vera can truly become NVIDIA’s new long-term growth engine depends on cloud providers’ procurement decisions, the speed of product deployment, and the outcome of competition with in-house chip solutions.