Nvidia Unveils Vera AI CPU, Taking Aim at AMD And Intel In The Battle For Data Centers

Nvidia is expanding beyond the graphics processors that fueled its rise to become the world's most valuable company, setting its sights on another critical piece of artificial intelligence infrastructure.

The company has revealed new technical details about its next-generation Vera central processing unit, a move that positions the chipmaker to challenge longtime server CPU leaders AMD and Intel in the rapidly growing AI data center market.

The announcement marks a significant milestone in Nvidia's broader strategy to control more of the hardware powering artificial intelligence. While the company's graphics processing units, or GPUs, remain the backbone of AI training and inference, Nvidia believes the CPU is becoming increasingly important as AI systems evolve toward autonomous "agentic AI" capable of handling complex tasks with minimal human oversight.

According to the company, the first Vera processors were delivered to customers in June, including OpenAI, Anthropic and SpaceX. The company also disclosed detailed specifications, architectural information and performance benchmarks that potential buyers typically require before deploying new processors in production environments.

Unlike previous server processors that relied on Arm's off-the-shelf core designs, Nvidia said Vera is the first server CPU it has developed from the ground up. The chip was engineered specifically to eliminate bottlenecks that slow AI agents, with the company claiming up to 50% better performance for AI agent workloads than traditional x86 processors from Intel and AMD.

Nvidia executives argue that the shift reflects changing demands inside AI data centers. While earlier AI systems primarily relied on CPUs to manage a handful of GPUs, next-generation AI agents require processors capable of rapidly feeding information to accelerators and responding to millions of independent requests.

"Agents have made CPUs much more integral," Ian Buck, Nvidia's vice president of hyperscale, said during a company presentation, according to CNBC. "Particularly how fast a CPU can answer one question."

Rather than focusing on maximizing the number of processor cores, Nvidia said Vera emphasizes high single-core performance, memory bandwidth and low latency, enabling AI agents to return workloads to GPUs more quickly and keep those expensive accelerators operating at maximum efficiency.

The company is also integrating Vera into its broader AI infrastructure strategy. The processor will be sold as a standalone chip, as part of dual CPU server configurations, in liquid-cooled racks containing 256 Vera processors, and alongside Nvidia's upcoming Rubin GPUs in the company's Vera Rubin AI systems.

The launch also represents Nvidia's latest effort to vertically integrate its AI platforms. Instead of selling individual chips, the company increasingly packages complete racks of computing hardware that combine CPUs, GPUs, networking equipment and software into fully optimized AI systems.

That approach could reshape the competitive landscape. Before the AI boom, CPUs represented the most valuable component inside enterprise servers. Although Nvidia's GPUs shifted that balance after ChatGPT's debut in 2022, the resurgence of AI agents has renewed investor interest in processors.

AMD and Intel have benefited from that trend. Their shares have surged 128% and 149%, respectively, in 2026, significantly outperforming Nvidia's more modest 8% gain during the same period.

Despite Nvidia's momentum, breaking into the server CPU market will not be easy. Gartner analyst Kevin Knox told CNBC that AMD currently leads innovation in enterprise AI server processors and has built strong relationships with hyperscale cloud providers. AMD controls roughly one-third of the server CPU market, while Intel still maintains nearly two-thirds, although AMD continues gaining ground.

Industry analysts believe Nvidia is carving out a specialized niche rather than directly replacing traditional server processors. Cambrian AI Research founder Karl Freund said Vera is designed for intensive AI workloads rather than conventional enterprise computing such as hosting websites or business applications.

OpenAI plans to begin deploying Vera processors at scale this quarter, but Nvidia acknowledged adoption remains in its early stages. Oracle was the only major cloud provider listed among launch partners, suggesting the company still faces work convincing hyperscalers to move away from Intel and AMD platforms.