Nvidia has entered the CPU market with its custom-designed Vera processor, directly challenging Intel and AMD's dominance in data center chips and opening a new front in the competition for AI infrastructure. The company released detailed specifications and benchmarks on Tuesday for Vera, which it says delivers 50% better performance than traditional x86 processors for AI agent workloads. Vera chips have already been delivered to major customers including OpenAI, Anthropic, and SpaceX as of June.

Why Is Nvidia Moving Into CPUs Now?

For years, Nvidia dominated AI infrastructure through its graphics processing units, or GPUs, which became the standard chip for training and running artificial intelligence models. But the rise of agentic AI, which can operate independently in the background with minimal human oversight, has shifted the calculus. These AI agents require powerful CPUs to manage data flow and coordinate tasks between the agent and the GPU, making the CPU far more critical than it was during the early ChatGPT era.

Nvidia's move reflects a broader strategy to vertically integrate its systems, selling complete racks of computing power rather than individual chips. The company argues this approach helps customers squeeze maximum performance from their GPUs by optimizing the entire system architecture. Financial markets have taken notice; AMD and Intel stock prices have surged 128% and 149% respectively in 2026, outpacing Nvidia's 8% gain, signaling investor confidence in CPU demand.

"Agents have made CPUs much more integral, particularly how fast a CPU can answer one question," said Ian Buck, Nvidia's vice president of hyperscale, at a presentation last week.

Ian Buck, Vice President of Hyperscale at Nvidia

What Makes Vera Different From Intel and AMD Chips?

Nvidia designed Vera from the ground up specifically for AI agent workloads, rather than adapting an existing architecture. The chip prioritizes single-core speed, high memory bandwidth, and low latency, enabling AI agents to return results to GPUs as quickly as possible and keep those expensive processors fully utilized.

Traditional CPUs from Intel and AMD focused on maximizing core count, the number of processing units on a chip. Nvidia took a different approach, betting that speed on individual cores matters more for AI agents than raw parallelism. The company claims this design philosophy delivers 50% better performance for AI agent tasks compared to x86 chips, the architecture that powers Intel and AMD processors.

Vera also supports an enormous amount of low-power memory, up to 1.5 terabytes per chip, the same type used in laptops and phones. However, this power efficiency comes with a trade-off: the CPU itself is power-hungry, consuming between 250 and 450 watts of electricity per chip.

How to Evaluate Vera's Market Potential?

• Market Size Estimates: Nvidia claims the server CPU market could eventually reach $200 billion in value, though current estimates place the mature server CPU market at roughly $37 billion in 2025, suggesting significant growth potential.
• Pricing and Volume Forecasts: Wolfe Research estimated Vera chips would sell for approximately $5,000 each, with Nvidia shipping about 1.3 million units in 2026, though Nvidia declined to confirm pricing details.
• Competitive Positioning: AMD currently holds about 33% of the server CPU market and is gaining share through deep relationships with hyperscalers, while Intel maintains 66.8% market share, making Nvidia's entry into an established duopoly a significant challenge.

Gartner analyst Kevin Knox noted that AMD has built a strong ecosystem around its chips and remains the company to beat in enterprise AI server CPUs. Nvidia will need to convince cloud providers and hyperscalers to adopt Vera despite the established relationships between AMD, Intel, and their customers.

What Are the Adoption Challenges?

Despite Nvidia's market dominance in GPUs, selling CPUs presents a different challenge. The company has only publicly listed Oracle as a major cloud provider partner, though it announced that OpenAI plans to deploy Vera chips in large quantities starting this quarter. Coutand, a Vera product marketer at Nvidia, acknowledged the chip is in "early innings" of adoption.

"The CPU is something they've done to kind of unhook their customers from using Intel or AMD CPUs, and they covet that revenue. What they've done is they decided to focus on a unique CPU that isn't available in the market from anyone right now," said Karl Freund, founder of Cambrian AI Research.

Karl Freund, Founder of Cambrian AI Research

Some analysts believe Nvidia has created an entirely new class of CPU that Intel and AMD don't yet have a direct answer to. Unlike traditional server CPUs designed for web serving and general workloads, Vera is purpose-built for intense AI tasks, which could limit its addressable market but also reduce direct competition.

Nvidia is offering Vera in multiple configurations: as a standalone chip, in a liquid-cooled rack of 256 chips, as a two-chip server configuration, and paired with Nvidia's GPUs in a system called Vera Rubin. This flexibility aims to accommodate different customer needs and deployment scenarios.

The power consumption profile of Vera, ranging from 250 to 450 watts per chip, underscores the ongoing challenge of data center energy demands as AI infrastructure scales. This power draw will be a critical consideration for cloud providers evaluating whether to adopt Vera alongside their existing GPU deployments, particularly as electricity costs and grid constraints become increasingly important factors in infrastructure planning.