Nvidia has entered the Central Processing Unit (CPU) market with its new data center product, Vera. This move positions the chipmaker as a direct competitor to long-established industry leaders Advanced Micro Devices (AMD) and Intel within the rapidly expanding field of artificial intelligence servers.
Strategic Shift into CPU Manufacturing
Nvidia originally achieved its status as a major tech powerhouse due to overwhelming demand for its Graphics Processing Units (GPUs), which are foundational chips in AI development. However, the company is now actively manufacturing and promoting its own CPUs—the Vera line—which cloud service providers may choose to use instead of those from AMD or Intel.
The release included detailed specifications, architectural data, and performance benchmarks necessary for potential customers to thoroughly evaluate the chip. Representatives reported that Vera chips were delivered to major clients, including OpenAI, Anthropic, and SpaceX, in June.
Vertical Integration and Market Opportunity
Nvidia has historically been recognized as an industry trendsetter. By expanding into CPUs, however, it is challenging two deeply entrenched players: Intel and AMD. This investment reflects Nvidia’s broader strategy of vertical integration—aiming to control more aspects of its systems from chip design through the entire computing rack.
Instead of simply selling discrete chips, the company plans to sell complete computational racks. According to Nvidia, this approach will enable engineers to extract greater performance from their GPUs, helping the company maintain its position as a preferred choice for leading AI research laboratories amid increasing competition from custom silicon and rivals like AMD.
While CPUs were once critical and expensive components in servers, the focus has recently shifted toward GPUs. However, the emergence of “agentic AI”—systems capable of running independently with minimal human intervention—has renewed attention on the CPU’s role in managing data flow to these agents. Ian Buck, Nvidia’s vice president of hyperscale, noted that agents have made CPUs “
much more integral
,” particularly concerning how quickly a processor can respond to a single query.
Performance and Market Projections
The financial market has reacted positively to this shift; in 2026, AMD’s stock rose by 128%, while Intel’s saw gains of 149%, significantly outperforming Nvidia’s increase of 8%.
- Market Size: Nvidia suggested the total server CPU market could eventually reach $200 billion. This contrasts with an earlier Bernstein estimate, which placed the mature server CPU market value at about $37 billion in 2025.
- Analyst Forecasts: Wolfe Research projected that the average selling price of a Vera chip would be around $5,000 and anticipated Nvidia shipping approximately 1.3 million units this year.
Technical Advantages of the Vera Chip
Nvidia highlighted several technical advantages with the Vera line. The company stated that Vera is the first server CPU it has designed entirely from its own architecture, rather than utilizing an off-the-shelf design provided by Arm.
The chip was engineered specifically to address bottlenecks inherent in AI agents. Nvidia claimed that Vera offers 50% better performance for AI agents compared to traditional x86 chips used by Intel and AMD. This improvement stems from the fact that previous CPUs emphasized core count, while Vera focuses intensely on single-core speed, high memory bandwidth, and low latency.
Hannah Coutand, a Vera product marketer at Nvidia, said in a presentation that the chip focused on per-core speed, high memory bandwidth, and latency, “so that agents can return to their GPUs as quickly as possible and keep those GPUs, which are a very expensive and a highly valuable asset in the AI factory, as highly utilized as possible.”
Deployment Options and Competitive Landscape
The Vera chip is available for purchase both standalone and paired with Nvidia’s GPUs within a system named Vera Rubin. Furthermore, it can be deployed in liquid-cooled racks containing 256 Vera chips, or configured with two Vera chips in a single server.
Technically, the CPU draws between 250 watts and 450 watts of power and supports substantial low-power memory capacity, up to 1.5 terabytes per chip.
Regarding competition, Kevin Knox, an analyst from Gartner, noted that while AMD reportedly holds about 33% of the server CPU market and Intel commands 66.8%, AMD has been successfully increasing its share and maintaining strong relationships with cloud providers. He stated, “
AMD’s done a great job building their ecosystem around their chips
.”
However, Nvidia representatives indicated that Vera is still in the “early innings” of adoption. Although OpenAI plans to deploy large quantities of Vera chips starting this quarter, the company has limited its list of partners publicly.
