Nvidia Announces Major AI Firms as Early Adopters of New Vera Processor

Nvidia has revealed that leading AI developers Anthropic, OpenAI, and SpaceX will be among the first major customers for its upcoming Vera microprocessor, securing key early adopters for its latest push to expand its already dominant presence in AI data centers.

Company co-founder and CEO Jensen Huang named these firms during a presentation at the Computex conference in Taiwan, stating they will be early implementers of the Vera central processing units in their data center operations.

The new processor is scheduled to enter full-scale production in the third quarter of this year.

This announcement is part of Nvidia's concerted effort to demonstrate to customers and investors that it is not only aware of evolving trends in AI data center technology but is actively leading them.

As the industry's focus shifts increasingly from training AI models to deploying and running AI services, there has been growing discussion about the use of more general-purpose CPUs and speculation that Nvidia's specialized AI accelerators might become less critical.

The Vera chip represents Nvidia's first standalone data center microprocessor designed to compete directly with established products like Intel's Xeon processors, AMD's Epyc chips, and in-house solutions such as Amazon's Graviton.

Last month, Huang asserted that Nvidia continues to gain market share with customers like Amazon, despite their efforts to develop independent component solutions.

He maintains that Nvidia is uniquely positioned as the only company providing a full suite of integrated data center components, engineered for rapid deployment even by operators with limited technical expertise.

Huang claimed that the Vera processor delivers performance that is 1.8 times faster on key AI workloads compared to chips based on the industry-standard x86 architecture from Intel.

This marks the first time Nvidia has publicly issued a direct performance challenge against the established benchmarks in this segment.

In a related development, Nvidia has also updated and expanded its software suite for planning, deploying, and monitoring data center systems.

The company stated that its open-source DSX platform offers flexible adoption, allowing users to implement as much or as little of the software as needed.

Nvidia claims this new software enables significantly more efficient management and monitoring of data center power consumption.

The company asserts this efficiency could allow data centers to utilize up to 40% more of Nvidia's accelerator chips within the same power budget, representing a substantial operational advantage.

Furthermore, Nvidia is introducing new high-end workstation computers developed in partnership with PC manufacturers.

The Nvidia DGX Station for Windows is designed to assist companies that rely on the Microsoft Windows ecosystem for developing and deploying AI software.

Dell Technologies and other computer makers are expected to begin sales of these systems in the fourth quarter of this year.

In a move to stimulate the emerging market for humanoid robots, Nvidia announced a collaboration with Chinese robotics company Unitree to mass-produce advanced humanoid machines.

The goal is to accelerate research and development in robotics technology and produce devices that are practical for real-world applications.

Nvidia noted that research labs currently spend excessive time configuring and fine-tuning prototype "Frankenrobots" before they can be used for substantive research.

The product of this partnership will feature sophisticated five-fingered hands and will come with integrated chips and software, ensuring it is operational immediately upon unboxing.