Nvidia Corp said that Anthropic, OpenAI and SpaceX are among the first major customers for its upcoming microprocessor, marking an early win as it expands deeper into AI data center infrastructure.

At the Computex Taipei conference, Nvidia CEO Jensen Huang highlighted the companies during his presentation, saying they will be among the first to deploy Nvidia’s Vera CPUs in their data centers.

The chip is scheduled to enter full production in the third quarter of this year.

The move underscores Nvidia’s push to reassure investors and customers that it is adapting to shifts in AI computing, as workloads gradually move from model training toward inference and service delivery, areas where general-purpose CPUs from rivals like Intel Corp and Advanced Micro Devices Inc, as well as in-house chips from firms like Amazon, are increasingly important.

Huang said Vera delivers up to 1.8 times the performance of traditional x86-based chips on key AI workloads, positioning it as a direct competitor in a market long dominated by Intel and AMD.

He also reiterated Nvidia’s claim that it offers a full-stack AI data center platform, combining hardware and software designed for rapid deployment.