Related videos:
NVIDIA publicly thanked SpaceX and Elon Musk for being one of the first clients to test its new processor Vera CPU, designed for agent-based artificial intelligence, in a post on X that garnered tens of millions of views within a few hours.
The message, posted on Monday, May 18, by NVIDIA's official account, stated: "Thanks @SpaceX and @elonmusk, excited for you to try out the NVIDIA Vera CPU." The company's AI infrastructure account, @NVIDIAAIInfra, expanded on the announcement with another post: "This is just the beginning for Vera, our CPU specifically designed for agent-based AI. Thanks to @elonmusk and the SpaceX team."
Musk responded this Tuesday with a play on words: “Vera nice, Vera nice...”, referencing the name of the processor and the Italian expression meaning “very good.”
The delivery to SpaceX was part of a tour conducted on Friday, May 15, by Ian Buck, vice president of NVIDIA, who personally delivered the first units of Vera CPU to four key clients: Anthropic in San Francisco, OpenAI in Mission Bay, SpaceX in Palo Alto, and Oracle Cloud Infrastructure in Santa Clara, where the tour concluded on Monday, May 18.
During the visit to the SpaceX facilities in Palo Alto, Musk examined the system and specifically inquired about the cores, memory, and cooling system of the processor, according to the official NVIDIA blog on delivery.
The Vera CPU was officially announced on January 5, 2026 as NVIDIA's first custom processor for data centers. It features 88 custom "Olympus" cores compatible with Arm architecture, up to 1.5 TB of LPDDR5X memory, up to 1.2 TB/s of memory bandwidth, and second-generation NVLink-C2C connectivity with 1.8 TB/s of bidirectional bandwidth between CPU and GPU, seven times faster than PCIe Gen 6.
The processor is part of the Vera Rubin platform, NVIDIA's next generation of AI systems for data centers, where Vera acts as the orchestration CPU and Rubin as the computing GPU. NVIDIA positions this chip for advanced training workloads, AI agents, analytics, cloud, and storage in what the company refers to as "AI factories."
The relationship between Musk and NVIDIA is close and multifaceted: SpaceX is a direct customer of the company's hardware, while xAI —the artificial intelligence company founded by Musk in 2023— operates the Colossus supercluster in Memphis, Tennessee, utilizing tens of thousands of NVIDIA GPUs to train its Grok models. The delivery of Vera CPUs to SpaceX, rather than to xAI, positions the aerospace company as a pioneer in adopting NVIDIA's new generation of CPU infrastructure for agent-based AI.
The NVIDIA AI Infrastructure account summarized the significance of the moment with a phrase aimed at the future: "This is just the beginning for Vera."
Filed under: