NVIDIA publicly thanked SpaceX and Elon Musk for being among the first clients to test its new Vera CPU processor, a custom chip designed specifically for agent-based artificial intelligence. The announcement, posted on May 18, 2026, garnered tens of millions of views within hours, highlighting the significance of SpaceX's role in validating NVIDIA's next-generation data center infrastructure.

The Vera CPU represents a major shift in how companies approach AI infrastructure. Unlike graphics processing units (GPUs), which have dominated AI training for years, the Vera CPU is optimized for orchestrating complex AI agent systems, where artificial intelligence systems can autonomously plan and execute tasks. NVIDIA's official AI Infrastructure account stated: "This is just the beginning for Vera, our CPU specifically designed for agent-based AI. Thanks to @elonmusk and the SpaceX team".

What Makes the Vera CPU Different From Traditional AI Hardware?

The Vera CPU was officially announced on January 5, 2026, as NVIDIA's first custom processor built from the ground up for data centers. The chip features specifications that set it apart from consumer and enterprise processors:

• Core Architecture: 88 custom "Olympus" cores compatible with Arm architecture, providing significant processing power for coordinating AI workloads across multiple systems
• Memory Capacity: Up to 1.5 terabytes of LPDDR5X memory, allowing the processor to handle massive datasets and complex AI models simultaneously
• Memory Bandwidth: Up to 1.2 terabytes per second of memory bandwidth, enabling rapid data movement between the processor and other components
• GPU Connectivity: Second-generation NVLink-C2C connectivity with 1.8 terabytes per second of bidirectional bandwidth between CPU and GPU, seven times faster than PCIe Gen 6 standard connections

The Vera CPU is part of the larger Vera Rubin platform, where Vera acts as the orchestration CPU and Rubin serves as the computing GPU. NVIDIA positions this combination for advanced training workloads, AI agents, analytics, cloud infrastructure, and storage applications in what the company calls "AI factories".

Why Did SpaceX Receive the First Vera CPUs Instead of xAI?

The decision to deliver Vera CPUs to SpaceX rather than to xAI, Musk's artificial intelligence company founded in 2023, reveals a strategic distinction between Musk's various enterprises. xAI operates the Colossus supercluster in Memphis, Tennessee, which utilizes tens of thousands of NVIDIA GPUs to train its Grok language models. However, SpaceX's aerospace operations present different computational challenges that align with the Vera CPU's agent-based AI focus.

SpaceX's work in autonomous rocket landing, satellite deployment, and space logistics requires real-time decision-making systems that can coordinate multiple subsystems simultaneously. These agent-based AI applications differ fundamentally from the large-scale language model training that xAI performs. The Vera CPU's architecture, optimized for orchestration and coordination rather than pure computational throughput, makes it particularly suited to SpaceX's operational needs.

During NVIDIA Vice President Ian Buck's personal delivery visit to SpaceX's Palo Alto facilities on May 15, 2026, Musk examined the system closely and specifically inquired about the processor's cores, memory capacity, and cooling system, according to NVIDIA's official blog. This hands-on evaluation underscores SpaceX's role as a technology pioneer within Musk's portfolio.

How to Understand NVIDIA's Hardware Delivery Strategy

• Multi-Client Rollout: Ian Buck's tour delivered Vera CPUs to four key clients simultaneously, including Anthropic in San Francisco, OpenAI in Mission Bay, SpaceX in Palo Alto, and Oracle Cloud Infrastructure in Santa Clara, demonstrating NVIDIA's coordinated approach to introducing new hardware to major technology companies
• Executive Engagement: Direct involvement from NVIDIA's vice president in hardware delivery signals the importance NVIDIA places on these relationships and ensures that technical specifications are properly understood by each client's engineering teams
• Public Recognition: NVIDIA's public acknowledgment of SpaceX and Elon Musk on social media platforms amplifies the credibility of the Vera CPU by associating it with a company known for pushing technological boundaries in aerospace and autonomous systems

Musk responded to NVIDIA's announcement with characteristic wordplay, tweeting "Vera nice, Vera nice..." on May 19, 2026, referencing both the processor's name and the Italian expression meaning "very good." This lighthearted exchange highlighted the collaborative relationship between the two companies while drawing additional attention to the hardware announcement.

Musk

The relationship between Musk and NVIDIA extends across multiple dimensions. SpaceX is a direct customer of NVIDIA's hardware, while xAI operates one of the world's largest GPU clusters powered by NVIDIA technology. This multifaceted partnership demonstrates how NVIDIA has positioned itself as the essential infrastructure provider for companies pursuing advanced artificial intelligence applications, whether those applications involve training massive language models or coordinating autonomous systems in real-world environments.

The Vera CPU's introduction marks a significant moment in AI infrastructure evolution. As NVIDIA's AI Infrastructure account concluded: "This is just the beginning for Vera," suggesting that the processor will play an increasingly important role in how companies deploy agent-based AI systems across their operations in the coming years.