Nvidia Ships First Vera CPUs to Anthropic, OpenAI, SpaceXAI and Oracle Cloud

Why It Matters

Vera represents Nvidia’s first foray into custom server‑CPU design, a strategic move that could diversify the company’s revenue beyond GPUs and networking. By targeting agentic AI—a class of models that require continuous reasoning, tool use and orchestration—Nvidia is addressing a growing compute bottleneck that traditional CPUs struggle to meet. If successful, the chip could accelerate the deployment of autonomous agents in fields ranging from autonomous vehicles to scientific simulation, expanding the overall AI market. The commitment from Oracle Cloud to deploy hundreds of thousands of Vera CPUs signals that cloud providers view the CPU as a critical lever for next‑generation AI services. This could force rivals like AMD, Intel and emerging Chinese vendors to accelerate their own AI‑centric CPU programs, intensifying competition and potentially driving down costs for enterprises seeking end‑to‑end AI infrastructure.

Key Takeaways

• Nvidia delivered its first Vera CPUs to Anthropic, OpenAI, SpaceXAI and Oracle Cloud in a coordinated rollout.
• Vera features 88 custom Olympus cores, 1.2 TB/s memory bandwidth and ~50 % higher per‑core performance under sustained load.
• Oracle Cloud plans to deploy hundreds of thousands of Vera CPUs starting in 2026, the first large‑scale cloud commitment.
• Ian Buck highlighted the rising CPU demand for AI models that generate code and manage multi‑step tasks.
• The launch marks Nvidia’s entry into custom server‑CPU market, challenging AMD and Intel’s dominance.

Pulse Analysis

Nvidia’s Vera rollout is more than a product launch; it’s a strategic bet that the next wave of AI will be defined by agents that need continuous, low‑latency compute beyond what GPUs can provide. By bundling a custom CPU with its existing GPU and networking portfolio, Nvidia is positioning itself as a one‑stop shop for full‑stack AI infrastructure. This could lock in customers who prefer a tightly integrated ecosystem, reducing the friction of mixing third‑party CPUs with Nvidia GPUs.

Historically, Nvidia’s strength has been in accelerating matrix‑multiply workloads, while CPUs have remained the domain of Intel and AMD. Vera blurs that line, suggesting a future where the distinction between “CPU” and “accelerator” becomes less relevant. If the performance and efficiency claims hold up in real‑world data centers, Nvidia could capture a new revenue stream that offsets the cyclical nature of GPU demand. However, the success of Vera will hinge on pricing, software tooling, and the ability to scale production without supply‑chain hiccups that have plagued the broader semiconductor industry.

Competitors are unlikely to sit idle. Intel’s Xeon Scalable line and AMD’s EPYC processors are already being tuned for AI workloads, and Chinese firms such as Alibaba are unveiling their own AI‑focused CPUs. The coming months will likely see benchmark battles and early‑adopter case studies that will determine whether Vera becomes the de‑facto CPU for agentic AI or remains a premium niche offering for the most demanding customers.