CoreWeave adds NVIDIA Vera CPU to its cloud portfolio

By Splitfeed EditorialLast updated Sep 30, 2026, 15:01 UTC

On September 30, 2026, CoreWeave announced it would add the NVIDIA Vera CPU to its cloud computing portfolio, offering bare-metal access at rack scale. Cognition, an AI company, was named as an early customer, specifically for CoreWeave's availability of the NVIDIA Vera Rubin NVL72 system. One day earlier, on September 29, 2026, Los Alamos National Laboratory separately received an NVIDIA Vera CPU, marking another early deployment of the chip outside the commercial cloud context.

The Vera CPU is positioned by NVIDIA as the first processor built specifically for AI agent workloads, distinguishing it from general-purpose CPUs that have historically dominated both consumer and data center markets. CoreWeave's role as a launch partner means it is among the first commercial cloud providers to offer the chip, which matters because access to specialized AI hardware at scale has become a meaningful competitive differentiator in the cloud industry. The broader question in dispute is whether a CPU purpose-built for AI agents represents a genuine architectural leap or a marketing repositioning of incremental improvements, and whether centralized cloud access to such chips concentrates AI capability in ways that raise dependency or access concerns.

None of the three camps, Pro-AI, Anti-AI, or Middle Ground, have published reactions to this story yet. Pro-AI voices would typically welcome a purpose-built AI CPU as validation that the infrastructure layer is maturing to match ambitions around autonomous agents. Anti-AI voices would likely raise concerns about the accelerating build-out of AI-specific hardware, arguing it deepens commitment to systems whose risks are not yet well understood. Middle Ground observers would probably focus on questions of who controls access to this infrastructure and whether the concentration of advanced AI compute among a small number of cloud providers and national laboratories is a dynamic worth scrutinizing.

The clearest signal to watch for next is how quickly additional cloud providers or enterprise customers announce access to the Vera CPU, and whether NVIDIA or CoreWeave release performance benchmarks that substantiate the claim that the chip meaningfully outperforms existing alternatives for AI agent workloads. Pricing and availability details at rack scale would also help clarify whether this offering is broadly accessible or effectively limited to well-resourced customers.