CoreWeave launches NVIDIA Vera Rubin NVL72 cloud service at production scale

By Splitfeed EditorialLast updated Oct 1, 2026, 08:00 UTC

On September 30, 2026, CoreWeave announced that NVIDIA's Vera Rubin NVL72 GPU system is now available as a production-scale cloud service. AI company Cognition was identified as the first customer to use the new offering. Alongside the hardware launch, CoreWeave also unveiled related software and disclosed plans for a separate Vera CPU cloud service designed specifically for AI agent workloads. CoreWeave's stock rose on the day of the announcements.

The significance of this launch lies in what it represents for the broader AI infrastructure market. The Vera Rubin NVL72 is NVIDIA's latest generation GPU system, and CoreWeave's deployment marks its first appearance in a commercial cloud environment at scale. That distinction matters because access to cutting-edge compute has become a central competitive variable for AI developers, and whoever brings new hardware to market first as a managed service gains leverage over customers who cannot or do not want to build their own infrastructure. The planned Vera CPU service adds another layer, signaling that CoreWeave is positioning itself not just for model training workloads but for the emerging and fast-growing category of AI agent deployment, which has different performance and cost characteristics than training.

None of the three camps, Pro-AI, Anti-AI, or Middle Ground, have published reactions to this story yet. Pro-AI commentators would typically treat this as straightforward evidence of rapid and healthy progress in AI infrastructure, pointing to faster hardware access and growing commercial adoption as validation of the technology's trajectory. Anti-AI voices would likely raise concerns about the concentration of power among a small number of compute providers and the energy and resource demands that come with deploying systems at this scale. Middle Ground observers would probably acknowledge the genuine capability advances while pressing for more transparency around pricing, access equity, and the environmental costs of expanding GPU cloud capacity.

The details worth watching going forward include how quickly other cloud providers match or announce competing access to Vera Rubin hardware, whether Cognition publicly reports performance or cost outcomes from its use of the service, and what pricing and availability terms CoreWeave sets for the planned Vera CPU agent-focused offering.