OpenMetal adds two private GPU server lines to its v5 hardware catalog, built around NVIDIA RTX PRO 6000 Blackwell and H200

OpenMetal said it is expanding its v5 hardware catalog with two new private GPU server offerings built on NVIDIA accelerators, a move aimed at customers that want managed access to GPU compute without relying on shared public cloud resources. The new lines are the RP6000, described as based on the NVIDIA RTX PRO 6000 Blackwell Server Edition, and the H200, described as based on the NVIDIA H200 NVL-class platform for enterprise use.

According to the announcement, the company is marketing both configurations as “private” server lines, meaning OpenMetal supplies dedicated hardware setups rather than simply pointing customers to on-demand instances. In practice, that distinction matters most for organizations with constraints around data handling, workload isolation, or procurement and operational control, though the company did not provide further details in the published post.

The RP6000 is positioned as an implementation of NVIDIA’s RTX PRO 6000 Blackwell Server Edition. OpenMetal did not outline performance targets, memory capacity, networking layout, or software stack components in the information provided through the published report, but it tied the configuration directly to NVIDIA’s RTX PRO 6000 Blackwell server-class GPU platform.

The second addition is the H200 server line, described by OpenMetal as built on NVIDIA’s H200 platform. In the same announcement, OpenMetal presented H200 as the basis for its private GPU servers, again without disclosing full system specifications in the published excerpt. The company also did not provide pricing, availability dates, or minimum contract terms in the report.

For NVIDIA, the appearance of its RTX PRO 6000 Blackwell and H200 in a third-party private hardware catalog is another announcement of how its data center GPUs are being productized through different channels. Instead of only being sold as direct server equipment or consumed through cloud providers, NVIDIA’s chips are increasingly being “wrapped” into managed infrastructure offerings by firms that focus on procurement, deployment, and operational support.

For OpenMetal, expanding a named v5 catalog suggests it is standardizing compute platforms around repeatable configurations. Customers searching for private GPU capacity typically weigh factors such as deployment timelines, reproducibility of environments, and how quickly systems can be brought online for model development and inference, alongside cost.

Still, key questions remain unanswered in the published report. OpenMetal did not disclose system-level details such as CPU pairing, storage, interconnect type, cooling approach, service-level commitments, or the exact software components included in the private server configurations. It also did not clarify whether the private servers are offered as fully managed services, bring-your-own-deployment models, or a combination.

What to watch next is whether OpenMetal follows up with additional technical documentation for the RP6000 and H200 offerings, including expected time-to-provision, supported frameworks, and any operational guarantees. Separately, the market will likely look for indicates that these configurations are being taken up by enterprises or managed service providers looking to scale GPU workloads outside the constraints of shared environments.