OpenMetal GPU Servers are expanding with new NVIDIA-powered systems designed for artificial intelligence, machine learning, high-performance computing, and enterprise GPU workloads.
OpenMetal has added two private GPU server lines to its v5 hardware catalog: the RP6000 and the H200. The RP6000 uses NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, while the H200 is built around NVIDIA H200 NVL GPUs.
The launch gives customers access to dedicated, single-tenant GPU infrastructure with predictable monthly billing. It also strengthens OpenMetal’s v5 platform, which is based on Intel Xeon 6000-series processors, DDR5-6400 memory, PCIe 5.0 connectivity, NVMe storage, and 40Gbps private networking.
OpenMetal GPU Servers Target AI and HPC Demand
OpenMetal GPU Servers are aimed at teams running demanding AI, ML, and HPC workloads.
The company said the new systems are built for customers who need full hardware control rather than shared cloud resources. Each server is delivered as bare metal infrastructure, meaning there is no shared hypervisor between the customer’s workload and the machine.
This setup is useful for organizations that need predictable performance, stronger isolation, and direct access to GPU power.
OpenMetal GPU Servers Add RTX PRO 6000 Blackwell Power
The RP6000 server supports one or two NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs.
Each GPU includes 96GB of GDDR7 memory, 1.79 TB/s of memory bandwidth, and 24,064 CUDA cores. OpenMetal said this configuration is suited for inference, model fine-tuning, rendering, visualization, and mixed AI production pipelines.
The large VRAM capacity makes the RP6000 attractive for teams that need strong GPU performance without moving to a fully shared public cloud model.
OpenMetal GPU Servers Bring H200 for Large AI Models
The H200 server line is designed for larger and more memory-heavy AI workloads.
Each NVIDIA H200 NVL PCIe GPU offers 141GB of HBM3e memory and 4.8 TB/s of memory bandwidth. The servers can support one or two H200 GPUs per system.
OpenMetal said the H200 option is built for large-model training, memory-bound inference, and advanced enterprise AI deployments. Each H200 server also includes a five-year NVIDIA AI Enterprise software subscription.
OpenMetal GPU Servers Support Popular AI Frameworks
OpenMetal GPU Servers can run common AI and machine learning tools, including PyTorch, TensorFlow, JAX, and Hugging Face Transformers.
Customers can deploy the servers as standalone systems or as part of multi-node GPU clusters over OpenMetal’s private network. This gives teams flexibility when scaling AI training, inference, and technical computing workloads.
The new servers also match the wider v5 hardware catalog, allowing customers to build CPU, memory, storage, and GPU environments on the same modern platform.
Why OpenMetal GPU Servers Matter
OpenMetal GPU Servers arrive as businesses seek more control over AI infrastructure costs and performance.
Many companies want dedicated GPU capacity but do not want unclear metered pricing or shared cloud limitations. OpenMetal is positioning the new RP6000 and H200 systems as an alternative for customers that want private infrastructure with transparent monthly pricing.
The RP6000 and H200 servers are available now in OpenMetal’s US East data center in Ashburn, Virginia. Additional locations are planned, and proof-of-concept deployments are available for teams that want to test workloads before committing.
With demand for AI infrastructure continuing to grow, OpenMetal GPU Servers give enterprises, SaaS companies, hosting providers, and managed service providers another option for building private, high-performance AI and HPC environments.
