For teams running compute-heavy workloads, the choice has usually come down to two uncomfortable options: rent capacity from a hyperscaler and watch the bill swing unpredictably, or buy hardware outright and take on all the operational work that comes with it. OpenMetal‘s new v5 hardware lineup aims to remove that tradeoff entirely, pairing Intel’s Xeon 6000 P-series processors with DDR5-6400 memory across both bare metal servers and hosted private cloud configurations.
The memory upgrade matters more than it might initially sound. High-core-count processors often sit idle waiting on data, and faster memory keeps those cores fed instead of stalled. For workloads like AI inference, large-scale analytics, and high-performance databases, that bottleneck is frequently where real-world performance actually breaks down, not in raw processing power.
OpenMetal built v5 as three distinct tiers rather than a single flagship option, which lets teams match hardware to their actual workload instead of paying for capacity they will never use. Medium v5 offers 24 cores and 48 threads with 256GB of memory, aimed at development environments and production workloads with moderate demands. Large v5 sits in the middle and carries the highest base clock speed in the lineup at 3.2GHz, making it well suited for latency-sensitive applications where single-threaded performance matters. At the top of the lineup, the XL v5 packs 64 cores, 128 threads, and a full terabyte of memory for workloads that can’t afford to hit a ceiling.
Every tier ships with 40Gbps of private bandwidth, which matters for distributed clusters and storage replication where network throughput often becomes the hidden constraint nobody planned for.
The deployment flexibility is where OpenMetal differentiates itself most. Teams can take v5 as bare metal, getting direct hardware access with no hypervisor overhead, ideal for databases and AI inference that benefit from running close to the metal. Alternatively, the same hardware is available as a hosted private cloud, arriving as a three-node hyper-converged cluster running OpenStack and Ceph, ready within minutes rather than weeks.
Jamie Tischart, CTO of OpenMetal, framed the release around predictability. Customers need infrastructure that scales with modern AI and analytics workloads without costs scaling unpredictably alongside them.
v5 hardware is available now in Ashburn, Virginia, with pre-orders open for Los Angeles, Amsterdam, and Singapore.
