Fujitsu CELSIUS R970 Power Workstation

The Fujitsu CELSIUS R970 is a dual-socket rack-mounted professional workstation engineered for the most demanding computational environments. Built on the Intel Xeon Scalable processor platform, the R970 supports two high-core-count processors simultaneously, enabling massive parallel processing throughput suitable for AI model training, large-scale numerical simulation, and real-time rendering pipelines. The chassis accommodates substantial memory expansion, supporting registered ECC DDR4 memory across multiple channels to ensure data integrity under sustained enterprise workloads.

Designed with professional GPU integration at its core, the CELSIUS R970 provides multiple full-length PCIe expansion slots capable of hosting NVIDIA RTX or Quadro-class graphics cards alongside high-bandwidth compute accelerators. This architecture allows organisations to deploy the system as a centralised AI development node, a physics simulation engine, or a high-fidelity visualisation server accessible remotely via virtualisation and remoting technologies. The system's power delivery is handled by redundant hot-plug power supply units, minimising unplanned downtime in mission-critical rack environments.

The R970 integrates Fujitsu's iRMC (integrated Remote Management Controller) for out-of-band server lifecycle management, enabling remote monitoring, BIOS configuration, firmware updates, and power control without requiring physical access to the hardware. With support for Fujitsu ServerView Suite management software and compatibility with standard IPMI and Redfish interfaces, the CELSIUS R970 slots naturally into existing enterprise IT management frameworks. Its rack-optimised 4U form factor balances density with the thermal headroom needed to sustain dual high-TDP processors and multiple GPU accelerators under continuous workload conditions.

Ideal for

• AI and deep learning model training using dual CPU parallelism combined with multi-GPU acceleration for research and enterprise data science teams
• Computational fluid dynamics and finite element analysis simulations in automotive, aerospace, and industrial engineering environments
• High-resolution 3D rendering and visual effects production requiring sustained GPU throughput across extended render sessions
• Medical imaging processing and DICOM data analysis workflows demanding ECC memory integrity and certified ISV application support
• Seismic data interpretation and geoscientific modelling for energy sector organisations requiring high floating-point compute density
• Centralised remote workstation deployment serving multiple concurrent users via GPU virtualisation in VDI and cloud workstation architectures

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.