The flagship sovereign AI and research-computing platform for organisation-scale shared services and the largest Hyperion workloads.
This brief explains where the baseline fits, the software and operating model it supports, and the evidence Axiotech should use to validate a final configuration. Indicative specifications remain subject to component availability, export compliance and formal quotation.
Baseline architecture
Four 4U nodes; 512 AMD EPYC cores; 4.6TB ECC DDR5; sixteen NVIDIA RTX PRO 6000 96GB GPUs; more than 1.5TB aggregate VRAM; 100GbE compute and dedicated storage fabrics.
The current compute node is based on the Supermicro AS-4125GS-TNRT / CSE-418G2TS 4U rack platform, integrated with matched rails, power, management, networking and rack infrastructure as required.
Best-fit workloads
Organisation-scale sovereign LLM/RAG, multi-team AI, distributed fine-tuning, large research models and mixed long-running services plus scheduled campaigns across sixteen 96GB GPUs.
Four nodes allow maintenance-aware service placement, large distributed runs and multi-team capacity under one engineered rack and support model.
Recommended software stack
Kubernetes and/or Slurm, NVIDIA AI Enterprise components where licensed, CUDA/NCCL, Triton, TensorRT-LLM, vLLM, PyTorch, MLflow, observability, identity, policy and protected shared storage.
Pin host drivers and infrastructure separately from versioned application containers. Record source, image, dataset/model and hardware allocation with each benchmark or production release.
Deployment pattern
Design an operating model before commissioning: Kubernetes and/or Slurm boundaries, tenant identity, policy, signed images, shared storage, backup, 100GbE fabric, observability and audited model/data lifecycle.
Define monitoring, identity, backup, change control and workload ownership at the same time as compute. Multi-user platforms require resource allocation and quotas; production services require health, overload and rollback behaviour.
Sizing boundary
This is a platform programme, not merely a server purchase. Workload admission, data governance, export review, backup, power/cooling and operations ownership must be agreed before commissioning.
Final sizing should use representative code, data, concurrency and service objectives. Aggregate core, RAM or VRAM figures do not by themselves predict application performance.
Commissioning and acceptance
Validate all-node soak and distributed jobs, service failover, scheduler isolation, model/data permissions, backup/restore, power/cooling headroom, change rollback and workload-specific evaluation under peak concurrency.
Axiotech should retain the resulting configuration, firmware/driver baseline, environment manifest, benchmark data and recovery procedure as the system acceptance pack.
Primary technical references
- NVIDIA Container Toolkit overview
- Prometheus overview
- Slurm Quick Start Administrator Guide
- NVIDIA CUDA C++ Best Practices Guide
- NVIDIA NCCL documentation
References are provided for software architecture and implementation planning. Validate the versions, licences, support matrix and regulated-use requirements applicable to the final deployment.