DRACO 8× H200 SXM GPU Server

Overview

The DRACO 8× H200 SXM GPU Server is a Rack 8U rack server built to bring training and high-throughput inference of the largest models into your own data centre. Eight NVIDIA H200 SXM5 (HGX H200) GPUs on a single HGX baseboard deliver 1,128 GB HBM3e of HBM3e, linked by NVLink and NVSwitch into one tightly-coupled accelerator — sized to train and serve up to 405B-class models, behind your firewall, in INR, on a GST invoice.

Engineered for AI platform teams bringing large-model training in-house, it pairs the eight GPUs with a 2× Intel Xeon 6 host, 2 TB DDR5 ECC and 60 TB NVMe, with 8× 400G InfiniBand for scale-out, redundant power and full BMC/IPMI — a data-centre training node that racks and runs.

Key highlights

• 1,128 GB HBM3e of HBM3e across 8× H200 SXM — train and serve up to 405B-class models on a single node.
• NVLink + NVSwitch fabric — full all-to-all GPU bandwidth for efficient tensor-parallel training, the defining advantage of an HGX node.
• 2× Intel Xeon 6 + 2 TB DDR5 ECC — high core count and memory bandwidth to feed eight data-centre GPUs.
• Rack 8U, liquid-cooled, redundant PSU, BMC/IPMI — sustained all-GPU clocks, hot-swap drives, lights-out management.
• 60 TB NVMe + 8× InfiniBand NDR 400G — large local dataset/checkpoint capacity with a high-bandwidth scale-out fabric; no egress fees.
• On-prem data sovereignty — training data and weights stay in-house; DPDP-friendly, air-gappable.
• Make-in-India OEM — predictable INR pricing, GST tax invoice (HSN 8471), pan-India onsite support, GeM-procurable.
• Scale path — multiple HGX nodes scale out over InfiniBand into RDP rack-scale systems and superclusters.

AI workload fit (what it actually runs — honestly)

• Training & fine-tuning: full pre-training/fine-tuning and parameter-efficient (QLoRA/LoRA) training of 405B-class models, tensor- and pipeline-parallel across the eight NVSwitch-linked GPUs.
• Inference: serve 405B-class models at high throughput, or host several large models concurrently.
• RAG, vision, multimodal & agentic: production pipelines on the 60 TB NVMe and multi-agent back-ends.
• Engineering note: the 8 SXM GPUs sit on an HGX baseboard with NVLink + NVSwitch — full all-to-all GPU bandwidth for efficient tensor-parallel training of the largest models, exactly what frontier-scale training needs.

AI workload positioning

This sits at the train-and-serve stage of the AI lifecycle, at data-centre scale. With 1,128 GB HBM3e of HBM3e on an NVLink+NVSwitch fabric, a 2× Intel Xeon 6 host and 400G InfiniBand, it is sized to sustain real large-model training and production serving — where renting eight H200-class cloud GPUs around the clock becomes the dominant line in an AI budget.

Industry use cases

• BFSI — train and serve private large models for risk, fraud and document intelligence under data-residency rules.
• Healthcare & life sciences — on-prem training of medical and imaging models, PHI in-house.
• Government & PSU — sovereign large-model training on GeM-procurable infrastructure.
• Neocloud / AI providers — a node to build a GPU cloud or training service.
• Telecom & manufacturing — large-scale model development near operations.
• Research & higher-ed — an institutional large-model training node.

Performance — and how to be sure

We don’t publish inflated peak numbers. The honest picture: 1,128 GB HBM3e of HBM3e across eight NVSwitch-linked GPUs is sized to train/fine-tune and serve up to 405B-class models on-prem. Want certainty? Request a free benchmark of your model and dataset on this exact configuration before you buy — we’ll send back real tokens/sec and training/fine-tune timings for your workload.

Series & upgrade path

• DRACO (flagship training tier) — this.
• GPU-count ladder: 4-GPU → 8-GPU within the line; step up to Blackwell B200/B300 HGX nodes for higher memory and the largest models.
• When to step up: for multi-rack scale, move to RDP Rack-Scale AI Systems and AI SuperClusters — talk to an architect about the fabric.

On-prem vs cloud — the TCO case

For sustained training and inference, owning beats renting: eight continuously-running H200-class cloud GPUs add up fast, and on-prem removes egress fees and keeps data and weights in-house. RDP pricing is fixed in INR with a GST input-credit-eligible invoice — ask for a 3-year TCO comparison against your current cloud spend.

Software & day-one readiness

Ships pre-configured to train and serve: NVIDIA driver, CUDA, cuDNN, NCCL, the InfiniBand stack, Docker and NVIDIA Container Toolkit, with PyTorch, vLLM / Triton / TensorRT-LLM on Ubuntu LTS. Optional Slurm/Kubernetes, managed AI-stack and observability setup available.

Power, cooling & rack integration

An Rack 8U liquid-cooled node with redundant PSUs and substantial power draw — this node is liquid-cooled; plan CDU/manifold and facility water, plus InfiniBand cabling. (Exact PSU rating, BTU, flow, fabric cabling and rack-depth figures confirmed on the build sheet.) BMC/IPMI provides remote power, console and health monitoring.

Deployment, warranty & support

• Made to order, built, racked-and-stacked, cabled and burned-in in India; realistic lead time confirmed at quote.
• In the box: server, rails, power cables, quick-start, and the pre-installed AI software stack; fabric switches and cabling scoped at quote.
• Onsite warranty + AMC with pan-India coverage and an RMA/escalation path (exact term & response window confirmed at quote).

Why RDP

14 years of Make-in-India infrastructure and 300,000+ devices shipped. Indian OEM, INR pricing, GST tax invoice (HSN 8471), pan-India onsite engineers, GeM availability, and DPDP / sovereign-AI-ready deployment.

Buy with confidence

This is a data-centre training-and-serving server, made to order — talk to an RDP solution architect, get a configuration and 3-year TCO tailored to your workload, and benchmark your own model on it before you commit. Request a quote to begin.

Specifications