New The CAT NeoVERSE OT-Cloud hardware line is here — from the 1U edge node to the 7U OT-OCN Giant Unit. View Catalogue ›
Home / NeoVERSE Catalogue / Full Rack · Micro-OT-Cloud Cluster
CAT NeoVERSE · Vortex OT-Cloud

Full Rack · Micro-OT-Cloud Cluster

NEO-RACK-31U

A composed sovereign rack — cluster compute, Pentascale storage and Grace-Hopper AI.

  • 3U cluster · 8 processors
  • NVIDIA GH200 Grace-Hopper
  • Pentascale NVMe storage
  • ~31U example fill

Download Datasheet   Request a demo

Full Rack Micro-OT-Cloud Cluster

A fully composed, sovereign OT-Cloud rack — combining the 3U Micro-Cloud cluster, 1U Pentascale storage, and NVIDIA Grace-Hopper AI systems into a single ~31U deployment. Cluster compute, deterministic storage and coherent-memory AI, delivered as one immune-grade unit.

Rack Building Blocks

3U Micro-OT-Cloud Cluster

8 processors
  • CPU: 8 × single Intel® Xeon® 6300/2400/2300 or AMD EPYC™ 4005/4004 (LGA-1700, up to 95W TDP)
  • Memory: up to 128GB ECC UDIMM DDR5-4400 per node (4 DIMM sockets × 8)
  • Drives: 2 × front hot-swap 3.5" SATA3 or hybrid 2.5" SAS/SATA3/NVMe U.2 per node
  • PCIe: 8 × PCIe 5.0 x8 LP + 8 × PCIe 4.0 x8 Micro-LP
  • Power: 2200W redundant Titanium high-efficiency

1U Pentascale Storage

NVIDIA Grace CPU Superchip
  • Single NVIDIA Grace™ CPU Superchip — up to 144 Arm v9 cores, 600W TDP
  • Up to 960GB LPDDR onboard coherent memory
  • 16 front hot-swap E3.S (7.5mm) NVMe bays · 8 × E1.S · 2 × M.2
  • 2 × PCIe 5.0 x16 FHHL + 1 AIOM (OCP 3.0 SFF)
  • Redundant 1600W Titanium; BlueField®-3 / ConnectX®-7 support

1U CPU+GPU Coherent Memory (AI/HPC)

NVIDIA GH200 Grace-Hopper
  • Up to 2 × NVIDIA GH200 Grace-Hopper Superchips (72-core Arm CPU + H100 Tensor Core GPU)
  • Up to 96GB HBM3 + 480GB LPDDR5X per superchip
  • NVLink® Chip-2-Chip (C2C) high-bandwidth, low-latency interconnect
  • 8 hot-swap E1.S + 2 M.2 · 2 × PCIe 5.0 x16
  • Air and liquid cooling options

Example Rack Fill (~31U)

Component
Qty
Role
3U Micro-OT-Cloud Cluster
4 units
48 cluster processors — OT-Edge, OT-Profile, OT-File, OT-Zone, OT-Cache
1U Pentascale Storage
5 units
OT-File and OT-Profile
1U Pentascale Storage (OT-AI)
6 units
Logs, profile utilization, OT-Cloud observations
1U CPU+GPU Coherent Memory
8 units
OT-AI implementation (load-dependent)

Gallery

Deploy Full Rack · Micro-OT-Cloud Cluster

Datasheet PDF Contact Sales