Request a quoteBuy GPUs
HYPERCUBE
HYPERQ
GPUs
Astra Cybercity
Sectors
Press Room
Company

info@worldoneindia.netWhatsApp +91 93244 08321

Modular and factory scale

Core Cell C32

A private AI cloud in twenty feet.

Four eight-GPU nodes on one liquid loop, with fabric, storage and protected power: an AI cloud of its own for a bank, an insurer, a hospital group or a manufacturer. Blackwell Ultra today, Vera Rubin in the same layout.

Illustrative exterior rendering of Core Cell C32, shown in full.
32Blackwell Ultra B300 GPUs
576 PFFP4 sparse; 1.6 EF NVFP4 in the Vera Rubin edition
8.4 TBHBM3e across four node domains
60 to 90 kWIT design envelope, liquid-cooled

StatusBudgetary architecture; HGX B300 shipping, node quote required. Vera Rubin edition: production ramp, allocation and certification.

Questions first? Ask the team about Core Cell C32

A closer look.

One twenty-foot module and one cooling package, on one liquid loop.

Behind headquartersA private AI cloud for a bank, an insurer or a hospital group, placed behind its own building and kept under its own control.
Four nodes, one loopTwo compute racks holding four eight-GPU nodes, one fabric and storage rack, a coolant distribution unit and protected power, all on one liquid loop.
Inside the private cloudThe coolant unit in the foreground, then two compute racks of two eight-GPU nodes each, the fabric rack, the storage rack and the power cabinet by the door.

Styling concepts from the HYPERCUBE Product Family Design Collection, Edition 3. Configuration not finalised; renders do not depict a supplied configuration.

Two editions.

Blackwell Ultra today and Vera Rubin next, in the same module and the same layout.

SpecificationB300 editionVera Rubin edition
GPUs32 Blackwell Ultra B30032 Rubin
Nodes4 HGX B3004 HGX Vera Rubin NVL8
CPUs4 dual-socket x86 hosts4 Vera CPUs: 352 Arm cores
FP4 inference576 PF sparse1.6 EF NVFP4
FP4 dense or training432 PF dense1.08 EF NVFP4 training
FP8144 PF dense; 288 PF sparse520 PF
GPU memoryAbout 8.4 TB HBM3e: 2.1 TB per node8 TB HBM4: 2 TB per node at 130 TB/s
Scale-upNVLink 5: 14.4 TB/s per node; four domainsNVLink 6: 24 TB/s per node; four domains
Scale-out1.6 TB/s per node; InfiniBand or Spectrum-X1.8 TB/s per node
Storage0.25 to 1.0 PB usable NVMe0.25 to 1.0 PB usable NVMe
IT power60 to 90 kWPlant sized for about 120 kW, planning
CoolingDirect liquid, mandatory at 1,400 W per GPUDirect liquid at 45 °C inlet
StatusHGX B300 shipping; node quote requiredProduction ramp; allocation and certification

The 32 GPUs are four separate NVLink domains, not one. NVIDIA has not published HGX Vera Rubin NVL8 node power; the 120 kW plant allowance is a planning figure. Vera Rubin figures are NVIDIA's system values for HGX Vera Rubin NVL8. Reference figures from the HYPERCUBE Product Family Design Collection, Edition 3, and NVIDIA published platform data. They are budgetary and indicative, not a supplier quotation. Final configuration, pricing, availability and delivery are confirmed by quotation for a named country and end user. Images are styling concepts.

What goes in.

One heat-rejection package beside the module, on the coolant loop. Site power and water are site-interface packages.

Modules and packages

1 Twenty-foot module
6.06 × 2.44 × 2.90 m
1 Heat-rejection package
Dry cooler, 2.5 m long, beside the module

What the site provides

  • Low-voltage supply for the module and its cooling package
  • Fibre uplink to the customer network
  • No water in operation: a closed liquid loop to a dry cooler
  • Concrete pad for the module and the package, crane access
ItemQtyWhat it is
Eight-GPU compute nodesN1 to N44HGX B300 with x86 hosts, or HGX Vera Rubin NVL8 with one Vera CPU each; liquid-cooled
Compute racks2Two nodes per rack, with supply and return manifolds
GPUs32B300 with 288 GB HBM3e each, or Rubin with HBM4
Fabric rackF11Scale-out InfiniBand or Spectrum-X, management and out-of-band
Storage rackS110.25 to 1.0 PB usable NVMe
Power cabinetU11Distribution, metering and lithium UPS ride-through
Coolant distribution unit1Liquid-to-liquid exchanger, duty and standby pumps, filtration
Controls panel1Building management, access control and monitoring
Clean-agent fire suppression2Cylinders with detection, beside the entrance
Heat-rejection packageT11Dry cooler with open top, outside the module

Options.

Chosen at quotation and fixed in the bill of materials for the edition ordered. Prices are given by quotation.

Compute

  • B300 edition: four HGX B300 nodes
  • Vera Rubin edition: four HGX Vera Rubin NVL8 nodes
  • Two-node start, plant sized for four

Storage

  • 0.25, 0.5 or 1.0 PB usable NVMe
  • Immutable backup tier

Network

  • InfiniBand or Spectrum-X Ethernet scale-out
  • Dual-homed uplinks to the customer network
  • Out-of-band management network

Power and autonomy

  • Lithium UPS ride-through
  • Single or dual utility feed
  • Generator interface in the graphite end

Cooling and climate

  • Dry cooler sized for the site ambient
  • Adiabatic assist only where water is available
  • Package in louvre fascia or sleeved finish

Security

  • Access control and CCTV
  • Customer-held encryption keys
  • Telemetry under customer control

Site and services

  • Factory, site and workload acceptance
  • Remote monitoring
  • In-country support and spares

Finish

  • Pearl white sleeve, the reference finish
  • Graphite sleeve with a white logo
  • Titanium-grey sleeve for coastal and industrial sites

Best fit.

Where Core Cell C32 earns its place.

  1. Enterprise private AI cloud
  2. Banking, insurance, healthcare and industrial analytics
  3. Production inference, fine-tuning and batch training
  4. In-country data residency and operational control

Price Core Cell C32 for your site.

Tell us the quantity, the delivery country and the timing. We reply with a written quotation: price, lead time and delivery terms.

Developed by Quantum AI Connect, owner of HYPERCUBE and HYPERQ.
Core Cell C3232 GPUs · Four-node private AI cloud
Request a price