

NVIDIA · GB300 NVL72 (72 Blackwell Ultra GPUs, 36 Grace CPUs)
NVIDIA GB300 NVL72 Rack-Scale AI Factory
Liquid-cooled rack of 72 Blackwell Ultra GPUs and 36 Grace CPUs in one NVLink domain: 130 TB/s NVLink, 37 TB fast memory, 1,440 PFLOPS FP4 (sparse).
₹71,86,20,000
Inclusive of 18% GST. Invoice details confirmed with your order.
- Availability
- Built to order
- Delivery
- Built to order
- SKU
- NV-GB300-NVL72
- Warranty
- Per manufacturer
Key specifications
- CPU cores (total)
- 2,592
- CPU sockets populated
- 36
- CPU model
- NVIDIA Grace (Arm Neoverse V2)
- GPU count
- 72
- GPU model
- 72 NVIDIA Blackwell Ultra GPUs, 36 NVIDIA Grace CPUs
- Cooler type
- Fully liquid-cooled, rack-scale
- FP32 peak (vendor)
- 6,000 TFLOPS
- Interconnect
- 130 TB/s
GB300 NVL72 is NVIDIA's fully liquid-cooled, rack-scale platform for AI reasoning: 72 Blackwell Ultra GPUs and 36 Arm-based Grace CPUs in a single NVLink domain with 130 TB/s of NVLink bandwidth. The rack provides 37 TB of fast memory — 20 TB of GPU memory at up to 576 TB/s and 17 TB of LPDDR5X CPU memory at 14 TB/s — and 2,592 Arm Neoverse V2 CPU cores.
A GB300 NVL72 deployment is a facility project: power delivery, liquid cooling, fabric and software are specified together. Price is based on the published listing of an Indian NVIDIA partner (RDP GPU Mart, ₹42 crore plus GST) for the rack; the facility work is scoped separately. Specifications are NVIDIA's published figures (Tensor Core figures with sparsity unless noted).
Full specifications
Accelerator
| Configuration | 72 NVIDIA Blackwell Ultra GPUs, 36 NVIDIA Grace CPUs | Vendor-reported |
|---|---|---|
| GPU count | 72 | Vendor-reported |
Compute
| Grace CPUs | 36 | Vendor-reported |
|---|---|---|
| CPU cores | 2592 | Vendor-reported |
| CPU | NVIDIA Grace (Arm Neoverse V2) | Vendor-reported |
Networking
| NVLink bandwidth | 130 TB/s | Vendor-reported |
|---|
Memory
| Fast memory | 37 TB | Vendor-reported |
|---|---|---|
| GPU memory | 20000 GB | Vendor-reported |
| GPU memory bandwidth | Up to 576 TB/s | Vendor-reported |
| CPU memory | 17 TB LPDDR5X, 14 TB/s | Vendor-reported |
Performance
| FP4 Tensor Core (sparse | dense) | 1440 | 1080 PFLOPS | Vendor-reported |
|---|---|---|
| FP8 / FP6 Tensor Core (sparse) | 720 PFLOPS | Vendor-reported |
| INT8 Tensor Core (sparse) | 24 POPS | Vendor-reported |
| FP16 / BF16 Tensor Core (sparse) | 360 PFLOPS | Vendor-reported |
| TF32 Tensor Core (sparse) | 180 PFLOPS | Vendor-reported |
| FP32 | 6000 TFLOPS | Vendor-reported |
| FP64 / FP64 Tensor Core | 100 TFLOPS | Vendor-reported |
Thermal
| Cooling | Fully liquid-cooled, rack-scale | Vendor-reported |
|---|
Workload suitability
- calculated Parallel CPU work: 2592 cores.
Performance & benchmarks
No measured benchmarks are published for this product yet. GIRIVON does not publish estimated figures as benchmarks; use the workload performance estimator with your own measurements.
Power & thermal
Power data not specified. Use the power calculator with the manufacturer's figures.
Compatibility
The compatibility engine validates this product against other components using its structured attributes (9 recorded).
Warranty & support
Manufacturer warranty applies. For installation, RMA and technical help, contact GIRIVON support.
Documentation
Reviews
No reviews yet. Reviews are moderated before publication.
Write a review
Related products


Supermicro
Supermicro 8U GPU SuperServer (SYS-822GS-NB3RT)
- GPU count
- 8
- GPU model
- 8 Onboard NVIDIA HGX B300 GPUs
- CPU model
- Dual Intel Xeon 6700 Series Processors with P-cores
₹9,03,53,780
Built to order
