NVIDIA publishes no per-GPU board power for the NVL72 bin, only a full-rack figure of up to 142 kW. tdp_watts is deliberately NULL rather than carrying a contradicted third-party figure.

NVIDIA GB300 NVL72 GPU 288GB

Blackwell Ultra SXM 2025 4nm Spec confidence: Official
FP8 (dense)
5,000
TFLOPS
FP32
83
TFLOPS
VRAM
288 GB
HBM3e
Bandwidth
8.0 TB/s
memory
Download datasheet

Overview

This is a single Blackwell Ultra GPU as deployed inside a GB300 NVL72 rack, which is the unit cloud providers rent and quote. It carries 288 GB of HBM3e at 8 TB/s and 15 PFLOPS of dense NVFP4, about 11 percent above the air-cooled HGX B300 bin on every tensor row. Blackwell Ultra is a memory and inference part rather than a general uplift: dense NVFP4 rises by half again over Blackwell while sparse NVFP4 stays put, and INT8 and FP64 are both deliberately de-emphasised, so this is a poor choice for classical HPC in a way a GB200 NVL72 GPU is not. It is compute-identical to the DGX Station desktop superchip and differs only in memory capacity, bandwidth and cooling.

Performance

Peak theoretical throughput by precision type

Precision
Dense
2:4 Sparse
FP64
64-bit floating point
1.4TFLOPS
Structured sparsity is a tensor-core feature; this vector precision has no sparse form
FP32
32-bit floating point
83TFLOPS
Structured sparsity is a tensor-core feature; this vector precision has no sparse form
TF32
TensorFloat-32
1,250TFLOPS
2,500TFLOPS
BF16
Brain Float 16
2,500TFLOPS
5,000TFLOPS
FP16
16-bit floating point
2,500TFLOPS
5,000TFLOPS
FP8
8-bit floating point
5,000TFLOPS
10,000TFLOPS
FP6
5,000TFLOPS
10,000TFLOPS
FP4
4-bit floating point
15,000TFLOPS
20,000TFLOPS
INT8
8-bit integer
165TOPS
330TOPS

Specifications

Architecture

Blackwell Ultra

Form Factor

SXM

Launch Year

2025

Process Node

4nm

Memory

288 GB HBM3e

Bandwidth

8,000 GB/s

TDP

No verified data available

Max Power

No verified data available

Transistors

208.0 billion

Interconnect

1.8 TB/s NVLink

Spec Confidence

Official

Full Specifications

Chip Design
Transistors 208.0 billion
Process Node 4nm
Chiplets 2 (Two reticle-limited dies joined by a 10 TB/s die-to-die link, presented to software as one GPU)
Memory
VRAM 288 GB
Memory Type HBM3e
Bandwidth 8.0 TB/s
Interface Width No verified data available
Interconnect & I/O
GPU-to-GPU NVLink
Interconnect Bandwidth 1.8 TB/s
Power & Thermal
TDP No verified data available
Max Board Power The vendor publishes no maximum board power; the Max Power tile above estimates it from TDP
Cooling Liquid-cooled
Enterprise Features
ECC Memory Yes
Sparsity Yes
General
Form Factor SXM
Architecture Blackwell Ultra
Process Node 4nm
Launch Year 2025

Datasheet & Resources

Data Provenance

Every figure traced to a source

Primary Source

Publisher
NVIDIA
Published
2025-01-01

Data Quality

Spec confidence
Official
Clock basis
Boost
Core precisions with figures
9 of 9
Normalization
All values in TFLOPS

Compare cloud providers offering on-demand GPU instances for AI training, inference, and HPC workloads.

Browse GPU Cloud Providers

Similar GPUs

NVIDIA B300 SXM 288GB

SXM · 2025
FP32: 75 TFLOPS
Compare vs B300

NVIDIA B30A

· 2025
FP32: No verified data available
Compare vs B30A

Desktops Built on This Chip

AI desktops and workstations powered by the NVIDIA GB300

DesktopMemoryAI Performance
ASUS ExpertCenter Pro ET900N G3
ASUS · 2025
748 GB 20,000 TOPS View Desktop
MSI XpertStation WS300
MSI · 2025
748 GB 20,000 TOPS View Desktop
NVIDIA DGX Station GB300
NVIDIA · 2025
748 GB 20,000 TOPS View Desktop
Supermicro Super AI Station ARS-511GD
Supermicro · 2025
748 GB 20,000 TOPS View Desktop
Dell Pro Max with GB300
Dell · 2025
784 GB 20,000 TOPS View Desktop

Frequently Asked Questions

How many TFLOPS does the NVIDIA GB300 have?

The NVIDIA GB300 delivers 83 TFLOPS FP32, 2,500 TFLOPS FP16 and 5,000 TFLOPS FP8 at peak.

What is the power consumption of the NVIDIA GB300?

Flopper does not currently have a verified TDP (Thermal Design Power) figure for the NVIDIA GB300.

How much memory does the NVIDIA GB300 have?

The NVIDIA GB300 is equipped with 288 GB of VRAM with 8,000 GB/s of memory bandwidth.

What architecture is the NVIDIA GB300 based on?

The NVIDIA GB300 is based on the Blackwell Ultra architecture, launched in 2025.

Stay Updated on GPU Releases

Get notified when new GPUs are added or specifications are updated.

Loading verification...

No spam, unsubscribe anytime.

Back to GPUs