Compute Comparison
vs
All comparisons →

Blackwell Ultra

NVIDIA GB300 SXM6 288GB

wins

4

The GB300 SXM6 (Grace Blackwell Ultra) is a CPU+GPU superchip with 288 GB HBM3e and 8 TB/s memory bandwidth.

Frontier AI trainingInference at scaleMemory-bound LLM serving

Volta

NVIDIA V100 16GB

wins

1

The V100 PCIe 16GB was NVIDIA's flagship AI GPU before the A100. It delivers 112 TFLOPS FP16 (dense) and 900 GB/s HBM2 bandwidth.

Legacy ML trainingBudget HPCMature framework support

Performance comparison

Visual bars — winner highlighted

NVIDIA GB300 SXM6 288GB

Metric

NVIDIA V100 16GB

FP16 Tensor TFLOPS (dense)

2.8k T
112 T

Bandwidth

8.0k GB/s
900 GB/s

VRAM

288 GB
16 GB

TDP (lower=better)

1.2k W
250 W

TFLOPS/watt

2.33
0.45

GB/s per watt

6.67
3.60

Full specification table

SpecNVIDIA GB300 SXM6 288GBNVIDIA V100 16GB
ArchitectureBlackwell UltraVolta
VRAM288 GB HBM3e16 GB HBM2
VRAM typeHBM3eHBM2
Memory bandwidth8,000 GB/s900 GB/s
FP16 Tensor TFLOPS (dense)2,800 TFLOPS112 TFLOPS
TDP1,200 W250 W
TFLOPS/watt2.33 T/W0.45 T/W
GB/s per watt6.67 GB/s/W3.60 GB/s/W
NVLinkYesYes
Release year20252018

Live cloud pricing

On-demand pricing across providers — updated in real time

Per hourPer GB VRAM/hr
Loading prices…

Power efficiency analysis

TFLOPS/watt and GB/s/watt — critical for data center TCO

NVIDIA GB300 SXM6 288GB

FP16 TFLOPS2,800 TFLOPS
TDP1,200 W
TFLOPS/watt2.333 T/W
GB/s per watt6.67 GB/s/W

NVIDIA V100 16GB

FP16 TFLOPS112 TFLOPS
TDP250 W
TFLOPS/watt0.448 T/W
GB/s per watt3.60 GB/s/W

TFLOPS/watt measures compute efficiency — how much AI throughput you get per watt of power consumed. For data centers with PUE of 1.2–1.5, a 10% improvement in TFLOPS/watt translates directly to lower electricity costs and cooling requirements. GB/s/watt measures memory bandwidth efficiency, which is the binding constraint for memory-bound LLM inference workloads.

When to choose each GPU

Choose NVIDIA GB300 SXM6 288GB for:

  • Frontier AI training
  • Inference at scale
  • Memory-bound LLM serving

Weaknesses / Limitations

  • Very limited cloud availability — still ramping supply
  • Extremely high cost — enterprise/hyperscaler pricing only
  • Requires specialized NVLink-C2C infrastructure

Choose NVIDIA V100 16GB for:

  • Legacy ML training
  • Budget HPC
  • Mature framework support

Weaknesses / Limitations

  • Volta architecture — significantly lower TFLOPS than A100 or H100
  • 16 GB VRAM ceiling limits modern LLM workloads
  • No BF16 support — less efficient for modern transformer training