NVIDIA H100 vs NVIDIA V100: which is better for AI?

Choose NVIDIA H100 if you need 80GB+ on a single GPU (48GB more than NVIDIA V100). Choose NVIDIA H100 if you need FP16 tensor throughput (691% faster than NVIDIA V100).

What's the price difference between NVIDIA H100 and NVIDIA V100?

On Runcrate today, NVIDIA H100 starts at $4.29/hr and NVIDIA V100 starts at $3.04/hr; NVIDIA V100 is cheaper by 29%.

Can I deploy NVIDIA H100 or NVIDIA V100 on Runcrate today?

NVIDIA H100 is on demand in 4 regions from $4.29/hr, billed by the second on most GPUs. NVIDIA V100 is on demand in 1 region from $3.04/hr, billed by the second on most GPUs.

How much VRAM does each GPU have?

NVIDIA H100 has 80GB HBM3. NVIDIA V100 has 32GB HBM2.

GPU comparison

NVIDIA H100 vs NVIDIA V100.

Specs, performance and live cloud availability, side by side. The NVIDIA V100 is cheaper today, from $3.04/hr.

Launch V100 Get a reserved quote

The short answer

Choose NVIDIA H100 if you need 80GB+ on a single GPU (48GB more than NVIDIA V100).
Choose NVIDIA H100 if you need FP16 tensor throughput (691% faster than NVIDIA V100).

Compare another pair

Hopper

NVIDIA H100

from$4.29/hr

Available now · 4 regions

Memory: 80 GB HBM3
FP16 tensor: up to 989 TFLOPS
Bandwidth: up to 3.4 TB/s

Launch H100 Get a reserved quote

Full pricing details

Volta

NVIDIA V100

from$3.04/hrCheaper

Available now · 1 region

Memory: 32 GB HBM2
FP16 tensor: 125 TFLOPS
Bandwidth: 900 GB/s

Launch V100 Get a reserved quote

Full pricing details

At a glance

Spec by spec. The longer bar is the better number.

Memory

NVIDIA H10080 GB

NVIDIA V10032 GB

Memory bandwidth

NVIDIA H100up to 3.4 TB/s

NVIDIA V100900 GB/s

FP16 tensor

NVIDIA H100up to 989 TFLOPS

NVIDIA V100125 TFLOPS

Hourly price, from (live)

NVIDIA H100$4.29/hr

NVIDIA V100$3.04/hr

Full specs

The complete spec sheet.

Spec	NVIDIA H100	NVIDIA V100
Architecture	Hopper	Volta
Memory	80 GB	32 GB
Memory type	HBM3	HBM2
Memory bandwidth	up to 3.4 TB/s	900 GB/s
FP32	up to 67 TFLOPS	15.7 TFLOPS
FP16 tensor	up to 989 TFLOPS	125 TFLOPS
INT8	up to 2400 TOPS	125 TOPS
TDP	up to 700 W	300 W
GPUs per node	8	8
Hourly price (live)	from $4.29/hr	from $3.04/hr

Questions

H100 or V100. What buyers ask first.