GPU group

Multi-GPU training.

Datacenter cards designed to scale across NVLink — training and large-batch inference.

GPU	Tier	Architecture	VRAM	Providers	AI models	From
Nvidia Tesla V100 SXM2 32GB	datacenter	Volta	32GB	1 →	—	$0.023/hr	Compare →
Nvidia T4	datacenter	Turing	16GB	4 →	—	$0.070/hr	Compare →
Nvidia L4	datacenter	Ada Lovelace	24GB	5 →	—	$0.18/hr	Compare →
Nvidia Tesla V100	datacenter	Volta	16GB	9 →	—	$0.21/hr	Compare →
Nvidia A40	datacenter	Ampere	48GB	2 →	—	$0.28/hr	Compare →
Nvidia L40S	datacenter	Ada Lovelace	48GB	10 →	—	$0.28/hr	Compare →
Nvidia A30	datacenter	Ampere	24GB	1 →	—	$0.39/hr	Compare →
Nvidia NVIDIA A40	datacenter	Ampere	48GB	4 →	—	$0.42/hr	Compare →
Nvidia H200	datacenter	Hopper	141GB	10 →	—	$0.46/hr	Compare →
Nvidia A100	datacenter	Ampere	80GB	14 →	—	$0.48/hr	Compare →
Nvidia L40	datacenter	Ada Lovelace	48GB	2 →	—	$0.64/hr	Compare →
Nvidia A10	datacenter	Ampere	24GB	4 →	—	$0.75/hr	Compare →
Nvidia H100	datacenter	Hopper	80GB	20 →	—	$0.80/hr	Compare →
Nvidia Nvidia A100 80GB PCIe	datacenter	Ampere	40GB	11 →	—	$1.29/hr	Compare →
AMD AMD MI300X	datacenter	CDNA 3	192GB	3 →	—	$1.95/hr	Compare →
Nvidia GH200	datacenter	Hopper	96GB	1 →	—	$2.29/hr	Compare →
Nvidia B200	datacenter	Blackwell	192GB	8 →	—	$2.79/hr	Compare →
Nvidia B300	datacenter	Blackwell	288GB	2 →	—	$5.38/hr	Compare →