Transparent pricing

Private GPUs in Secure Cloud

Rent dedicated H100, A100 and L40S GPUs by the hour, billed per minute. For reserved capacity or private clusters, talk to us.

13 GPUs · $0.20–$7.50/hr

Sort

RTX A4000 ↗

1–8×

Secure Cloud

vRAM: 16 GB
Per GB vRAM: $0.013/hr

$0.20per GPU-hr

RTX 4090 ↗

1–8×

Secure Cloud

vRAM: 24 GB
Per GB vRAM: $0.019/hr

$0.45per GPU-hr

RTX A6000 ↗

1–8×

Secure Cloud

vRAM: 48 GB
Per GB vRAM: $0.011/hr

$0.55per GPU-hr

RTX 5090 ↗

1–8×

Secure Cloud

vRAM: 32 GB
Per GB vRAM: $0.017/hr

$0.55per GPU-hr

RTX 6000 Ada ↗

1–8×

Secure Cloud

vRAM: 48 GB
Per GB vRAM: $0.018/hr

$0.88per GPU-hr

NVIDIA L40S ↗

1–8×

Secure Cloud

vRAM: 48 GB
Per GB vRAM: $0.021/hr

$1.00per GPU-hr

NVIDIA A100 80GB PCIe ↗

1–8×

Secure Cloud

vRAM: 80 GB
Per GB vRAM: $0.019/hr

$1.49per GPU-hr

NVIDIA A100 80GB SXM ↗

1–8×

Secure Cloud

vRAM: 80 GB
Per GB vRAM: $0.019/hr

$1.49per GPU-hr

RTX PRO 6000 ↗

1–8×

Secure Cloud

vRAM: 96 GB
Per GB vRAM: $0.021/hr

$1.99per GPU-hr

NVIDIA H100 80GB PCIe ↗

1–8×

Secure Cloud

vRAM: 80 GB
Per GB vRAM: $0.034/hr

$2.75per GPU-hr

NVIDIA H100 80GB SXM ↗

1–8×

Secure Cloud

vRAM: 80 GB
Per GB vRAM: $0.037/hr

$2.99per GPU-hr

NVIDIA H200 ↗

1–8×

Secure Cloud

vRAM: 141 GB
Per GB vRAM: $0.030/hr

$4.20per GPU-hr

NVIDIA B200 ↗

1–8×

Secure Cloud

vRAM: 180 GB
Per GB vRAM: $0.042/hr

$7.50per GPU-hr

Need something else?

Reserved capacity, private clusters and multi-node interconnect are quoted directly.

Talk to us

All rates are per GPU hour, billed per minute. Every supported model runs at the same rate.

How our pricing works

One rate per GPU hour, whatever you run on it. No instance-type maths, no egress surprises.

Transparent, real-time pricing

No hidden fees. Rates update live as you use it, so the number on the card is the number you pay.

Built for discovery

Sort by price, memory or cost per gigabyte and find the instance that fits the job in seconds.

Flexible, fluid billing

Pay per minute on demand, or move to reserved and enterprise capacity once the workload settles.

Instant deployments

Launch fast, scale on demand and automate the whole thing from the CLI or API.

Give your workloads exactly what they need, without overbuying, guesswork or long-term lock-in.

Deploy a GPU Contact sales