Hardware Selection & Payment
Before onboarding your team to our platform, we want to ensure we have the proper hardware to support your AI workloads. Today, we offer the following hardware
| Chip | VRAM | RAM | RAM | Storage |
|---|---|---|---|---|
A100 (SXM) | 40GB | 15.5 | 200 GiB | 512 GiB SSD |
A100 (SXM) | 80GB | 30 | 225 GiB | 2.5 TiB SSD |
H100 (SXM)* | 80GB | 26 | 225 GiB | 3.25 TiB SSD |
We offer these GPUs at extremely competitive prices. See pricing below or learn more here.
| Chip Type | Predictable Spot Pricing | Guaranteed Uptime Pricing |
|---|---|---|
A100 40GB (SXM) | $1.05/gpu/hr | $1.42/gpu/hr |
A100 80GB (SXM) | $1.45/gpu/hr | $1.97/gpu/hr |
H100 80GB (SXM)* | $2.79/gpu/hr (*$22.32/hr for all 8 GPUs) | $3.85/gpu/hr ($30.80/hr for all 8 GPUs) |
Notes on chip types:
No payment required for data movement (ingress/egress)
Prices listed in the chart above are “Plus applicable sales tax”
GPUs come pre-installed with popular ML frameworks: Ubuntu, TensorFlow, PyTorch, NVIDIA CUDA, and NVIDIA cuDNN
*H100 GPUs must be reserved / used in quantities of 8 (entire server)
After selecting a chip type, you'll need to select a number of GPUs and length of time in weeks (ex. 4 A100 40GB GPUs for 3 weeks). From there, we'll be able to calculate your total all in price.
With your all in price set, we'll send you a payment link to prepay for your GPUs. After receiving your payment we'll begin onboarding you to the platform. Your reserve will period begin on the start date agreed upon with our team.
Last updated