Eleveight AI Neocloud pricing
Transparent, predictable pricing for your AI workloads, on the newest NVIDIA Blackwell B300.
Deploy on demand, or talk to our team to reserve large-scale capacity at lower rates.
NVIDIA GPU instances
| Resource | Configuration | GPU vRAM (GB) | PFLOPS (at FP4) | Bandwidth (TB/s) | NVLink (TB/s) | On-demand (GPU-hour) |
|---|---|---|---|---|---|---|
| NVIDIA B300 | HGX | 288 | 18 | 8 | 1.8 | $8.75 |
CPU instances
| Resource | vCPUs | CPU vRAM (GB) | CPU-hour |
|---|---|---|---|
| Intel Xeon 6 6527 | 2 - 96 | 16 - 512 | $3.2 |
Storage
| Resource | Unit | Price |
|---|---|---|
| Shared filesystem | Standard | GB/month | $0.115 |
| Shared filesystem | High performance (WEKA) | GiB/month | $0.180 |
| Object storage (volume) | GB/month | $0.019 |
| Boot disk | GB/month | $0.135 |
Other services
| Resource | Price |
|---|---|
| Managed service for Kubernetes | Free** |
| Managed Slurm on K8s | Free** |
| Networking · egress / ingress | Free** |
| Networking · public IP | Free** |
**for on-demand instances only
FAQ about payments
The information and prices on this page are for general reference only. They do not constitute an offer, an invitation to make an offer, or a binding commitment to buy, sell, or otherwise use any service, product, or resource referred to on this website. Eleveight AI may change the prices here at any time without notice.
All prices are shown in US dollars and exclude any applicable taxes, including VAT, which are added where required. On-demand rates are listed on this page; reserved and committed capacity rates are agreed on a case-by-case basis and may differ. Prices vary by configuration, commitment term and availability.