CPU instance types

vCPUs
Memory
Disk (SSD)
On-demand price (per hour)
AMD EPYC 7552
2 vCPU
4 GiB
80 GiB
$0.10
AMD EPYC 7552
4 vCPU
8 GiB
80 GiB
$0.20
AMD EPYC 7552
8 vCPU
16 GiB
80 GiB
$0.40
AMD EPYC 7552
12 vCPU
24 GiB
80 GiB
$0.60
AMD EPYC 7552
16 vCPU
32 GiB
80 GiB
$0.80
AMD EPYC 7552
20 vCPU
40 GiB
80 GiB
$1.00
AMD EPYC 7552
24 vCPU
48 GiB
80 GiB
$1.20

Storage types for most demanding needs

Block Storage
$
0.04
/ GB / month
High throughput and low latency
Persistent and reliable
SSD or HDD (prices differ)
Ideal for data reuse and long-term archive
High-Speed File Storage
$
0.10
/ GB / month
Highest performance
Flexible and infinitely scalable
Data platform for AI with VAST Data
Ideal for multi-node trainings on large datasets
Object Storage
$
0.03
/ GB / month
S3-compatible
Elastic and highly scalable
Ideal for ML datasets and backups
Local Storage
Pricing depends on the configuration
High performance and IOPS
Low latency and highly available
Ideal for applications requiring fast and reliable data access
Disk Snapshots
$
0.02
/ GB / month
Quick and efficient
Safe point-in-time copies
Ideal for quick backups, versioning, and recovery

Premium Network for speed and scale

Public Internet Connectivity
Multi Gbps and redundant
IPv4 (paid) + IPv6 (free)
Configurable stateful firewall
No NAT
No traffic charges
RDMA
Networking
400 Gbps per GPU
Non-blocking
Low-latency
No traffic charges
Private
Networking
200 Gbps
Non-blocking
Any L3 protocol supported
No traffic charges
Firewall Security
Inbound and outbound groups
Avoid unauthorized access
Included with our service
Secure SSH Access
Fully encrypted SSH access
Included with our service
Serial Console
Monitor system operations
Included with our service

Additional features

Icon
Instance snapshots
Icon
Security groups
Icon
High internet bandwidth
Icon
Preconfigured images
Icon
Public APIs, Terraform and Pulumi
Talk to our solution experts
Comprehensive full-stack expert support. Contact us for a free consultation
Arrow

*Prices are indicated per GPU component per hour.
The minimum rental for NVIDIA HGX H100 and NVIDIA HGX H200 is one node with 8x GPUs.
All prices are shown without any applicable taxes, including VAT.