V100 GPU available now. Pay by the hour, use as much as you need.
Launch immediately from CUDA/driver pre-installed templates. CPU/RAM/NVMe can be changed freely at any time
Free 10Gbps private network included. Multi-node inference also supported when applying for multiple instances simultaneously
※All prices shown are exclusive of tax
V100G08 NODE-A
For redundancy, we recommend ordering one server each from NODE-A and NODE-B
Within the same node, we provide units based on available inventory (GPU slots). Multiple server orders are also possible
| instance | vCPU | memory | storage | data transfer | hourly rate | |
|---|---|---|---|---|---|---|
| NODE-A / V100-G08A（Lite） | 8 Cores | 64 GB | 250 GB NVME SSD | 6TB per month | $0.64/hour | Order |
| NODE-A / V100-G08A（Standard） | 12 Cores | 96 GB | 500 GB NVME SSD | 6TB per month | $0.77/hour | Order |
| NODE-A / V100-G08A（Pro） | 16 Cores | 160 GB | 1000 GB NVME SSD | 6TB per month | $0.98/hour | Order |
| NODE-A / V100-G08A（Max） | 16 Cores | 250 GB | 2000 GB NVME SSD | 6TB per month | $1.18/hour | Order |
| NODE-A / V100-G08B（Lite） | 8 Cores | 64 GB | 250 GB NVME SSD | 6TB per month | $0.64/hour | Order |
| NODE-A / V100-G08B（Standard） | 12 Cores | 96 GB | 500 GB NVME SSD | 6TB per month | $0.77/hour | Order |
| NODE-A / V100-G08B（Pro） | 16 Cores | 160 GB | 1000 GB NVME SSD | 6TB per month | $0.98/hour | Order |
| NODE-A / V100-G08B（Max） | 16 Cores | 250 GB | 2000 GB NVME SSD | 6TB per month | $1.18/hour | Order |
V100G08 NODE-B
For redundancy, we recommend ordering one server each from NODE-A and NODE-B
Within the same node, we provide units based on available inventory (GPU slots). Multiple server orders are also possible
| instance | vCPU | memory | storage | data transfer | hourly rate | |
|---|---|---|---|---|---|---|
| NODE-B / V100-G09A（Lite） | 8 Cores | 64 GB | 250 GB NVME SSD | 6TB per month | $0.64/hour | Order |
| NODE-B / V100-G09A（Standard） | 12 Cores | 96 GB | 500 GB NVME SSD | 6TB per month | $0.77/hour | Order |
| NODE-B / V100-G09A（Pro） | 16 Cores | 160 GB | 1000 GB NVME SSD | 6TB per month | $0.98/hour | Order |
| NODE-B / V100-G09A（Max） | 16 Cores | 250 GB | 2000 GB NVME SSD | 6TB per month | $1.18/hour | Order |
| NODE-B / V100-G09B（Lite） | 8 Cores | 64 GB | 250 GB NVME SSD | 6TB per month | $0.64/hour | Order |
| NODE-A / V100-G08B（Standard） | 12 Cores | 96 GB | 500 GB NVME SSD | 6TB per month | $0.77/hour | Order |
| NODE-B / V100-G09B（Pro） | 16 Cores | 160 GB | 1000 GB NVME SSD | 6TB per month | $0.98/hour | Order |
| NODE-A / V100-G08B（Max） | 16 Cores | 250 GB | 2000 GB NVME SSD | 6TB per month | $1.18/hour | Order |
Client Portal Features
Standard Features and Infrastructure Characteristics
Click “Learn More” on each feature description to navigate to the detailed explanation page
10Gbps Private Network (Free)
GPU-VPS Private Connection / Cloud Private Network
Low-Latency Data Integration
Multi-node inference with multiple server orders
When installing and using applications that support multi-node distributed inference such as GPUStack, we can also accommodate scale configurations across nodes (for distributed inference and distributed processing).
pay-as-you-go
You can use a fully pay-as-you-go model by depositing$31.87 (tax excluded) in advance at the time of sign-up.
Network・Bandwidth
・1Gbps shared line (best effort)
・Up to 6TB monthly transfer included
・Automatic email notification when 80% of transfer capacity is reached
・Bandwidth limited to max
2.5MB/s until end of billing period when limit is exceeded
GPU templates for immediate use (CUDA/Driver pre-installed)
No driver installation hassle. Check with nvidia-smi immediately after startup.
24/7/365 support available
All plans include standard support.
You can open a support ticket from the client portal at any time to request support from us.
V100 GPU VPS Service Specifications (Ubuntu 24.04 LTS)
Dedicated 1 GPU per 1 VM with NVIDIA V100 32GB. Launch your GPU environment immediately from an Ubuntu template with CUDA / NVIDIA drivers pre-installed.
- AI inference / LLM experimentation
- Image generation with Stable Diffusion, etc.
- Machine learning & research applications
- GPU-dependent batch processing
Just register your SSH key and log in to start using GPU.
Pricing structure designed to easily “start small and expand when needed.”
Key Features
Connects with non-GPU cloud instances within the same tenant, making configurations such as preprocessing→inference simple.
Multi-node inference / distributed processing configurations are also possible.
Notifications are sent when balance or transfer volume approaches thresholds, supporting easy-to-control operations.
Pricing Overview (Hourly Billing)
Automatically calculated based on uptime and allocated resources during operation (vCPU / RAM),
as well as allocated storage (NVMe).
| Configuration Example | Typical Resources | Estimate (円 / h) | 730h Equivalent (円 / month) |
|---|---|---|---|
| Lite (Light inference / SD trial) | 8 vCPU / 64GB RAM / 250GB NVMe | approx. 100.45 | approx. 73,329 |
| Standard (SD generation / 7B–13B inference) | 12 vCPU / 96GB RAM / 500GB NVMe | approx. 121.30 | approx. 88,549 |
| Pro (LoRA Training / Batch Processing) | 16 vCPU / 160GB RAM / 1000GB NVMe | Approx. 153.00 | Approx. 111,690 |
| Max (High-Capacity Memory / Long-Running Jobs) | 16 vCPU / 250GB RAM / 2000GB NVMe | Approx. 185.00 | Approx. 135,050 |
Network & IP Addresses
Supports configurations for multi-node inference/distributed processing with multiple machines.
- Free Allowance: 6TB (6,000GB) / VM per month
- Overage: Amount exceeding free allowance is $0.01 / GB/month (additional purchase)
- Notification: Automatic email notification when 80% of free allowance is reached
- Global IP: 1 free by default, additional IPs are $1.91 / IP/month (up to 5 IPs)
Backup & Snapshot (Optional)
- Backup Capacity: $0.01 / GB/month (HDD storage)
- Backup Generations: $12.75 / generation/month (billed according to number of retained generations)
- Snapshot: $1.91 / generation/month (charged according to the number of generations)
- 1 generation: $1.91 / month, 3 generations: $5.74 / month
Billing & Payment
How to Get Started
Frequently Asked Questions
Do you have any questions? Feel free to contact us even if your inquiry is not listed here.
Yes. The NVPS series uses NVMe SSD.
The standard is 100GB NVMe, and you can add more as needed.
KVM-based virtualization environment.
Yes. You can flexibly customize resources not only at the time of order but also after ordering.
Yes. You can reinstall the OS from the control panel.
(Reinstalling will initialize all data on the server.)
Ubuntu 24.04 (GPU template) comes with CUDA/drivers pre-installed, so you can start using the GPU immediately after startup.
Yes. A 10Gbps private network is included free of charge. Private connections are possible between GPU instances and with regular cloud instances. When applications such as GPUStack are installed, multi-node inference and distributed processing are supported with simultaneous orders of multiple instances.
