GPU Cloud Server
On-demand cloud GPU computing, supports CUDA, cuDNN, suitable for AI training, inference, scientific computing, graphics and image processing, video encoding/decoding and more.
Full GPU Driver Support
Supports CUDA, cuDNN, perfectly adapted for AI training, inference, and scientific computing needs.
Flexible Configuration
From single GPU to multi-GPU, dynamically adjust GPU resources based on workload.
Rapid Deployment
Cloud GPU servers can be started within minutes, start computing immediately.
Pay-as-you-go
Usage-based billing, no long-term contracts, flexibly support project needs.
Single Instance
Up to 5 PFLOPS
Ultra-powerful GPU computing cards, combined with high-performance CPU platform, mixed-precision computing performance up to 5 PFLOPS, suitable for various high-performance computing scenarios.
- AI Artificial Intelligence Training
- Machine Learning & Deep Learning
- Scientific Computing Simulation
- Graphics & Image Processing
- Video Encoding/Decoding
Available from a single instance
GPU Cloud Server Plans
Flexible cloud computing, meeting diverse AI workload needs
RTX-4090 24GB x1 Standard
Use Cases
Specifications
- 16 Core CPU
- 64GB RAM
- 100GB HD
RTX-4090 48GB x1 PRO
Use Cases
Specifications
- 14 Core CPU
- 120GB RAM
- 100GB HD
A100 80GB x1 M-Type
Use Cases
Specifications
- 14 Core CPU
- 120GB RAM
- 100GB HD
H100 80GB x1 SXM
Use Cases
Specifications
- 128 Core CPU
- 512GB RAM
- 500GB HD
H200 141GB x1 SXM
Use Cases
Specifications
- 128 Core CPU
- 1TB RAM
- 500GB HD
GB300 NVL72 Super Flagship
Use Cases
Specifications
- 72x Blackwell GPU
- 14.4TB RAM
- 800Gb InfiniBand
RTX-4000 20GB Ada x1
Use Cases
Specifications
- 4 Core CPU
- 16GB RAM
- 512GB HD
Use Cases
AI Training
Large-scale model training, quickly build training environment
AI Inference
Real-time inference service, low latency and high throughput
Visual Processing
Image recognition, video analysis, computer vision applications