Hosting Opportunity
Trusted by 10,000+
Researchers, AI Startup Teams, Engineers
and AI Endeavors across the world.
HPC-AI's Core infrastructure Colossal-AI already powers global leading enterprises.
Why Host Your GPU Clusters with Us?
Monetize Idle Hardware
Turn underutilized servers into a predictable revenue stream.
Access a Global Market
Access to international customers across various industries and use cases.
Seamless Integration
Our expert team handles the technical integration, billing, and customer support for you.
A Clear Path to Partnership
Review & Qualify
Check our hosting requirements to ensure your resources fit all requirements.
Formal Agreement
Eastablish an agreement that covers technical specifications, SLAs, and profit-sharing model.
Go Live & Earn
Our expert team manages the technical setup, security, and maintenance, and provide you with dashboard for performance tracking.
Hosting Requirements
Hardware and software requirements to join our platform.
See Details >| Component | Minimal Requirements |
|---|---|
| GPU | Nvidia GPUs such as B200, H200, H100, A100, RTX-4090, etc. Minimum 1 server with 8 GPU cards and ≥ 4TB local storage. |
| CPU | 3 physical nodes, each with 16 cores, 32GB RAM, 300GB system disk, and 1TB data disk |
| Network & Storage | Network: Public IP for CPU nodes, ≥200 Mbps shared bandwidth, and ≥10 Gbps internal speed Local Storage: Minimum 4TB of performant local storage (NVMe recommended) per GPU. (Required) Shared Storage: High-performance distributed file system (e.g., WEKA, DDN, Lustre) is a requirement for multi-node clusters and must include redundancy to ensure data security. (Preferred) |
| Software | OS: Ubuntu 22.04 (preferred) or 20.04 | GPU Driver: 550.124.06 | CUDA: 12.8 All GPU nodes must pass standard nccl-test benchmarks to verify interconnect performance. |
Note: Consumer-grade setups and homelab GPUs are not accepted as they do not meet enterprise reliability and security standards.
Partnership Terms
Minimum Commitment
3 months with flexible renewal options
Testing Period
2 business days full stress test for full hardware validation
Deployment
Production environment testing for immediate launch
Frequently Asked Questions
We prioritize high-performance, enterprise-grade GPUs such as the NVIDIA H200, H100 and A100, along with large-scale clusters designed for demanding AI workloads. Specific requirements will be confirmed during the technical consultation.
Join Our Partner Network Today.
Join leading hardware providers who are already monetizing their resources on the HPC-AI platform.