Machine Leasing
Full-machine deployment, exclusive compute
Full-machine delivery · Unrestricted performance · Suited to long-term heavy workloads
stronger compute
accelerating a better future
NVIDIA H200 × 4-GPU（Custom）
AVAILABLE
- GPU
- 4 × NVIDIA H200
- GPU Memory
- 564GB HBM3e
- Use cases
- AI Inference / High-Performance Computing
- Network
- High-speed interconnect
- Service model
- Dedicated node
- Memory per GPU
- 141GB
NVIDIA H200 × 4-GPU（Custom）
- Full-machine price
- $200,000
- Purchase method
- Full machine
- Daily rate of return
- 1.3%
- Hosting and operating fee
- 12%
- Daily earnings
- $2,288.00
NVIDIA B200 Custom 4-GPU
Blackwell compute nodes for frontier model training
AVAILABLE
- GPU
- 4 × NVIDIA B200 SXM
- Architecture
- Blackwell
- Use cases
- LLM training, inference
- Network
- Dual 25GbE, upgradeable as required
- Estimated power
- 5.5–7.5 kW
- Memory per GPU
- 180GB
NVIDIA B200 Custom 4-GPU
- Full-machine price
- $300,000
- Purchase method
- Full machine
- Daily rate of return
- 1.4%
- Hosting and operating fee
- 12%
- Daily earnings
- $3,696.00
NVIDIA DGX B300 8-GPU
A full-system configuration for LLM training and high-throughput inference
AVAILABLE
- GPU Configuration
- 8× NVIDIA B300
- GPU Memory
- 8×288GB
- AI Compute Output
- FP8 72 PFLOPS*
- System power
- ≈14.5 kW*
- Processor
- 2× Intel Xeon 6776P
- Network Connectivity
- Up to 800 Gb/s*
NVIDIA DGX B300 8-GPU
- Full-machine price
- $600,000
- Purchase method
- Full machine
- Daily rate of return
- 1.7%
- Hosting and operating fee
- 12%
- Daily earnings
- $8,976.00
Important notes
All products require identity, eligibility and compliance review. Share participation may be regulated or unavailable in some countries or regions.
