GPU BARE-METAL & SPOT COMPUTE (算力现货)
REAL-TIME SPOT $/HRBare-metal NVIDIA H100 SXM5, H200, B200, and A100 80GB spot instances across Tier-3/4 data centers and decentralized compute grids with InfiniBand and NVLink interconnects.
TRACKED NODES 645 GPUs
H100 SXM5 SPOT
$1.50 / hr
Spot & Reserved Rates
Dynamic market pricing updated every 60 seconds with 24h spread variance.
NVLink & InfiniBand Fabric
900 GB/s bidirectional NVLink 4 and 3.2 Tbps Quantum-2 InfiniBand networking.
Global Regional Distribution
Direct low-latency edge clusters located across US-East, US-West, EU, and Asia-Pacific.
Pre-Configured Environments
Instant SSH, JupyterLab, PyTorch 2.5, vLLM, SGLang, and Docker root containers.
⚡ GPU BARE-METAL INSTANCE SCREENER
Compare hourly spot rates, interconnect architecture, VRAM capacity, and provider hosts.
SORT BY:
🔍
CATEGORY:
| GPU ARCHITECTURE & VRAM | HOST PROVIDER | HOURLY SPOT RATE | ON-DEMAND RATE | INTERCONNECT FABRIC | AVAILABLE INSTANCES | LOCATION EDGE | ACTION |
|---|---|---|---|---|---|---|---|
| 01 NVIDIA H100 SXM5 (80GB) 80GB | RunPod | $2.69 / hr | $3.29 / hr | NVLink 900 GB/s | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 02 NVIDIA H200 (141GB HBM3e) 141GB | RunPod | $3.59 / hr | $4.59 / hr | NVLink 900 GB/s | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 03 NVIDIA B200 (180GB HBM3e) 180GB | RunPod | $5.98 / hr | $6.79 / hr | NVLink 900 GB/s | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 04 AMD Instinct MI300X (192GB) 192GB | RunPod | $0.50 / hr | $2.39 / hr | NVLink 900 GB/s | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 05 NVIDIA A100 SXM4 (80GB) 40GB | RunPod | $1.00 / hr | $1.35 / hr | NVLink 900 GB/s | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 06 NVIDIA L40S (48GB) 48GB | RunPod | $0.79 / hr | $0.99 / hr | PCIe Gen5 | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 07 NVIDIA RTX 5090 (32GB) 32GB | RunPod | $0.69 / hr | $0.99 / hr | PCIe Gen5 | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 08 NVIDIA RTX 4090 (24GB) 24GB | RunPod | $0.34 / hr | $0.74 / hr | PCIe Gen5 | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 09 NVIDIA RTX 6000 Ada (48GB) 48GB | RunPod | $0.74 / hr | $0.84 / hr | PCIe Gen5 | 45 Ready | US East / EU Central (RunPod Secure) | CALCULATE → |
| 10 NVIDIA B200 (192GB HBM3e) 192GB HBM3e | Vast.ai | $5.31 / hr | $7.17 / hr | NVLink 900 GB/s (8x Cluster) | 1 Ready | Oregon, US | CALCULATE → |
| 11 NVIDIA H200 (141GB HBM3e) 141GB HBM3e | Vast.ai | $2.97 / hr | $4.50 / hr | NVLink 900 GB/s (8x Cluster) | 2 Ready | Virginia, US | CALCULATE → |
| 12 NVIDIA H100 SXM5 (80GB) 80GB HBM3 | Vast.ai | $1.50 / hr | $2.99 / hr | NVLink 900 GB/s (8x Cluster) | 1 Ready | Global Cloud Cluster | CALCULATE → |
| 13 NVIDIA A100 SXM4 (80GB) 80GB HBM2e | Vast.ai | $0.27 / hr | $1.85 / hr | NVLink 900 GB/s (8x Cluster) | 1 Ready | Florida, US | CALCULATE → |
| 14 NVIDIA RTX 5090 (32GB) 32GB GDDR7 | Vast.ai | $0.32 / hr | $1.20 / hr | PCIe Gen5 64 GB/s | 17 Ready | Quebec, CA | CALCULATE → |
| 15 NVIDIA RTX 4090 (24GB) 24GB GDDR6X | Vast.ai | $0.11 / hr | $0.75 / hr | PCIe Gen5 64 GB/s | 8 Ready | , CN | CALCULATE → |
| 16 NVIDIA RTX 6000 Ada (48GB) 48GB GDDR6 | Vast.ai | $0.74 / hr | $1.60 / hr | PCIe Gen5 64 GB/s | 10 Ready | Utah, US | CALCULATE → |
| 17 NVIDIA H100 SXM5 (80GB) 80GB HBM3 | Lambda Labs | $2.49 / hr | $3.29 / hr | NVLink 4.0 900 GB/s | 32 Ready | US West / US East (Lambda Cloud) | CALCULATE → |
| 18 NVIDIA H200 (141GB HBM3e) 141GB HBM3e | Lambda Labs | $3.49 / hr | $4.50 / hr | NVLink 4.0 900 GB/s | 16 Ready | US West (Oregon) | CALCULATE → |
| 19 NVIDIA H100 HGX (80GB) 80GB HBM3 | CoreWeave | $2.85 / hr | $3.75 / hr | Quantum-2 InfiniBand 3.2 Tbps | 128 Ready | US East (New Jersey Tier-4) | CALCULATE → |
| 20 Cerebras CS-3 (Wafer-Scale Engine) 44GB SRAM On-Chip (900k AI Cores) | Cerebras Cloud | $3.50 / hr | $4.80 / hr | Swarm Fabric 214 PetaBits/s | 24 Ready | US West / US Central (Cerebras Supercomputing Clusters) | CALCULATE → |