NVIDIAIn stock
B200 SXM
frontier training · large-scale inference
- VRAM
- 180 GB
- vCPU
- 30
- RAM
- 240 GB
- NVMe
- 3.84 TB
AnchorGPU · Dedicated GPU rental
Rent GPUs with crypto.
By the week or month.
Compare NVIDIA and AMD GPUs for training, inference, and rendering. Configure your hardware and see the full price for 7 or 30 days.
Prices in USD No identity fields in signup 1 / 2 / 4 / 8 GPUs Weekly or monthly
15GPU models
24–192 GBVRAM per GPU
1 / 2 / 4 / 8GPUs per node
7 / 30fixed days per term
Catalog inventory
Choose the right GPU, not a bidding market.
NVIDIAIn stock
H200 SXM
LLM inference · fine-tuning
- VRAM
- 141 GB
- vCPU
- 24
- RAM
- 192 GB
- NVMe
- 3.84 TB
NVIDIAIn stock
H100 SXM
training · high-throughput inference
- VRAM
- 80 GB
- vCPU
- 24
- RAM
- 192 GB
- NVMe
- 3.84 TB
NVIDIAIn stock
H100 PCIe
fine-tuning · inference
- VRAM
- 80 GB
- vCPU
- 20
- RAM
- 160 GB
- NVMe
- 2 TB
AMDIn stock
MI300X
large models · ROCm workloads
- VRAM
- 192 GB
- vCPU
- 28
- RAM
- 256 GB
- NVMe
- 3.84 TB
NVIDIAIn stock
A100 SXM
training · batch inference
- VRAM
- 80 GB
- vCPU
- 20
- RAM
- 160 GB
- NVMe
- 2 TB
One fixed periodNo hourly metering
Private by defaultNo identity documents
Multi-GPU configurationsCompare 1 to 8 GPUs per node
Choose your environmentImages for AI and development