Rent cloud GPUs for AI & ML.
Choose from NVIDIA and AMD GPUs. Deploy in minutes with built-in auto-failover, no egress fees, and per-minute billing.
NVIDIA RTX 4090
24GB GDDR6X · Ada Lovelace
$0.35/hr
billed per minute
LLM inference, Stable Diffusion, real-time AI applications
NVIDIA RTX 3090
24GB GDDR6X · Ampere
$0.18/hr
billed per minute
Cost-effective inference, fine-tuning small models, batch processing
NVIDIA RTX Pro 6000
96GB GDDR7 · Blackwell
$1.55/hr
billed per minute
Large model inference and training on 96GB of VRAM
NVIDIA H100 SXM
80GB HBM3 · Hopper
$2.60/hr
billed per minute
Large model training, multi-GPU workloads, enterprise AI
NVIDIA A100 80GB
80GB HBM2e · Ampere
$1.10/hr
billed per minute
Mixed-precision training, 70B inference with 8-bit quantization, large batch inference
NVIDIA RTX 5090
32GB GDDR7 · Blackwell
$0.50/hr
billed per minute
Inference on 32GB of GDDR7, 33% more VRAM than the RTX 4090
NVIDIA B200
180GB HBM3e · Blackwell
$5.50/hr
billed per minute
180GB per GPU and FP4 Tensor Cores, for models that outgrow an H100
NVIDIA B300
288GB HBM3e · Blackwell Ultra
$7.49/hr
billed per minute
288GB per GPU and 13.5 PFLOPS of dense FP4, for the largest models
AMD Instinct MI355X
288GB HBM3E · CDNA 4
$5.50/hr
billed per minute
288GB per GPU on ROCm, for large models, 4-bit inference and FP64 work
Not sure which GPU to pick?
Our team can help you choose the right GPU for your workload.