What are you building?
Select the AI workloads you need to run. Choose all that apply.
H100 80GB × 4-8
Fine-tuning
Adapt foundation models to your domain with custom datasets
Training hoursDataset sizeEpochs
L40S or A10G
RAG / Retrieval
Embedding generation, vector storage, and retrieval-augmented generation
Queries/secLatency p99Index size
H100 or L4
Inference
Production model serving with low latency requirements
Tokens/secTTFTConcurrency
DGX SuperPOD
Pre-training
Train foundation models from scratch (rare, high compute)
Petaflop-daysCheckpoint frequency
Mixed (inference-heavy)
Agents
Multi-step AI orchestration with tool use and reasoning
Tasks/minTool callsSuccess rate
Industry
Select your industry to apply relevant compliance and workload presets