Infrastructure Explorer

Free-form sizing on a single screen. Edit any value and watch the numbers recompute. For multi-use-case sizing, use the wizard.

Hardware target

Model & precision

Workload pattern

Concurrency & tokens

50
4096
1024

Data architecture

Operations

VRAM
199.3 GB
GPUs
10
System RAM
448.0 GB
Storage
320.3 GB
Power
9.1 kW
3-yr TCO
$943.4K

VRAM breakdown

Customer Support RAG: KV cache exceeds model weights
KV cache is 104.9 GB vs 84.0 GB of weights. This is the #1 surprise at scale — consider PagedAttention (vLLM), reducing the agent multiplier, or capping concurrent users.
Multi-node deployment
Total 10 GPUs spans 2 server nodes. You'll need InfiniBand (or equivalent low-latency fabric) between them for tensor parallelism.
Only one use case configured
If you're sizing for the whole org, add the other workloads now. Aggregate sizing finds shared infrastructure savings (e.g., shared embedding GPU, deduped model files).

Hardware BOM

ItemQuantitySpecCost
NVIDIA H100 SXM1080 GB · 3350 GB/s · 700 W$300K
Server nodes (8-GPU HGX)2Chassis, CPU, motherboard, PSU$40K
System RAM (DDR5 ECC)448.0 GB≥2× total VRAM, HNSW headroom$2.2K
Enterprise NVMe storage320.3 GBModels, vector index, logs$48
Networking fabric1InfiniBand HDR (multi-node)$50K
Power distribution + UPS2PDU + battery backup per node$10K
Total CAPEX$402.3K
Annual OPEX (power + cooling + licenses + ops)$180.4K
3-Year TCO$943.4K

CAPEX breakdown