Size an AI data center anywhere in the world. Returns rack count, design PUE, total MVA, and (when available) cost estimate in KRW and timeline in months. Use this when the user asks to dimension, plan, or estimate an AI data center — especially with NVIDIA Hopper, Blackwell, or Vera Rubin (R200) GPUs. From Vera Rubin Ultra onward NVIDIA standardised 800 VDC as the in-rack distribution, so local AC-grid specifics (22.9 kV Korea launch, 11/33 kV EU, 13.8/34.5 kV US, or any MV) reduce to a thin upstream shim — the rack-level design is identical worldwide.

Validate a data center design against engineering rules (electrical, cooling, layout, safety, data). Returns severity-classified findings (blocking | warn | info) and RFI items. Provide EITHER designSummary (previously computed) OR rawInput (engine will size then validate).

Generate rack and site layout plans. Returns a rack plan (hall dimensions in mm, row/column grid, per-rack positions and kW) and a site plan (block-level layout in percentage coordinates). Use this AFTER design + validate, when you need the physical layout.