01 · BASIS
Engineering basis, not aesthetics
Every number on this site is grounded in an engine basis assumption — rack density, MV topology, CDU placement, validation rule family. We surface the basis, not just the headline.
ABOUT
An infrastructure design engine, written by people who size halls and plants.
AIDC-AI.IO exists because Rubin-era density makes the legacy AI data-center workflow — isolated spreadsheets for electrical, separate decks for cooling, a layout drawn after both — break in ways that reviewers can see. The site documents how we close that gap with a single engine and one source of truth for every consuming view.
PRINCIPLES
02 · GENERIC
Generic engine, profile-aware
There is no Phase-1 / Phase-2 split inside the engine. Future-vendor or site-specific behavior is represented as optional capability profiles that toggle on top of the same schema.
03 · EVIDENCE
Evidence over slides
We treat external validators (pandapower, OpenFOAM, EnergyPlus, Ecodial, SIMARIS) as the audience. The engine output is the artifact; the deck is a view of it.
SCOPE
What the engine covers — and what it does not.
We are deliberate about boundaries. Items outside the engine are tracked as evidence workflows or RFIs, not absorbed into vague pseudo-features.
| Inside the engine | External / evidence |
|---|---|
| Site totals & hall panels | Pandapower / pypowsybl power flow |
| Rack & EPOD layout | OpenFOAM CFD validation |
| CDU & secondary loop sizing | EnergyPlus / OpenStudio EnergyModel |
| 22.9 kV / RMU / transformer planning | Ecodial · SIMARIS · DOC · xSpider sizing |
| Validator rule families | Owner / vendor RFI ledger |
CONTACT
Want to compare your hall against the engine?
We work with architects, EPCs, hyperscaler infra teams, and owners on Rubin-era designs. Reach out with a basis brief and we'll come back with a model-driven view.