An independent read on whether the infrastructure is actually there

Why the initiative exists

AI-scale compute is being discussed across the Midwest faster than the underlying infrastructure questions are being answered. Announcements routinely precede any public accounting of firm generation, transmission deliverability, water availability, or the multi-year equipment lead times that pace every build.

That gap creates real risk on both sides. Developers commit capital against assumptions that may not survive an interconnection study. Communities form expectations — or objections — before anyone has laid out what a given site would physically require. And utilities and planners are left absorbing timelines that were set without them.

WAII exists to put a neutral, technical baseline into that conversation early, while it is still cheap to change course.

How we assess readiness

Our analysis is constraint-based. Rather than asking whether a region wantsAI-scale compute, we ask what would physically have to be true for it to work, then test each condition against the public record. Eight constraint domains carry most of the weight:

• Power capacity — firm generation and resource-adequacy margin against loads that could materially shift regional demand.
• Grid deliverability — transmission, substation, and interconnection capacity. Frequently the binding constraint, and rarely the one discussed first.
• Cooling — thermal load at AI rack density and what each site realistically needs: evaporative, closed-loop, or liquid.
• Water — sustainable supply, consumptive-use limits, and Great Lakes basin permitting and oversight.
• Land, zoning, and community — parcel and zoning fit, plus the social license that increasingly determines whether a project proceeds.
• Workforce and schedule — skilled-trade availability and equipment lead times, with large transformers running on multi-year horizons.
• Regulatory and planning — MISO resource adequacy, Public Service Commission tariff and cost-allocation rules, and who ultimately pays.
• Execution risk — documented failure modes, including cost misallocation, demand overestimation, permitting delay, and community backlash.

We work from public and verifiable sources: utility filings and integrated resource plans, MISO queue and planning data, Public Service Commission dockets, DNR and permitting records, and published equipment and construction lead times. Where the record is thin, we say so rather than filling the gap with an estimate that reads as fact.

What independence means here

We are not selling a site, a ranking, or a narrative. WAII does not represent developers, utilities, landowners, or economic development organizations, and our conclusions are not reviewed by any party with a position in a project. When our analysis indicates a constraint is binding, we publish that — including when it is inconvenient for a proposal that is otherwise well received.

This also means we take no position on whether Wisconsin should pursue AI-scale compute. That is a public policy question. Ours is the prior question: what the infrastructure can currently support, and what it would take to change that.