About AI Data Center Impact
For twenty years, this work meant helping stores, banks, hotel brands and plasma collection networks choose sites. Within twelve months, we could see whether the choice was right. A wrong choice had a clear, costly result. Learning from those results is how a strong site-selection method is built.
AI data center siting doesn't have that loop yet. Sites are selected on power and price, and the verdict arrives two years later in a local hearing room.
AI Data Center Impact applies that method after power and land screening. We return a sourced record for the site's coordinates. It covers nearby sensitive uses, service capacity, local review steps, hearing records and changes that may affect approval or underwriting.
This work adds to—not replaces—energy analysis, land research, engineering, legal review and expert studies. Those teams assess what may be possible. Place diligence records the people, local bodies, services and public process a project must work through.
No commission. No brokerage. No side of the table.
We do not broker land, earn a commission, operate a marketplace, or have a stake in whether a transaction closes. The assessment says what the data supports, including when the answer is to pause, condition, or not advance.
Independence does not make this a legal conclusion, engineering feasibility study, insurance opinion, or permit determination. It makes the evidence package suitable for deciding which of those specialists to engage next.
What our site diligence examines
Community Readiness Score
We keep three things separate: underlying civic patterns, dated local or project activity, and the site's exposure to land-use change. These help guide review. They do not predict a permit decision.
Emergency Response Capacity
We screen local fire and hospital response for the site's needs. Rural land may have power and a low price but still need a closer look at local services.
Community & Economic Impact
We model what changes when a site opens, using similar operating sites as a comparison. This work uses our own sub-market planning geographies.
Acoustic Diligence Trigger
Nearby homes or other sensitive uses may call for a noise study before design or permit work. We flag when an expert should measure and model noise. We do not sell acoustic modeling.
Grid Interconnection
We check nearby power lines, substations and pressure in the generator interconnection queue. Each finding has an evidence tier. These are screening signals, not proof of spare substation capacity or a place in a utility's large-load service queue.
Water Stress
We review cooling water, aquifer reliance, water rights and scarcity. We compare the baseline, current 30-day conditions and documented drought-and-heat stress cases. Yearly averages can hide peaks: high cooling demand and drought can arrive together.
Climate & Air Quality
Temperature stress and air quality exposure that surface in permitting.
Legislative & Power Independence
Active data center regulation, ratepayer-cost legislation, tariff exposure, and behind-the-meter optionality.
Who We Serve
Underwriters and insurers. Lenders and credit teams. Infrastructure and private capital diligence. Entitlement and land use counsel. CRE development teams and land assemblers. Construction and corporate security leadership. REITs. Landowners evaluating their own acreage.
Any stakeholder whose process benefits from a structured, independently derived site assessment sitting alongside developer-supplied projections.
Jack Hall — Founder
Jack has spent two decades building data and map-based tools from scratch. He has led analytics teams and built first-of-kind predictive models for retail, hotels, finance and healthcare. As a SaaS product manager, he helped turn those tools into multi-year enterprise partnerships.
He built the hospitality vertical that produced a 14-year Marriott partnership, later extending to IHG, Sonesta, and Peregrine Hospitality. He built the plasma vertical that produced partnerships with Grifols, ImmunoTek, BioLife Plasma, Octapharma, and Kamada Plasma. He led the strategic account and analytics team whose whitespace and household-level member analysis underpinned the S-1 filing of a major U.S. fitness brand.
You work directly with him. The methodology isn't handed off to a junior analyst, and the person who built it is the person who signs the report.
Credentials
- AI/ML for LeadersUT Austin, McCombs School of Business (2026)
- SMU Big Data Advisory BoardMember (2017–2018)
- Bisnow Lodging Investment SummitKeynote Speaker (2021)
- Wharton Customer Analytics ConferenceSpeaker (2014, 2015)
Public methodology
How the screen works, and what it cannot tell you
This page describes the method at the level needed to read a report responsibly. It does not publish coefficients, thresholds or formulas, and it is not a guarantee of any outcome.
Site Viability: 0 to 100, higher is better
A weighted screen that summarizes several evidence groups. It is not a probability of approval, financing or interconnection.
Exposure sub-scores: lower is better
Climate, water and similar exposure measures run the other way: a lower number means less exposure. Each report labels the direction next to the value.
What contributes to the composite
- Climate and air
- Heat stress, drought and air-quality exposure that can surface in permitting and operations. Inputs can include modeled weather data (Open-Meteo), NOAA National Weather Service and USGS feeds, EPA AirNow observations and EPA county annual AQI as a labeled fallback.
- Grid and power context
- Proximity to transmission and substations, generator-interconnection queue pressure and regional energy records (Berkeley Lab Queued Up, EIA). Proximity is not proof of spare capacity, a queue position or time to power.
- Water
- Baseline water stress and cooling-water context, drawing on mapped layers such as WRI Aqueduct. Annual averages can hide seasonal peaks, so stress cases are shown separately where supported.
- Regulatory and community
- Active data center regulation, ratepayer and tariff policy, local review steps and documented public activity. Contextual area proxies are kept separate from dated, project-linked records.
Reading rules
Missing is not favorable
If a top-level input is unavailable, it can be omitted and the remaining contributions rebalanced. Missing evidence is not evidence of good conditions, even if the summary score looks favorable. Check the report for coverage and fallback notes; if a source or date is absent, treat it as unknown rather than assuming it was verified.
Close scores are not a ranking
Two sites a few points apart may differ mainly in which inputs were available. Inspect coverage, scale and dates before reading a difference as meaningful. Compare only reports with coherent scope, date and coverage. We do not publish a ranking universe we cannot prove is complete.
Supplemental signals stay outside the score
Emergency response capacity, workforce and labor context, and nearby receptor or noise-study triggers are shown where available. They do not change the composite Site Viability score or the compliance certificate.
Fields to inspect on each cited source
Use these fields to judge a citation. Not every source or report populates every field; treat a missing field as unknown, not as current or reliable, and do not read a generic confidence label as a substitute for these details.
| Field | Why it matters |
|---|---|
| Publisher and URL | Who published the dataset and where it can be inspected. |
| Scale | Site point, block group, county, utility territory, ISO/RTO or state. Area data is not parcel data. |
| Evidence type | Observed, mapped or modeled. Each is labeled; they are not interchangeable. |
| Vintage | The period the publisher says the data describes. |
| Retrieved and scored dates | When we pulled the data and when it entered a score. These differ from vintage. |
| Expected cadence vs. last refresh | How often the publisher updates, and when our last successful refresh ran. A source can be behind schedule. |
Community proxies vs. project evidence
Area-level civic and demographic patterns are contextual proxies. Dated hearing records, filings and project-linked public activity are evidence about a specific process. Neither proves local support or predicts an approval.
Automated report vs. analyst-assisted work
The scorecard is generated automatically. Analyst-assisted review and the Community Engagement Playbook are scoped separately. Not every report is checked by a person unless that review is part of the agreed scope.
What independence means
We earn no sale commission and hold no brokerage or transaction stake. Independence is not an external certification, audit or legal opinion.
Validation research
Our retrospective case study applies the current model to past sites. It is descriptive, not proof that the model forecast those outcomes at the time. See each study on the research page for its own design and limitations.
Disclosure context: report fields are organized to be usable alongside climate-disclosure frameworks such as IFRS S2 and California SB 253 and SB 261. Use in a disclosure does not make the report a formal disclosure opinion. See the illustrative sample report for how these rules appear on the page.