Site-control funnel
Buildable acres by stage
Est. acquisition cost by stage
Candidate parcels — site control
Synthetic shortlist. Click a row to locate it on the map. Verdicts are authored parcel-by-parcel, not generated; flag thresholds are illustrative.
| Parcel | County | Status | Buildable ac | $/ac | Title | Suitability | Speed to power | Verdict | Updated |
|---|
Deal sheet illustrative · basis-stated
All deal figures are synthetic and illustrative; assumptions are stated per panel. Not investment advice.
Power & timeline illustrative · basis-stated
Speed-to-power is a disclosed phase-model estimate gated on the interconnection queue — not a scraped commissioning date. Synthetic + illustrative.
Executive memo illustrative · basis-stated
An IC-style one-pager for the parcel. Download PDF opens your browser's print dialog — choose Save as PDF.
Option-budget portfolio illustrative · basis-stated
Synthetic + illustrative. Option capital deployed now; diligence staged as power validates. Not investment advice.
Methodology & About
The model
Every rural parcel ≥ 50 acres is scored with a transparent weighted overlay: each criterion is normalized
to 0–100 against absolute, domain-anchored breakpoints, weighted, summed, and ranked. Interconnection leads,
because in ERCOT that is the constraint that makes or breaks a site. All weights and breakpoints live in a
single config.yaml — illustrative and tunable; a production-tuned model is discussed in
conversation.
| Criterion | Weight | How it scores |
|---|---|---|
| Interconnection proximity | 35% | Distance to nearest ≥138 kV substation/line (≤1 mi → 100, ≥10 mi → 0) |
| Buildable acreage | 20% | Acres after exclusions/setbacks (50 ac → 0, ≥640 ac → 100) |
| Terrain | 15% | Mean slope (≤2% → 100, ≥5% → 0; very steep excluded) |
| Land cover + soils | 10% | Open/marginal land favored; prime farmland (NRCS LCC 1–2) penalized |
| Hazard-free | 10% | % of parcel outside floodplain/wetlands |
| Road access | 5% | Distance to paved public road |
| Contiguity / shape | 5% | Polsby-Popper compactness |
Exclusions & setbacks (subtracted before scoring): FEMA SFHA high-risk floodplain, NWI wetlands, PAD-US protected lands, plus Eagle Ford setbacks from Texas Railroad Commission data — 150 ft around active wellheads, 50 ft around pipelines — and 100 ft parcel-boundary / 150 ft right-of-way setbacks.
Two more lenses
The same parcels can be re-scored for flexible / large load (interconnection headroom, proximity to generation, demand-response value) and for whole-site dual use (agrivoltaics — grazing compatibility, marginal-soil preference, pollinator habitat, soil-and-water stewardship). The base model and flex-load lens reward open, buildable, non-prime land; the agrivoltaic lens rewards working ag land. Same ground, three value systems.
Why this geography
San Antonio's data-center load is on track toward 3,300+ MW by 2033, with CPS Energy planning upwards of $1.3B in transmission to keep up. Moving up the corridor into Milam County, Rockdale (Riot Platforms) leveraged its ~700 MW on-site substation — historically used for crypto mining — to anchor a new AI data center hub via a ~$311M, 10-year AMD lease signed in January 2026. Interconnection-aware land screening is the live problem across this entire central Texas grid infrastructure corridor.
Data sources & restricted-data discipline
All layers are public and cited with retrieval dates in data/SOURCES.md (TxGIO parcels; HIFLD
transmission & substations; FEMA NFHL; USGS 3DEP/NLCD; USDA-NRCS soils; USFWS NWI; USGS PAD-US; Texas RRC
wells/pipelines; ERCOT/EIA generation context; USDA NAIP imagery). No MLS/IDX data is used. The
license-restricted ERCOT network model is never displayed raw — any capacity signal is abstracted to a 0–100
"Grid Connectivity Index"; the public map overlay uses HIFLD.
| Layer | Source | In this build |
|---|---|---|
| Parcels | TxGIO StratMap (county appraisal districts) | live |
| Substations · transmission | HIFLD · DOE-NETL Energy Atlas | live |
| Terrain (slope) | USGS 3DEP | live |
| Land cover | USGS / MRLC NLCD 2021 | live |
| Soils (capability class) | USDA-NRCS SSURGO · Soil Data Access | live |
| Floodplain | FEMA NFHL (SFHA zones) | live |
| Roads | U.S. Census TIGER | live |
| Generation | EIA power plants | live |
| Aerial basemap | USDA NAIP / Esri World Imagery | live |
What's new
Shipped, newest first — a running build log for the screen and tracker.
- Guided tour. A "Start here" walkthrough — the model ranks and I decide on the map, then the deal sheet, speed-to-power, and the memo, in four steps. First visit opens it once; the button in the header reopens it anytime.
- Pipeline on the map. The pipeline board IS the map: a fourth coloring mode shows every shortlist parcel colored by site-control stage, with a clickable stage strip to filter and everything else dimmed. Each pipeline parcel's popup now names its binding constraint — the one thing that kills or gates it right now — and the next action, authored parcel-by-parcel. Deep link:
?view=pipeline-map. - Executive memo. A per-parcel IC-style one-pager — verdict, the underwrite, speed-to-power, the diligence machinery, the feasibility gates, and an honest read on incentives + community posture. Built to print: "Download PDF" → save the memo. The artifact an interviewer can hold.
- Power & timeline. A per-parcel speed-to-power estimate — P50 months with an honest band — and the energization phase behind it. A disclosed phase model gated on the interconnection queue, modulated by each parcel's grid quality, with a co-location fast-path for parcels near existing generation. The restricted ERCOT queue stays backend-only; context is public.
- Data integrity. The screen audits itself: source freshness and license posture, run QA (all seven criteria live), field completeness, and known gaps — flagged, not hidden.
- Portfolio view. Given an option budget ($500K / $1M / $1.5M), how I'd deploy it across the shortlist — spread options to control the best parcels, stage diligence on the speed-to-power winners. The model ranks; I allocate.
- Deal sheet. Per-parcel development-margin economics — flip / ground-lease / JV side by side, a time-to-power × exit-value sensitivity table, and budget-vs-actual on the synthetic pipeline. Synthetic and illustrative; every assumption stated.
- Verdict layer. The model ranks; I decide. Authored pursue / pursue-if / pass calls on the shortlist, objective risk flags computed on every parcel, and a worked memo on why I'd pass on the #3 parcel in my own model.
- Data integrity. De-duplicated stacked parcel records from the statewide layer — the same physical parcel was being counted several times. 6,080 → 5,130 distinct parcels.
- Homepage + contact. consultruss.com now features the demo; the header links back to the homepage and to LinkedIn.
- Launch. Every parcel ≥ 50 acres in Wilson & Karnes scored from live public data; Ask-the-map; site-control tracker.
Data integrity
What's real, what's modeled, what's stale, and what I won't republish — computed from the run artifacts.