SiteScore™ & PowerMap™
The proprietary intelligence framework behind every Meridian engagement — four core domains, 100-point composite scoring, PRIME through WATCH tier classification.
Every site ranking, corridor recommendation, and advisory output Meridian produces is grounded in GridMind Inc.'s proprietary intelligence systems. This page documents the methodology — the scoring criteria, classification logic, and demand forecasting inputs that determine which nuclear-adjacent AI power sites Meridian advises clients to pursue and which to avoid.
Across 4 proprietary institutional domains
Nuclear-adjacent parcels in the 2026 dataset
PJM Core · Midwest · Southeast · Sun Belt
Delaware C-Corp · Licensed to Meridian Nuclear LLC
A 100-point composite system.
Four domains. No narrative substitutes.
SiteScore™ evaluates every nuclear-adjacent AI power site on four institutional domains. Each domain is broken into specific scored sub-criteria with defined ranges. The composite score determines a site's position in the GridMind national ranking and drives every advisory recommendation Meridian delivers.
Power Infrastructure
Evaluates the quality, accessibility, and capacity of transmission and power delivery infrastructure at and near the site. This domain rewards parcels with direct, near-term access to high-capacity transmission — eliminating the single most common failure mode in AI power site development.
Proximity to existing high-voltage transmission lines (230kV, 345kV, 500kV), direct substation access, and documented interconnection pathway. Tier 1 (direct access, ≥500kV within 2 miles) scores maximum. Tier 3 (no direct access, new line required) scores 0–2.
Analysis of the site's position in the relevant RTO/ISO interconnection queue (PJM, MISO, SERC, WECC). Sites with existing queue positions, completed studies, or legacy interconnection rights score highest. New greenfield queue filings in congested markets score 0–2.
Whether the adjacent nuclear facility is operating (maximum score), in license extension review, in decommissioning transition (partial credit), or fully decommissioned (infrastructure value only). Operating plants provide ongoing power procurement and grid stability benefits that decommissioned plants cannot.
Rated capacity of nearest accessible substation and available headroom for AI-scale load additions (typically 200–600 MW). Sites adjacent to nuclear plant substations with documented excess capacity score maximum. Sites requiring new substation construction score 0.
Quality of the regional utility's relationship with AI infrastructure operators and history of PPA execution for large-load customers. Utilities with documented hyperscaler off-take agreements score maximum. Utilities with no history of AI infrastructure engagement score 0–1.
A top-tier Power Infrastructure site has direct substation access, a clear interconnection pathway, an operating adjacent nuclear facility, and a utility with documented AI off-take history. This profile represents fewer than 12% of evaluated sites.
AI Demand Proximity
Evaluates the site's proximity to existing and planned AI compute demand — hyperscaler campuses, major metro fiber networks, and the workforce pools required to operate and maintain large-scale data centers. This domain distinguishes sites that can be monetized in the near term from those with only long-horizon optionality.
Distance to the nearest existing hyperscaler or large-scale AI data center campus (Amazon, Google, Microsoft, Meta). Sites within 50 miles of an operating campus benefit from established fiber routes, operator familiarity with the market, and precedent off-take structures. Sites >200 miles from any operating campus score 0–3.
Availability and density of long-haul fiber routes within 10 miles. AI data centers require multi-100Gbps fiber connectivity for inference workloads and data replication. Sites on or adjacent to major fiber trunk routes (e.g., I-95, I-90, I-70 corridors) score highest. Sites requiring new fiber construction score 0–2.
Volume of publicly announced or credibly rumored hyperscaler expansion activity within a 100-mile radius. Active expansion markets with multiple announced projects score maximum. Markets with no announced activity and no precedent score 0–1.
Proximity to metro areas with established data center and electrical engineering labor pools. MSA population within 90-minute commute, presence of relevant technical programs (community colleges, universities), and existing data center employer density in the market. Rural isolation without workforce precedent scores 0–1.
Whether a large-scale (100 MW+) co-location or PPA agreement has been executed within the same nuclear plant cluster or corridor. Precedent transactions reduce counterparty risk perception and accelerate deal structuring timelines. This criterion does not add points but can elevate a borderline site from one tier to the next in the final ranking.
Development Readiness
Evaluates how quickly and at what cost a parcel can be converted into an operating AI infrastructure campus. This domain penalizes environmental risk, permitting complexity, and water scarcity — the three most common development-stage failure modes in nuclear-adjacent AI power projects.
Annual available water volume for cooling purposes. AI liquid-cooled clusters require 50–120 million gallons/year per campus. Sites adjacent to nuclear plants on major rivers or reservoirs (with established water rights infrastructure nearby) score highest. Arid sites requiring municipal water expansion score 0–2. Water rights documentation and existing intake infrastructure are positive modifiers.
Current zoning classification and entitlement progress for industrial/utility-scale development. Sites with existing industrial or heavy commercial zoning score maximum. Sites in agricultural or residential zoning requiring full rezoning with environmental review score 0–2. Partial entitlements, conditional approvals, and industrial park adjacency score in the 3–5 range.
Assessment of wetlands, floodplain, endangered species habitat, NEPA review requirements, and proximity to protected areas. Clean sites with no identified constraints score maximum. Sites with identified wetland or NEPA exposure score 0–3 depending on severity and mitigability. This is a penalty criterion — identified constraints reduce scores.
Physical readiness: cleared and graded land, existing road access, preliminary utility stub-outs, and geotechnical studies. Sites with completed Phase I environmental assessments, existing road infrastructure, and cleared acreage score 3. Undeveloped woodland or uncleared agricultural land with no preliminary work scores 0–1.
Usable acreage for AI campus buildout. AI data center campuses targeting 500 MW+ draw typically require 500–2,000 acres including setbacks, cooling infrastructure, and expansion land. Parcels with ≥1,000 usable acres score maximum. Smaller parcels (<200 acres) score 0 unless adjacent acquisition is highly feasible.
In practice, water access is the single most differentiating Development Readiness criterion across the GridMind dataset. Sites without credible access to 50M+ gallons/year rarely score above 60 overall, regardless of their Power or Policy scores. Nuclear plant adjacency is the primary driver of water access eligibility.
Policy Environment
Evaluates the regulatory, legislative, and institutional conditions that determine how quickly and at what risk a site can be developed and financed. This domain captures the difference between states actively competing for nuclear-adjacent AI infrastructure and states where regulatory friction adds years and capital to development timelines.
Whether the state has enacted pro-nuclear legislation, established nuclear development incentive programs, or created regulatory fast-track pathways for nuclear-adjacent infrastructure. Illinois, Pennsylvania, South Carolina, and Georgia score highest. States with nuclear moratoria or hostile utility commissions score 0–2.
State public utility commission attitude toward AI infrastructure load growth, large-load co-location agreements, and nuclear power purchase arrangements. Commissions that have approved precedent hyperscaler PPAs or issued favorable large-load rulings score highest. Commissions with rate base protection conflicts score 0–3.
Eligibility for and probability of capturing federal incentives: DOE Title XVII Loan Guarantee Program, IRA Production Tax Credit for nuclear energy, EDA infrastructure grants, and CHIPS Act-adjacent data center provisions. Sites with documented application pathways or precedent awards score maximum.
The site's position relative to FERC transmission rules, particularly Order 1920 (December 2025) cost allocation and transmission planning requirements. Sites in regions where Order 1920 reduces interconnection cost exposure score highest. Sites in non-compliant or late-implementing regions score 0–2.
County and municipal authority posture on AI infrastructure development: active economic development recruitment programs, available tax abatement structures, and precedent industrial incentive agreements. Counties with active AI/data center recruitment programs score maximum. Jurisdictions with no industrial development history score 0.
GridMind's December 2025 dataset update incorporated the FERC Order 1920 compliance status of all sites. Parcels in PJM and MISO territories — which implemented Order 1920 earliest — received FERC compliance score upgrades that materially shifted corridor rankings. This was one of the largest single-cycle score updates in the dataset's history.
Not all nuclear-adjacent parcels score the same.
Distance and infrastructure inheritance matter precisely.
GridMind's nuclear adjacency classification determines the base tier of a site's Power Infrastructure and Development Readiness scores before individual sub-criteria are applied. Adjacency tier is not a separate score — it is a structural input that sets the upper bound on what a site can achieve in both domains.
Direct Nuclear Adjacency
Parcels with direct adjacency inherit the highest benefits of nuclear proximity: direct substation access to the plant's transmission infrastructure, established utility relationship with a nuclear operator, proximity to water intake infrastructure, and the reputational and regulatory context of a proven nuclear site. These parcels are the rarest and most competitively sought in the GridMind dataset.
Corridor Positioning
Corridor tier parcels maintain most of the strategic benefits of nuclear proximity: they fall within the same transmission planning zone, share utility relationship context, and are typically within the water infrastructure service area of the nuclear facility. These parcels represent the broadest opportunity set in the GridMind dataset and include most of the Top 50 ranked sites.
Proximity Positioning
Proximity tier parcels retain some nuclear infrastructure benefits — corridor-level transmission quality, state policy environment, and utility operator familiarity — but lose direct infrastructure inheritance advantages. Water access scores in this tier are determined primarily by independent hydrology rather than nuclear plant infrastructure proximity. These parcels can still rank in the Top 50 if their Demand Proximity and Policy scores are exceptionally strong.
Corridor position is a multiplier.
A high SiteScore in a Tier II corridor is not the same as a high SiteScore in Tier I.
PowerMap™ ranks the four active U.S. nuclear-adjacent AI power corridors against six criteria. Corridor tier is applied as a strategic context multiplier to SiteScore rankings — a site's raw score is interpreted differently depending on whether it sits in a Tier I or Tier II corridor. This prevents overvaluation of high-scoring sites in under-capitalized markets.
PowerMap™ corridor rankings are updated quarterly as transmission infrastructure develops, new hyperscaler demand signals are identified, and policy environments evolve. The Tier I / Tier II designation reflects current actionable opportunity, not long-term potential. A Tier II corridor site with a SiteScore of 82 may be more strategically valuable than a Tier I corridor site with a SiteScore of 74, depending on the client's mandate, time horizon, and counterparty access.
SiteScore is backward-looking. Demand forecasting is where the real intelligence lies.
The SiteScore composite reflects a site's current characteristics. Demand forecasting adds a forward-looking layer: which corridors and sites are likely to receive hyperscaler attention in the next 18–36 months, and what signals indicate that attention is already moving. This layer is what allows Meridian clients to position before the market sees what GridMind has already mapped.
Hyperscaler Capex Signal Tracking
GridMind monitors the capital expenditure programs of the five largest AI infrastructure operators (Microsoft, Google, Amazon, Meta, Oracle) and cross-references announced project locations against the SiteScore dataset. When a hyperscaler announces a campus in a corridor, adjacent sites that score 70+ receive elevated demand signal flags in the dataset — indicating near-term counterparty engagement probability.
- Quarterly capex announcement review across 5 primary hyperscalers
- Cross-reference against FERC large-load interconnection filings
- Campus location mapping against Top 50 site coordinates
- Demand signal flag assigned when campus announced within 100 miles of ranked site
- Flag intensity tiered by campus size announcement (100 MW / 500 MW / 1GW+ thresholds)
FERC Interconnection Queue Analysis
GridMind tracks all large-load (≥50 MW) interconnection queue filings in PJM, MISO, SERC, and WECC territories and maps them against nuclear-adjacent parcels within 30 miles. Queue activity near nuclear sites — even filings without named counterparties — is treated as a leading indicator of developer or operator interest in the corridor.
- Bi-annual queue scrape across all four RTO/ISO territories
- Geocoding of queue filing locations against nuclear plant coordinates
- Statistical clustering to identify corridor-level activity patterns
- Queue velocity metric: rate of new filings per corridor per quarter
- Withdrawal rate analysis to identify queue activity that converts vs. cancels
State Policy Leading Indicators
GridMind monitors state legislative calendars, utility commission docket activity, and economic development authority announcements for signals that precede infrastructure investment by 12–24 months. Nuclear investment tax credit expansions, large-load customer tariff reforms, and nuclear plant license extension rulings are all classified as leading indicators and trigger corridor score updates.
- Quarterly state legislative tracker across 15 nuclear-adjacent states
- Utility commission docket monitoring for large-load and co-location rulings
- Nuclear plant NRC license extension application tracking
- State economic development AI infrastructure recruitment program monitoring
- IRA and DOE program application deadline and award tracking
Land Market Velocity Indicators
GridMind tracks parcel-level transaction activity, listing velocity, and off-market inquiry signals in nuclear-adjacent corridors. Unusual transaction activity near nuclear sites — especially institutional buyer acquisition of multi-hundred-acre parcels — is treated as a leading indicator of AI power site positioning and triggers competitive pressure flags in the dataset for nearby ranked sites.
- Quarterly parcel transaction review in all four corridors
- Buyer type classification (institutional, developer, holding company, utility)
- Transaction size analysis relative to historical corridor averages
- Off-market inquiry signals from broker network contacts
- Competitive pressure flag assigned when institutional buyer activity spikes ≥2 standard deviations above corridor average
How SiteScore™ translates to strategic classification.
Corridor-aligned, high-readiness sites with strong institutional viability across all domains
Strategically positioned sites with notable infrastructure and demand fundamentals
Sites with measurable nuclear-adjacent characteristics warranting continued monitoring
Early-stage corridor positions under active intelligence review
The full SiteScore rankings are available through the Top 50 AI Power Sites report and Meridian advisory engagements.
This page documents the methodology framework. The actual site rankings, corridor maps, demand signal flags, and parcel-level data are only available through GridMind intelligence products — the Top 50 report at $3,500 or a Meridian advisory engagement.
