CFE: Why We Built an Institutional-Grade Data Center Site Intelligence Platform
Data center conversion is a trillion-dollar land grab. We built CFE to give developers the same intelligence that billion-dollar funds use to find sites — substations, fiber, gas pipelines, and AI-powered scoring.
The data center industry is in the middle of a land grab. AI training clusters need 100MW+ of power, and every hyperscaler — Microsoft, Google, Amazon, Meta — is racing to secure sites with the right combination of power, fiber, water, and zoning.
The problem? Finding these sites is still done with spreadsheets, Google Maps, and relationships. A developer might spend 6 months evaluating a site only to discover the substation can't support the load, or the fiber is 12 miles away, or the gas pipeline doesn't have enough capacity for backup generation.
We built CFE to fix that.
What CFE Does
CFE (Conversion Feasibility Engine) is an intelligence platform for data center conversion site analysis. It combines geospatial data with infrastructure intelligence to score potential sites instantly.
Here's what it maps:
- **Substations** — Voltage, capacity, headroom status, and withdrawal limits. ERCOT data with PACES priority flagging.
- **Fiber POPs** — Carrier locations and proximity scoring. You need lit fiber within 5 miles or your site is dead.
- **Gas Pipelines** — Transmission and distribution lines, diameter, pressure class. Critical for backup generation.
- **Properties** — Commercial and industrial properties scored against conversion criteria: square footage, year built, zoning, proximity to infrastructure.
Each site gets a conversion feasibility score based on weighted proximity to power, fiber, and gas, combined with property characteristics and market conditions.
The Scoring Engine
When you sweep an area, CFE finds every commercial property within your radius and scores it:
- **Green** — Strong conversion candidate. Power within 2 miles, fiber within 5, adequate square footage, compatible zoning.
- **Yellow** — Feasible with upgrades. Maybe the substation needs a new feeder, or the property needs structural work.
- **Killed** — Fatal flaw. No power path, wrong zoning, flood zone, or site too small.
Each scored property gets a detailed breakdown: distance to nearest substation, voltage available, fiber carriers in range, gas pipeline access, and estimated conversion costs.
Pro Forma Generation
Beyond scoring, CFE generates site-specific pro forma financials:
- Power capacity estimates based on substation headroom
- Construction cost modeling for shell conversion
- Operating cost projections including power, cooling, and connectivity
- Revenue modeling across customer segments (hyperscale, colo, edge)
- Emissions calculations and sustainability scoring
The output is a PDF-ready scorecard that a developer can hand to an investor or lender.
Why We Built This
Greenbelt's founder spent years in the hyperscaler and energy space. The gap between how sites SHOULD be evaluated and how they actually ARE evaluated is massive. Billion-dollar funds have teams of analysts doing this work manually. Smaller developers fly blind.
CFE democratizes that intelligence. The same analysis that takes a team of analysts weeks to compile — infrastructure proximity, capacity assessment, financial modeling — is available in seconds.
The Market
Data center construction spending is projected to exceed $300 billion annually by 2028. Every MW of capacity needs a site, and every site needs this analysis. The developers who find and secure sites fastest win.
Right now, the tooling is fragmented: - Power data lives in ISO/RTO databases - Fiber maps are behind carrier paywalls - Gas pipeline data is scattered across DOT filings - Property data requires commercial real estate subscriptions - Financial modeling happens in Excel
CFE brings it all together in a single map-based interface.
What's Next
CFE already covers ERCOT and PJM markets, with MISO in development. The platform includes transmission upgrade tracking, failed project analysis (why sites were abandoned), and nodal price data for power cost modeling.
The goal is simple: make it possible to evaluate a data center site in 5 minutes instead of 5 months.
If you're developing data center sites, or evaluating conversions, try it at cfe.projectgreenbelt.com.