Developers vs. the Neighbors: His Company Wants to Put a Number on It

Before choosing a site, developers can model power, water, land, labor, taxes, incentives and transmission. What residents will do once the project is announced usually gets handled by consultants, public meetings and instinct.

Jeremy Solomon's company, Learnewable, is trying to turn that into a number. The Portland, Ore., software firm scores locations for solar, wind, battery storage, transmission and data center projects, and each month it scores every county in the country. After about a year of those county runs, Solomon says data centers stand apart. "It's gotten quite a bit worse," he said.

A user enters an address, parcel or county, picks a project type and size, and gets a favorability score. The platform calculates a decimal score and labels it on a scale from −2 to +2. Solomon says 40% of the score reflects community sentiment, 40% the regulatory picture, meaning zoning and permitting, and 20% the area's history with similar projects. He doesn't claim the weighting is settled. "You could come up with a different one that might be better," he said.

The sentiment piece comes from several AI language models reading Google results and social platforms including Facebook, Reddit and Nextdoor. Solomon says the models check one another to weed out errors. The platform also identifies local stakeholders. About six weeks ago it added a legislative tracker and Podium, a tool that analyzes video of public meetings.

Solomon started Learnewable about three and a half years ago, after selling Solomon Exam Prep, the securities-licensing test-prep company he co-founded, to CeriFi in 2022. At renewable energy conferences, he found developers had plenty of technical data and very little about the communities they were trying to build in.

Over the past year, Solomon says, sentiment toward solar and wind has improved and battery storage has stayed mostly positive. Data centers still lean favorable in most places. Learnewable's September run found favorable sentiment in two-thirds of counties. Only 42 of 3,144 counties rated favorable once local rules were factored in, though, and the share of counties with regulatory restrictions rose to 31% from 26% in July.

Two patterns stand out in his data. Counties with more existing data centers tend to be more negative about adding another, and Solomon said "that is for real." More populated counties are also less receptive than rural ones. Neither pattern proves cause and effect. The first matters for places trying to build a cluster, where the second project may be easier to land than the fifth.

Learnewable often finds local officials and residents pointing in different directions. In September, 363 counties had favorable sentiment and restrictive zoning. "Increasingly, local rules, not public sentiment, are the bigger barrier," Solomon said in the company's release. A zoning map shows what can be built, but it can't predict what happens when someone tries.

Solomon says the opposition and the worries behind it are one and the same: residents object to data centers because of what they fear about power, water and tax breaks. He also sees developers holding back what residents want to know early: how much water and electricity a project will use, what it means for their rates, and which incentives are on the table. He is also skeptical of the nondisclosure agreements local officials are often asked to sign, which he describes as a personal view and not something the data shows.

Solomon is careful to say he isn't a data center expert, since most of Learnewable's history is in renewables. There, he says, developers see less pushback when they avoid NDAs where they can, show up at public meetings, share their environmental studies, hire local crews and offer a community benefits plan.

He pointed to Mammoth Solar, Doral Renewables' 1,300-megawatt project in Pulaski and Starke counties in Indiana. Solomon noted that Doral isn't a Learnewable customer, and said his company's analyses suggest Doral won local support. Doral has made economic development payments to Pulaski County and grazes sheep among the panels. The project also drew organized opposition and a lawsuit over its zoning approval. Engagement lowered the temperature without ending the fight.

Site selection can put numbers on workforce, taxes, utilities, logistics and acreage. Most developers still judge community acceptance by feel, even though a strong site can lose years to hearings and lawsuits. Solomon says his platform tracks sentiment for each stakeholder over time and weighs how much influence each one has. Much of the opposition also goes beyond NIMBYism. Residents raise legitimate questions, developers communicate poorly, and some communities simply don't want a particular use.

Solomon is candid about the limits of social data. Social media skews negative, some of it is fake, and no model captures every local relationship. People rarely post to say they love solar, he noted. To reach residents who don't post online, Learnewable built Podium to analyze video of public meetings. Contested Ground will keep asking whether community acceptance can be measured well enough to sit beside power and water in a site search.