Community Toolkit · Sample
Every data center project meets the same questions. We measured which ones people are really asking: 130 search terms across ten concern categories, plus the hearing and docket record where search goes quiet, and wrote a model answer for each, written from the project’s side. Every category is labeled with the measure that ranked it. Take them, adapt them, put them on your project site.
Largest category by a factor of four. "data center water usage" alone runs 9,900/mo, up 654% year over year.
Less than most people expect. Virginia's legislative watchdog reviewed the state's data centers and found most use about as much water as a large office building. A few use much more, and some use less than a single household. [Project] is permitted for [X] gallons a day. That is the ceiling, not the daily draw.
No. That figure is off by roughly a hundred times. It came from an early estimate that folded in water used at distant power plants and assumed older cooling technology. A more accurate figure is closer to a teaspoon. We would rather you had the real number than the viral one.
[Project] draws from [source]. Data centers run on potable, reclaimed or non-potable water depending on the site, and reclaimed or air-cooled designs are increasingly common because they cost less to operate. Ask any developer which one they have chosen and where the water comes from.
Wells and municipal supply usually draw from different sources, but “usually” is not good enough when it is your water. [Project] connects to [system], not the aquifer your well taps. We have committed to baseline well testing before construction so there is a record to compare against. Call [contact] to have your well included.
The same restrictions as everyone else. [Project] is subject to [utility]'s drought contingency rules, and our cooling design can shift to [reduced draw] when Stage [X] restrictions are called. A data center that gets a carve-out during a drought is a data center that loses its community.
People search what data centers consume, not what it costs them. "data center electricity usage" is 2,900/mo; "do data centers raise electric bills" is 210.
[Project] is designed for [X] megawatts at full build, roughly the demand of [X] homes. That is a real number and it is large. What matters locally is that new large loads pay for the generation and transmission they require, which is set in [utility]'s rate case at [commission], a public proceeding you can read.
It should not, and here is the mechanism that decides it. Large loads like [Project] are served under [rate class], designed so we pay the cost of the capacity we use rather than spreading it across residential customers. Virginia's legislative review found data centers there currently pay their full cost of service. Watch [docket]. That is where it is settled.
We pay [rate class] rates and the terms are in [docket]. Some large-load contracts elsewhere are confidential, which is a fair thing to be suspicious about. Ours is [public at link]. If a developer will not tell you what rate class they are in, ask why.
No. [Project] connects under an interconnection agreement requiring the grid capacity to exist before we energize. We cannot draw power that is not there. [Utility] plans generation and transmission years ahead. Residential customers are not behind us in line during an outage.
We do, for the ones built to serve us. Under [utility]'s directly assigned transmission rules, the cost of interconnecting a large load is assigned to that load rather than the general ratepayer base. This is the most important question on this list and you should ask it of every project in [county], not just ours.
"data center noise" is up 1,995% year over year, the second-fastest climber in the whole set.
At the property line, about [X] decibels, comparable to [reference]. Inside the building it is much louder; outside it should not be. [Project] is designed to [X] dBA at the nearest residential boundary, below [county]'s limit of [Y]. We will publish the post-construction measurements, not just the model.
That depends on the equipment and the terrain, not the fence line. Cooling equipment is the constant source; generator testing is occasional and much louder. [Project] sites its chillers [X] feet from [boundary] behind [barrier]. If you can hear us from your property, we want to know. [contact] reaches someone who can act on it.
Low-frequency sound from cooling equipment, and it is real, not something residents imagine. Standard meters use A-weighting, which discounts low frequencies by more than 26 dB, so a facility can measure as compliant while people still hear it. [Project] commits to low-frequency measurement, a higher bar than [county] requires.
Testing is [monthly] for [30-60] minutes during [weekday daytime hours], never overnight. Beyond that they run only in an actual outage, which for most facilities is zero to two times a year. We will publish the test schedule in advance at [link] so nobody is surprised by it.
Call [number]. It reaches [role], not a general mailbox. [Project] is bound by [condition] at [X] dBA and we have agreed to independent measurement on complaint. A $500 ordinance fine does not fix anyone's sleep, so the remedy we committed to is [mitigation]. Hold us to it.
"data center moratorium" is up 17,900% year over year, the fastest-moving term in all 130.
Yes. [Board] votes on [application] on [date] and it can be denied. If the parcel is already zoned [X], some uses are permitted by right and the board's discretion is narrower. Either way the staff report is public at [link] and public comment is taken at [meeting].
No. [Project] is at [stage] of a [X]-step process with [board] action still ahead on [date]. We have been talking with [county] since [month] because a project this size takes years. Early conversation is not approval. Everything filed so far is at [link].
[Project] requires [rezoning / special use permit / by-right site plan review]. That distinction decides how much say the board has, so it is worth being precise. By right means the use is already allowed and review is administrative. A rezoning or special use permit means a public hearing and a vote. Ours is [X].
Notice went out by [method] on [date] to properties within [X] feet, which is [county]'s requirement, and it is often not enough. We added [extra notice] beyond what is required. If you are hearing about this for the first time now, that is a failure on our end and we would like to fix it: [contact].
More than most people expect: noise limits, setbacks, buffers, construction hours, lighting standards, road improvements, water source. These become enforceable conditions of approval, not promises. [Project] has proffered [list] at [link]. If something is missing that would matter to you, say so at the [date] hearing while it can still be added.
The head term is the most alarmist phrasing: "do data centers cause cancer," 480/mo and up 376%. Answering it directly is better than leaving it to circulate.
No credible evidence supports that. The concern traces to electromagnetic fields from high-voltage power lines, where research has looked at decades of residential proximity, not at data centers. [Project]'s substation and lines are [X] feet from the nearest home and meet [standard]. We would rather answer this directly than let it circulate unanswered.
Yes, and the measurable impacts are noise and light rather than emissions. Virginia's legislative review measured backup generators across the country's largest data center cluster and found they contribute under 4% of regional nitrogen oxides. The things worth watching are sound and lighting, and both are addressed at [link].
Combustion exhaust, during testing and outages only. [Project]'s [X] generators are permitted under [permit] with a cap of [X] hours a year, and actual use is typically [X]. They are backup, not a power source. A facility running generators as primary power is a different thing entirely and should be described that way.
We take that seriously enough to give you numbers rather than reassurance. [Project]'s permitted emissions are [X] against a [county] baseline of [Y], and the permit modeling is public at [link]. If [school] is the concern, [contact] can walk your school board through the dispersion analysis.
Fields fall off sharply with distance. At [X] feet, exposure from [Project]'s substation is comparable to common household sources. The research people cite involves living under high-voltage transmission for years, a different exposure. There is no US federal EMF limit; we design to [standard].
Light pollution (390/mo) outsearches every habitat and farmland term. The land concern is mostly a nighttime visual one.
Darker than a parking lot, if it is designed right. [Project] uses full-cutoff fixtures that throw light down rather than out, at [X]K color temperature, with reduced lighting after [hour]. DarkSky International publishes the standard we designed to. If you can see glow from your property after we open, that is a defect and we will fix it.
[X] acres, of which [Y] is building footprint and [Z] stays undisturbed buffer. For scale, all of Virginia's data centers sit on about 7,200 acres combined. The larger land question in most places is the generation and transmission built to serve them, and that is fair to ask about too.
[X] acres come out of production. That is a real loss and we will not argue otherwise. What we have done about it: [conservation easement / buffer / continued lease of X acres]. If [county] wants a farmland mitigation condition attached to this approval, that is a conversation we are willing to have.
[X] of the site's [Y] wooded acres come down; [Z] stays as buffer along [boundary]. The tree line facing [road] stays. Replanting is [X] trees under [condition]. The site plan showing exactly what is cleared is at [link].
Honest answer: nobody has decommissioned a facility this size yet, so anyone confident either way is guessing. What we can commit to is a [decommissioning bond / restoration condition] written into the approval, so the obligation survives whoever owns it later.
Almost no national search volume, but among the most frequent questions in hearing rooms and from real estate agents. Ranked by public-record frequency.
The research is mixed and you should see both halves. A George Mason study found homes nearer Northern Virginia data centers sold for more, reflecting the infrastructure and jobs that draw both. A George Washington study found homes within half a mile of a newly announced facility sold about 2.8% lower. Distance, noise and sightlines drive the difference.
Half a mile is where measurable effects show up in the research, and inside that ring what matters is whether you can hear it or see it. [Project]'s nearest residence is [X] feet, with [buffer]. If you are inside that ring, contact [name]. We would rather talk to you specifically than have you rely on a general answer.
In most states yes: proximity to a known development is a material fact. Ask your agent about [state]'s disclosure rules. It is also worth knowing that agents report homes near active construction more often take longer to sell than they sell for less, which is a different problem with different solutions.
That is between you and [county assessor], and it is a legitimate route: [process, deadline]. We do not control assessments and will not pretend to. If [Project] measurably affects your property, the appeal record is where that gets recognized, and we will not contest a homeowner's appeal.
[Buffer width] of screening along [boundary], building height capped at [X] feet, and a [berm or tree line] on the [direction] side. Visual simulations from [X] vantage points are at [link]. If the one from your road is not there, tell us where to stand and we will produce it.
No measurable search volume, heavy presence in tax and hearing records. Note the falsifier: “do data centers help schools” returned no data at all. The positive case has no organic demand and has to be carried by comms.
[Project] pays [personal property] tax on equipment and buildings, which in [county] flows [X]% to schools, about [$X] a year at full build under current rates. Separately, [state]'s sales tax exemption on equipment is a state-level break. Different tax, different pocket, and the two get conflated constantly.
The abatement is on [tax] and it is disclosed in [county]'s annual financial report under tax abatement disclosures. It is public, and worth reading. What [Project] adds to the local base is [$X]; what is abated is [$Y]. We would rather you saw both numbers than took either side's version.
[X] permanent positions, plus [Y] during construction. A facility this size typically employs around 50 people full-time. That is the honest number and it is smaller than most economic development announcements imply. The case for the project is the tax base, not the headcount, and we should be judged on that.
[Project] is [X] feet from [school]. The measurable impacts at that distance are construction noise and traffic on [road]. We have committed to no construction traffic during arrival and dismissal, and [X] dBA at the school boundary. [Contact] is available to present to the school board.
It can, and it is solvable if addressed before the first truck. [Project]'s haul route is [road], approved by [agency], and we have agreed to no truck movements between [X] and [Y] on school days. If [district] needs a route change, [contact] can commit to it.
Barely searched, heavily documented. The strongest single fact in this category: standard traffic studies model operations, while residents experience construction.
Peak is roughly [X] a day for about [Y] months, concentrated in [phase]. This is the part of the project people actually feel, and standard traffic studies do not model it. They measure operations, when the site will have [Z] employees and almost no truck traffic. The construction number is the one to hold us to.
[Route], approved by [agency], avoiding [residential road] and [school]. Drivers are contractually required to use it and [contractor] pays [penalty] for violations. If you see a truck off route, text [number], with a photo and a time we can identify the vehicle.
We do, under a road use agreement with [county] that includes a [$X] bond and a pre-construction pavement survey so there is a baseline to measure damage against. [Link]. Without that baseline the question cannot be settled later, which is why we did one before the first truck moved.
Construction traffic near schools is where this has gone badly elsewhere: on one Louisiana haul route, daily volume nearly tripled and crashes rose sharply after construction began. We committed to no movements during arrival and dismissal, flaggers at [X], and a speed reduction. If it is not working, [contact] can change the route.
[X] months for [phase] and [Y] years to full build. We will publish a monthly construction schedule at [link] so you know what is coming rather than finding out at the end of your driveway. When construction ends, daily traffic drops to about [Z] vehicles. Operations are quiet.
Small but the fastest-accelerating term in the set on a three-month basis, and the subject of bills in ten states. Ranked by public-record frequency.
Land is usually assembled through a project entity, which is standard practice and also opaque. [Project LLC] is [Parent Company]. Now you know. Any developer can tell you who they are once land is under contract, and if they will not at that point, that is worth asking about.
[Yes / No]. It covered [scope] during [period] and has [expired]. It is at [link]. NDAs are common in economic development and they are also why people do not trust these projects. Our view is that confidentiality should end when the application is filed, and ours [did on date].
[Project X]. Code names exist to keep land prices from spiking during assembly, which is a real commercial reason and a terrible look. We would rather tell you the name than have you find it in a records request.
[Link], or request it from [county clerk] under [state]'s public records law: [process]. Everything we filed with [county] is public and we are not asserting an exemption on [list]. If something you asked for comes back redacted, tell us and we will say why.
Fair question. Here is what we have not disclosed and why: [X] because [reason], and [Y], which we release at [milestone]. Here is what we are not required to disclose but will: [list]. If the answer to this question is ever “nothing,” nobody believes it, so here is the actual list.
The full Community Toolkit carries the complete question bank by project phase, response playbooks for when a question becomes an organized campaign, and language calibrated to your market rather than the national average. The Local Sentiment Index measures where you actually stand before the first hearing.
Score your market Work with us
Reuse is the point: take any answer here and put it on your project site. Attribution to The GridSight is appreciated, not required. These are model communications written from the developer's perspective and published as a demonstration of our community engagement work. They are not the Local Sentiment Index, which is scored to a fixed rubric independent of who subscribes.