A case study using the WattShed engine · Wythe County, Virginia

Community Bill Reducer

Why this study. Fluidstack is building a 210 MW campus in Wythe County, and the county is split over it. A 2026 Virginia law lets a customer that size fund measured peak cuts in its neighbors' homes and count them; New York and New Jersey passed their own versions this summer. Every new campus will meet the same question, and the answer has to be worked out site by site, on the local tariff, the local homes and the local politics, before the backlash and before the utility files. This is that work for Wythe County.

Fluidstack should create the Community Bill Reducer: free upgrades and a battery in the homes nearest its campus, paid for by Fluidstack, counted as capacity reduction credits under Va. Code 56-596.7. Over 14 years the $10M commitment costs it $5.6M in the base case, $3.7M under the best credit rule, and $8.5M if no credit is ever granted.

A vote on 8 Dec would limit data centers to Progress Park, where the campus already is. On 27 Oct Fluidstack told the board it is "interested in discussing opportunities to support the broader community as the project develops." Appalachian Power files its program under the law by 15 Jan 2027.

Read the full paper, with appendices (PDF, 38 pages)

A program company Fluidstack owns pays for upgrades and a battery in the homes nearest the campus. Households pay nothing.

Open to every home within 5 miles (4,744 homes, closest first). A home gets each upgrade that costs less than $3,000 per verified kW in that home, decided on its own numbers, and a 13.5 kWh battery. Installers from Appalachian Power's contractor list work at fixed prices.

The average program home
Gets
A 13.5 kWh battery; a smart thermostat in 73% of homes (every home heated with electricity), attic insulation in 37%, a cold-climate heat pump in 24% (the homes on resistance heat).
Takes off the winter peak
4.3 kW (1.9 from the upgrades, 2.4 from the battery measured to zero)
Saves its household
$595 a year, its bill from $2,270 to $1,675
Costs Fluidstack
$15,557 installed, $11,957 after the battery tax credit, plus $150 a year of upkeep
  • Homes in Phase 1, per Census block
  • 138 kV line
  • 500 and 765 kV
Phase 1 within 3 miles (367 homes at 50% sign-up), Phase 2 out to 5 miles. The closest homes go first: they live with the construction, the substation and the traffic.

$10M costs Fluidstack between $3.7M and $8.5M, depending on what it buys and how the credit is counted.

In September Fluidstack, Cipher and Anthropic committed $10M to Colorado City's water system, in Mitchell County, Texas. It returns nothing on a bill. Here the same $10M buys capacity credits off the demand charge ($9.96 per kW-month, one for one) for 14 years, at 7%.

What Fluidstack spends $10M onHomesMW creditedOff the demand charge a yearOver 14 years, today's $Battery tax creditNet costEach household a yearAll households a year
Upgrades plus a 13.5 kWh batteryNo credit. Returns community support, and whatever the Commission attaches to the standard, which the law says may include faster interconnection.6420.00$0$0$2.3M$8.5M$595$562 upgrades, $33 battery$382K
Upgrades onlyEvery verified kW1,7324.44$532K$4.7Mnone$6.1M$769$769 upgrades, $0 battery$1.3M
Upgrades plus a 13.5 kWh batteryBattery measured to zero. The base case: the battery credited only for the home's own load.6422.76$331K$2.9M$2.3M$5.6M$595$562 upgrades, $33 battery$382K
Upgrades plus a 13.5 kWh batteryExports count6423.51$421K$3.7M$2.3M$5.1M$595$562 upgrades, $33 battery (a $103 export credit and a $49 make-whole from the program, less $120 of recharging on peak)$382K
Upgrades plus 27 kWhExports count4634.20$504K$4.4M$2.5M$4.2M$661$562 upgrades, $100 battery (a $297 export credit and a $138 make-whole from the program, less $335 of recharging on peak)$306K
Upgrades plus 40.5 kWhExports count3624.59$550K$4.8M$2.6M$3.7M$664$562 upgrades, $103 battery (a $492 export credit and a $206 make-whole from the program, less $595 of recharging on peak)$240K
27 kWh batteries onlyExports count. Batteries in every home, no upgrades.5554.00$480K$4.2M$3M$4M$90$0 upgrades, $90 battery (a $227 export credit and a $104 make-whole from the program, less $241 of recharging on peak)$50K

Net cost is the installed cost, less the 30% battery tax credit (subject to tax review), plus upkeep at $150 a home a year and the make-whole payment, less the credits, all over 14 years in today's dollars. In the export rows the program pays each household the make-whole, so its battery is worth at least what it would be measured to zero: no household does worse by letting the program export. Upgrades alone run out of homes before money: the electric-heated homes that sign up within 5 miles cost $8.4M, so that row spends $8.4M, not $10M.

  • $212 per kW-year is the credit at which the "Upgrades plus 40.5 kWh" row pays for itself (the demand charge is $120).
  • $7,600 installed for 27 kWh is the price at which batteries alone pay for themselves, against $18,000 at the bulk price assumed here.

Why homes, and not one battery at the substation

Fluidstack can buy the same kW for less. Homes cost two to three times as much per kW, and three things still make them the right choice: the statute's route is reductions from other retail customers, so a battery Fluidstack owns at the substation would count, if at all, as the campus's own resource; the households keep lower bills; and a fenced battery yard does not answer the county's question.

OptionPer kW of winter peak, net of the battery tax creditWhat else it delivers
Utility-scale four-hour battery at the substation (NREL, 2025)about $924Nothing on household bills; a new fenced site in the county
This program, 13.5 kWh measured to zero$2,786 ($3,092 with 14 years of upkeep)$382K a year off 642 households' bills
This program, 40.5 kWh exporting$1,607 ($1,852 with upkeep and make-whole)$240K a year off 362 households' bills
Curtailing the campus on winter morningsThe value of the compute lostNothing for the community

Three terms in the filing move the net cost most. Fluidstack can ask Appalachian Power for them.

  1. Exports count. A battery is credited for what it sends to the grid over the 3-hour window, and the household is credited for those kWh at the on-peak rate (14.306 cents). That is the difference between the measured-to-zero row and the export rows above.
  2. Every verified kW counts. Insulation and heat pumps lower the peak every winter day but cannot be called; Dominion has proposed counting only dispatchable capacity.
  3. The credit applies below the floor. The table assumes the campus runs at full capacity. Appalachian Power bills new large loads on at least 80% of contract capacity, used or not; below that, a kW saved next door counts only if the filing makes it a bill credit.

In comments on the filing, these are the asks. Beyond the bill, the measured record of verified kW from the homes nearest the campus is the community case for Lot 32, the parcel next door that Fluidstack leases with an option to buy. The program design carries over to New York; there the obligation to fund it would fall on the site owner.

A resident applies for their home; Fluidstack follows it to verified kW.

Two working apps, built alongside the analysis. Try them on a program you set up yourself.

wattshed.co/sponsor
Community Bill Reduceropen
Applications
Homes verified
Verified peak reliefCredits of $4,128 a year
Applications
WW. OrtizPhase 1 · 2.4 miReview
LL. EvansPhase 1 · 0.3 miInstalled
9:41
Phase 1 · 2.4 mi from the campusWhat your home would get, free
$2,193saved a year8.6 kWless at peak$0your cost
ApplySending
Heat pump in place of resistance heat13.5 kWh home battery, backup in an outageAttic insulation up to R-38Smart thermostat

Set the area, the cap, the battery and the credit terms; the wizard gives the money table for your choices and runs your program in the sponsor console. The homeowner app reads the heating nameplate from a photo with a vision model. Nothing either stores.

Design decisions

  • The credit is an ask, not a claimed saving. The law never mentions the campus's demand charge, so counting a verified kW against it is the first term to request in the filing.
  • Every home near the campus is eligible, and each home gets the upgrades that cost less than $3,000 per kW in it: a heat pump replaces resistance heat, while a working heat pump, even an old one, stays.
  • Batteries are sized by capacity (13.5, 27 or 40.5 kWh) and credited over the 3-hour peak window. Measured at the meter, a battery covers only the home's own load, so the base case credits that; exports are the filing's ask.
  • The household bill is computed hour by hour on Appalachian Power's four residential schedules, and the cheapest is chosen. Most homes are already cheapest on the smart demand schedule, so the battery saves $33 a year on the bill.
  • Exported kWh belong to the household, credited at the on-peak rate if the filing grants it. In any year exporting would leave a household with less than its battery saves measured to zero, the program pays the difference.
  • The program averages three kinds of home and four kinds of heating, from Census, building footprints and Street View: most homes near the campus are about 1,500 sq ft, and about one in five is a manufactured home.
  • Upkeep is part of the cost.

What this shows

  • A model with every input open, which anyone can rerun: the wizard above gives the money table for any area, cap, battery and credit terms.
  • Tariff 28, the SCC's large-load order and the statute read in full, and the three terms that move the result named.
  • County minutes, Census blocks, building footprints and a tariff turned into a one-page recommendation.
  • The number the decision turns on: the credit at which the program pays for itself, against the demand charge it offsets.
  • Four days in a market I had not worked in: Appalachian Power, PJM, Virginia.

Omer Bese, energy systems engineer (M.S., Columbia). Ten years in energy programs and startups, from LADWP's rebate program to WattShed. omer@wattshed.coLinkedIn

WattShed is my side project: the engines, the apps and the two case studies on it are mine, built to do studies like this one. This page sells nothing.

The full paper, with appendices (PDF)Google DocEvery assumption, the physics and one home hour by hourEarlier case study: the same $10M at the Barber Lake campus, Mitchell County, TexasThe engines (degree day and modified bin)How a job gets paid

Sources
  • Va. Code 56-596.7
  • SCC order PUR-2025-00057, Appalachian Power large-load terms (1 Jun 2026)
  • Appalachian Power Tariff 28: Schedule L.P.S.; residential R.S., R.S.-T.O.D., R.S.-S.T.O.U., R.S.-S.D.
  • Wythe County board minutes, 9 Jun, 28 Jul, 14 Sep, 28 Sep and 27 Oct 2026
  • PJM 2027/28 base residual auction results
  • Dominion Energy, Large Load Demand Flexibility working group materials (2026)
  • 2020 Census blocks; ACS 2019 to 2023 and 2020 to 2024; RECS 2020
  • Map: VGIN imagery and statewide parcels, lot numbers per VDOT (Oct 2025), lines and substations from OpenStreetMap
  • Battery and upgrade costs: EnergySage (2026), Tesla Powerwall 3 expansion pricing, NREL ATB 2025, Virginia contractor pricing
  • Weather: Marion-Wytheville Mountain Empire AP typical year (TMY3)
  • Upgrade kW and kWh: WattShed's engines, Wythe County (FIPS 51197)