buildmoredatacenters.com · feasibility assessment

Kootenai County, Idaho

PRIMARY UTILITY ANALYZED // Kootenai Electric Cooperative (member-owned co-op, BPA region, non-RTO)
SCENARIO // 25 MW primary · 100 MW scale-tested · RIGHT-SIZED MAX ON KEC // ~24 MW
PREPARED // July 25, 2026 · METHOD // public data + documented assumptions (shown in appendix)

County verdict: Conditional — bring-your-own-power

Kootenai County can host data centers, at the right sizes, on the right terms. The county's opportunity comes in two lanes: up to roughly 20–25 MW on Kootenai Electric Cooperative (headroom score 0.0/4 — every regional supply signal is tight), and mid-scale projects in Avista's service territory, which covers much of Coeur d'Alene and has roughly fifteen times KEC's system size. In both lanes the same condition applies, because both draw on the same constrained Pacific Northwest supply: the developer brings or funds its own generation and pays full marginal power costs. On that structure, the math is strongly positive for local ratepayers — a data center covering its own demand doesn't need to sell a single watt back to the grid; its fixed-cost payments simply spread the poles-and-wires costs that members currently carry alone. On a standard tariff, the same facility would raise bills. Structure decides everything.

1. The utilities serving this county

Utility	Type	Scale	Headroom	Right-sized max	Required structure
Kootenai Electric Cooperative	Co-op	33,486 meters · 588 GWh/yr · $61.6M revenue	0.0/4 (RED)	~24 MW	Bring-your-own-power + fixed-cost contribution; benefits flow to members automatically
Avista Corp.	Investor-owned	~420k electric customers regionwide (≈15× KEC)	Not separately scored — same constrained BPA region	Mid-scale possible w/ system study	BYO-power + a negotiated tariff mechanism; IOU margin does not flow to ratepayers automatically

The county is the unit of the verdict; the utility is the unit of the math. A rate impact can only be computed against a specific utility's sales, revenue, and tariff — there is no "county blended rate" to compute against. So this report scores each lane and then renders one county-level answer above. KEC is computed in full below because its cooperative structure makes it the cleanest demonstration of the mechanism (every dollar of margin returns to members automatically) and because its published cost data is unusually good. The Avista lane is assessed directionally — same regional supply constraint, same required structure — pending Avista-specific tariff and load data, which we would welcome.

1b. Why ~25 MW? The right-sizing logic

The 25 MW scenario is not arbitrary — it falls out of a sizing rule, stated here so it can be evaluated and challenged:

The rule: a distribution-served data center should not exceed roughly 30% of the host utility's total annual retail sales. Beyond that, the load stops being a large customer and starts being a second utility riding on the first one's system: it dominates power supply planning, requires transmission-level service, and concentrates default risk on the remaining members.
The arithmetic: KEC sells ~588 GWh/yr. 30% = ~176 GWh. A data center at 85% load factor consumes 8,760 × 0.85 = 7,446 MWh per MW per year. 176,000 ÷ 7,446 ≈ 23.7 MW. We rounded to a 25 MW analysis scenario.
Sanity checks: at 25 MW the load is 31.7% of KEC's sales — already at the edge, which is why the report's contribution estimates are capped and flagged. At 100 MW the load would be 127% of KEC's entire sales, which is why that scenario is marked non-analyzable on this system.
What would move the number: KEC's actual peak load, BPA Tier 1 allocation headroom, and substation capacity near candidate sites — all data the cooperative holds and this analysis does not. If KEC's system can comfortably serve more (or less), the right-size moves with it. This is the specific number we'd most like the cooperative to pressure-test.
The industry context: facilities are trending smaller and more distributed — edge (1–5 MW) and small-scale (5–25 MW) builds are a growing share of deployments. A ~20–25 MW recommendation is not a consolation prize; it's the size the market is increasingly building, matched to the system that would serve it.

2. Grid headroom — the four signals

Signal	Finding	Score
Sales trend	County population up >40% since 2010; member base growing steadily	0.0
Resource adequacy	BPA White Book projects regional energy deficits from 2027 (−370 MW) to −1,861 MW by 2030; E3 (Apr 2026) finds elevated shortage risk beginning 2026, especially winter; PNUCC 2026: resource development not keeping pace	0.0
Power cost trend	KEC's 2025 rate notice: BPA demand charges +24%, transmission +20%, attributed to scarcity; demand charges rising from 18% to ~25% of power costs within three years	0.0
Rescuable retirements	Region already short; new load worsens adequacy rather than preserving supply	0.0

The decisive fact: KEC buys the majority of its power from BPA at cost — among the cheapest in the nation — but the allocation is capped. New large load is served at the expensive margin, not the blended rate members pay. A RED score with a viable generation pathway (KEC joined the 17-co-op PNGC Power consortium in 2024 to build new resources) maps to CONDITIONAL, not NOT-HERE.

3. Rate impact on members (25 MW)

Structure	Rate change	Typical household
A — standard tariff (gap between marginal cost and blended rate socialized)	+4.2% to +14.8%	+$61 to +$213/yr
B — bring-your-own-power (developer self-supplies; pays fixed-cost contribution)	−7.1% to −14.2%	−$102 to −$205/yr

The swing between rows is the entire decision — same facility, opposite outcomes, determined by tariff design.

Caveats stated plainly. (1) The fixed-cost share is unusually well sourced: KEC publishes that power supply is "42 to 45¢ of every dollar" members pay, implying ~56% fixed. (2) The marginal power cost (7.5¢/kWh, range 6–9.5¢) is an ASSUMPTION and drives Structure A — this is the number we'd most like KEC to correct. (3) The Structure B result exceeds the 3–6% band in published research because a 25 MW load is a third of this co-op's volume, outside anything those studies measured; realistically the contribution would split across rate reduction, capital credits, and reserves. Treat the low end as the planning number.

100 MW scale test: 745 GWh/yr = 127% of KEC's total current sales — cannot be analyzed as a KEC distribution customer. At that size the project belongs on Avista or dedicated generation.

4. Water

The county sits atop the Spokane Valley–Rathdrum Prairie Aquifer, an EPA sole source aquifer. Precedent matters: in 2002, cogeneration plants sought ~18M gallons/day and most of those rights were denied after public opposition.

Cooling design (25 MW)	Gallons/day	Local equivalent
Traditional evaporative	~242,000	~800 homes
Hybrid evaporative	~47,000	~160 homes
Closed-loop / air-cooled	~2,700	~9 homes

North Idaho's cool, dry climate supports closed-loop and air cooling with minimal efficiency penalty — a genuine siting advantage. Recommendation: closed-loop or air cooling as a non-negotiable permit condition with annual public reporting.

5. Jobs and taxes (25 MW)

Construction: ~600 job-years — roughly 200 workers on site annually across three years. Without a cluster pipeline these do not become permanent; we say so rather than implying otherwise.
Permanent: ~9 direct FTE (hyperscale staffing) to ~15 (colocation); 18–30 with indirect effects. Published research finds naive multipliers overstate by ~3×, so use the low end publicly. The fiscal case, not headcount, is the local benefit.
Property tax: ~$1.4–2.4M/yr (midpoint $1.9M) against the county's $144M FY2026 budget — about 1.3%, flowing to county, cities, schools, fire, and highway districts. Meaningful in a county that warned in June 2026 that even a 3% levy increase might not avoid cuts — but this is a contribution, not Loudoun County's 38%.
Incentive trap: Idaho's sales-tax exemption requires $250M invested and 30 jobs at county-average wages. A 25 MW hyperscale facility clears the investment but yields ~9 direct jobs — it would take ~85 MW hyperscale or ~50 MW colocation to qualify. A developer promising a small footprint and full incentives should be asked to reconcile that.
Local lever: Idaho lets county commissioners grant up to 100% property tax exemption for up to five years at their discretion. If granted, the tax case above is zero for those years — the single biggest decision the county controls.

6. What a good deal looks like here

Adopt a large-load tariff before entertaining any proposal — negotiating after a developer is at the table is the weakest position.
Full marginal power-cost pass-through; the members' capped federal hydro is never blended into the data center's rate.
Developer-funded generation contracted through PNGC Power, so the scale benefits all member co-ops.
Closed-loop or air cooling in the permit, with annual public water reporting.
Fixed-cost contribution above proportional share via demand and facilities charges — the specific mechanism that lowers member rates.
Take-or-pay minimums, exit fees, and collateral; a 25 MW departure without protection strands costs on 33,000 members.
A deliberate commissioners' decision on the property-tax exemption — granting it fully largely erases the fiscal case during the construction-impact years.

7. What would change the answer

KEC's actual BPA Tier 1 headroom — the highest-value unknown in this analysis. If more allocation headroom exists than assumed, the verdict improves materially. We would welcome correction on this and on the marginal power cost assumption.
A site in Avista territory — different scale limits, different benefit mechanism; deserves its own run.
Idaho incentive reform — narrowing bills failed in 2025 and 2026 but sponsors have said they'll return.
A developer declining closed-loop cooling, or proposing 50+ MW on KEC without dedicated generation → NOT HERE.

Appendix — key inputs and arithmetic

Input	Value	Basis
KEC annual retail sales (S)	587,998 MWh	EIA-derived utility profile
KEC annual revenue (R)	$61.56M	EIA-derived utility profile
Blended rate	10.47¢/kWh (residential 12.68¢)	computed / profile
Fixed-cost share (F)	0.56	KEC: power supply = 42–45¢ of every $1
Marginal supply cost	7.5¢/kWh (6–9.5¢)	ASSUMPTION — please correct
Fixed-cost contribution factor (k)	0.6 (0.4–0.8)	ASSUMPTION
County FY2026 budget	$144M	county commission, Aug 2025

E_dc(25MW) = 25 × 8,760 × 0.85 = 186.2 GWh = 31.7% of S · right-size cap (30% of S) → ~24 MW
Structure A: (0.075 − 0.0461) × 186.2M ÷ 588M = +0.92¢/kWh = +8.8% (mid)
Structure B: c_f = 0.56 × 0.1047 × 0.6 = 3.52¢ → $6.55M/yr → −1.11¢/kWh = −10.6% (mid)

Principal sources: Kootenai Electric Cooperative published rate materials and news releases (kec.com) · EIA-derived KEC utility profile · BPA Pacific Northwest Loads & Resources Study · E3 Greater Northwest Resource Adequacy assessment (Apr 2026) · PNUCC 2026 Regional Forecast · Kootenai County budget records and assessor materials (kcgov.us) · Idaho Commerce data center exemption terms · EPA/Idaho DEQ sole source aquifer designation · BoiseDev and Idaho Conservation League legislative coverage (2025–26).

This assessment uses public data and documented assumptions to estimate outcomes. It is not a substitute for a utility system impact study, a formal fiscal impact analysis, or legal and engineering review. Verify figures with the serving utility and county assessor before citing them publicly. Where a figure could not be sourced, it is marked as an assumption rather than presented as fact. Prepared by buildmoredatacenters.com — corrections actively solicited, especially from Kootenai Electric Cooperative.