Does a microgrid pay off, and for whom?
Price an outage, size a campus microgrid against it, then see who pays, who gains and who carries the risk
Price an outage, then size Prairie Campus's microgrid against it (2025 dollars)
| Length | A household | A business | Household, per kWh not delivered |
|---|---|---|---|
| 2 hours | $7.08 | $3,772 | $3.10 |
| 8 hours | $29.42 | $7,898 | $3.22 |
| 24 hours | $53.68 | $15,106 | $1.96 |
| Length | Outages a year | Hours a year | Cost a year |
|---|---|---|---|
| 1 h | 0.8 | 0.8 | $7,200 |
| 4 h | 0.4 | 1.6 | $14,400 |
| 12 h | 0.15 | 1.8 | $16,200 |
| 36 h | 0.06 | 2.16 | $19,440 |
| 72 h | 0.02 | 1.44 | $12,960 |
- The hospital's expected outage cost: $70,200 a year (illustrative: $9,000 an hour). Outages of 12 hours or more are 16% of the outages but 69% of the hours.
- A battery: its first hour saves $12,870 a year and costs $14,975, so no battery pays from outages alone.
- At $27,000 an hour, hours up to 4 pass, but 4 hours cost $97,340 against $89,640 avoided.
- A generator: the best tank holds 72 hours, +$14,219 a year.
| Energy for one hour of the hospital | 300 kWh |
|---|---|
| Battery at $400 per kWh | $120,000 |
| Spread over 15 years at 7% (× 0.1098) | $13,175 a year |
| Upkeep, 1.5% a year | $1,800 a year |
| One more hour of battery | $14,975 a year |
| Switch and controls | $150,000 |
|---|---|
| Battery inverter (with any battery) | $150,000 |
| Generator | $200,000, upkeep $6,000 a year |
| Fuel tank, per hour it holds | $120 a year |
| Fuel, per hour the generator runs | $85 |
| Design | Annual cost | Net, outages only | Net, with bill savings |
|---|---|---|---|
| Battery, 2 hours | $67,389 | −$48,849 | +$5,151 |
| Generator, 72 hours | $55,981 | +$14,219 | +$14,219 |
| Battery 2 hours and generator 72 hours | $104,476 | −$34,276 | +$19,724 |
Campus view: the bill saving is revenue the utility no longer collects. It saves the whole system money only if the smaller peak also cuts the utility's own costs.
| Case | Panels | Battery | Per kWh |
|---|---|---|---|
| Households only | 11 kW | 40 kWh | $0.31 |
| Households, mill at midday | 17 kW | 40 kWh | $0.23 |
| Households, mill in the evening | 17 kW | 70 kWh | $0.28 |
| Households, mill at midday and a cold store | 23 kW | 55 kWh | $0.20 |
A toy comparison: the same average day repeats all year; cloudy spells, seasons and battery power limits are left out.
With JavaScript on, you work through this step by step: three outages to price, a battery and a generator to size, three designs, four owners and a village.
Real, 2025 dollars: what outages cost rural co-op members in a 2025 survey (Berkeley Lab and partners) and EIA's 2025 retail prices.
Illustrative, in 2025 dollars: Prairie Campus, the hospital's cost per hour, the outage pattern, every equipment cost, the borrowing rates and the village. Costs are spread over 15 years at 7% unless an owner borrows at another rate.
What you just did, as diagrams
- Value of lost load
- What an hour or a kWh without power costs someone; it differs by customer and by outage length.
- Expected outage cost
- Outages a year × their length × the cost of an hour, summed over outage lengths: $70,200 a year for the hospital here.
- The next hour
- Buy another hour of backup while it saves more a year than it costs; then check that the whole design covers its fixed costs.
- Value stacking
- Everyday savings can carry a design that outage savings alone cannot, if one battery can do both: here it is assumed full when each outage starts.
- Who owns it
- At a fixed borrowing rate, changing who pays moves the gains between parties, not the combined total. Here each owner also borrows at its own rate, so the net changes too.
| Owner | Who pays | Carries the risk | Annual cost |
|---|---|---|---|
| The utility (7%) | All the utility's customers, through rates its state regulator approves | The customers, if the regulator lets the cost into rates; the utility's shareholders, if not | $104,476 |
| A rural co-op (5%) | The co-op's members, through their rates | The members: the co-op borrows, and has no shares to sell | $94,521 |
| The county (4%) | County taxpayers, from its budget | The taxpayers | $89,784 |
| A private company (10%) | The campus: a fixed yearly fee in a long contract | The company carries cost overruns and breakdowns; the campus pays the fee even in calm years | $120,518 |
Go deeper · AnalystSpreading a cost over the years
A sum paid now becomes a cost a year through the capital recovery factor, CRF:
\[\text{CRF} = \frac{r}{1-(1+r)^{-n}}\]
Go deeper · ModelerLetting a solver choose the sizes
- This lesson tests one size at a time. A sizing model chooses battery, generator and solar together, hour by hour, and can treat the hospital’s survival as a rule rather than a cost.
- A rule (“survive 72 hours”) and a cost (“an hour without power costs $9,000”) can give different designs: the second weighs the rare long outages by how rarely they come.
- Modeler lesson M5 sizes a microgrid that must survive an outage.
Check yourself, then practise
Three myths and one question, then a practice set of about 30 minutes.
Myth or fact? “A microgrid that keeps a hospital open always pays for itself.”
- Myth
- Fact
Answer
Myth. Here a 2-hour battery comes out at −$48,849 a year on outages alone, and at +$5,151 with the campus's bill savings, if it is full at every outage.
Myth or fact? “The value of lost load is one number.”
- Myth
- Fact
Answer
Myth. In the 2025 co-op survey, an 8-hour outage cost a household $3.22 and a business $41 per kWh not delivered.
Myth or fact? “If a microgrid saves money overall, its owner comes out ahead.”
- Myth
- Fact
Answer
Myth. The owner pays the capital and gains only through rates, a fee or the savings it keeps. Its borrowing rate moves the net too: +$3,682 a year at 10%, +$34,416 at 4%.
Outages of 4 hours hit the campus 0.4 times a year, and the hospital loses $9,000 an hour. What do they cost the hospital in a year, on average?
- $3,600
- $14,400
- $36,000
Answer
$14,400: 0.4 outages × 4 hours × $9,000 an hour.
Practice · about 30 minutes on your own
Use the numbers in the steps above (2025 dollars; the campus's are illustrative).
Expected outage hours. Outages of 1, 4, 12, 36 and 72 hours happen 0.8, 0.4, 0.15, 0.06 and 0.02 times a year. How many hours without power a year, on average?
Answer
7.8 hours: 0.8 × 1 + 0.4 × 4 + 0.15 × 12 + 0.06 × 36 + 0.02 × 72.
The fifth hour of battery. At $9,000 an hour, what does a battery's fifth hour save a year? Does it pay at $14,975?
Answer
Outages longer than 4 hours happen 0.23 times a year, so it saves $2,070 a year: far below $14,975.
When the first hour passes. At what cost of an hour without power would the battery's first hour save as much as it costs?
Answer
$14,975 ÷ 1.43 outages a year = $10,472 an hour. The fixed costs would still need covering.
Value against price. A household's 24-hour outage costs $1.96 per kWh not delivered. How many times the 2025 U.S. residential average of 17.23¢ per kWh?
Answer
About 11 times.
The cost of borrowing. The battery-and-generator design costs $740,000 to build, spread over 15 years. Its capital cost a year at 4% and at 10%?
Answer
$66,556 and $97,291: the same equipment, $30,734 a year apart.
The mill's hours. Why does the mill at midday lower the village's cost per kWh, while the mill in the evening helps less?
Answer
At midday the panels run the mill directly: $0.31 falls to $0.23. In the evening the battery must grow from 40 to 70 kWh: $0.28.
Numbers, sources and what’s next
Numbers to remember
| Number | What it means | Source |
|---|---|---|
| $29.42 · $7,898 | What an 8-hour outage costs, on average, a household and a business on a rural co-op (2025 survey). | Berkeley Lab and partners, ICE Calculator 2 co-op report, Tables 4-2 and 5-2 · November 2026 |
| $3.22 · 17.23¢ | A household's loss per kWh not delivered in an 8-hour outage, against the average U.S. residential price of a kWh in 2025. | Berkeley Lab report, Table 4-4; EIA, Electric Power Monthly, Table 5.3 · released 23 Oct 2026 |
| almost 900 · more than 1,700 | Electric cooperatives and municipal utilities in the United States. | FERC, 2024 Energy Primer, p. 36 · December 2023 |
| 19.5% | Share of the surveyed co-ops' business members with backup generation on site. | Berkeley Lab report, p. 59 · November 2026 |
Keep going
In the real world
In the 2025 survey of rural co-op members, 65% of businesses said a three-day outage would shut them down, and 16% would keep going on backup generation (Berkeley Lab and partners, p. 11).
Beyond outages, on-site energy can also change where a business chooses to build:
Because the transactive energy market brought a new revenue stream, locations with stronger renewables became competitive to establish facilities despite higher fixed and logistics cost.
Contrary to intuition, energy storage systems, which balance intermittent power and uncertain load, had limited benefits for cost saving.
- Holds when: stylized networks with synthetic distances and costs, candidate sites in California, Arizona and Nevada, and one year, planned week by week for products and hour by hour for energy.
- Leaves out: grid outages, which the studies do not model.
The full case: case file CF1.
Sources and further reading
- Berkeley Lab with NRECA and partners (2026). ICE Calculator 2 Initiative: Customer Interruption Costs for Select National Rural Electric Cooperative Association (NRECA) Members (PDF), Tables 4-2, 4-4, 5-2 and 5-4; pp. 11 and 59. November 2026.
- EIA. Electric Power Monthly, Table 5.3: Average Price of Electricity to Ultimate Customers. Released 23 October 2026.
- FERC (2023). 2024 Energy Primer: A Handbook of Energy Market Basics (PDF), pp. 35, 36 and 130. December 2023.
- U.S. DOE and NREL (2019). Microgrids (PDF), slides 4 and 6: what microgrids are built for, and why their resilience is hard to value. 11 December 2019.
- Li et al. (2023). Facility and microgrid location-allocation for integrated supply chain and transactive energy operations. Applied Mathematical Modelling 119, 119–136.
Spotted an error, or stuck? Write to hlli@ksu.edu. Corrections are made in place and dated here.
Text licensed CC BY 4.0, code MIT: teachers and students may reuse and adapt it, with credit to Honglin Li.