How to read this case study. The engagement summary reflects the published record. Client names, actual costs, loan terms and proprietary data are confidential, so the project budget, market sizing and financial projections below are a modeled example built from public data for this project type and location. They show the analysis a lender-grade study performs, not the client's figures.
Data current as of November 2026.
Engagement summary
This engagement assessed a roof-mounted solar array on a cold-storage operation in Tifton, Georgia. It was prepared in 2023 for a CDFI lender, and its defining feature was right-sizing: the array was documented against the facility's own consumption. The rules have since changed. USDA rescinded the FY2025 to FY2027 REAP funding notice in April 2026 [2], and a new REAP rule took effect on 16 October 2026 [1]. This page describes the engagement type and models a south Georgia cold store and its rooftop array. It then shows how such a project is financed now and what a study must prove under the rule in force.
| Item | Published record |
|---|---|
| Program | USDA Rural Energy for America Program (REAP) |
| Asset class | Cold storage |
| Project | Roof-mount solar on a cold-storage facility |
| Location | Tifton, Georgia (Tift County) |
| Region | Southeast |
| Loan size tier | Under $5M |
| Year prepared | 2023 |
| Lender type | CDFI lender |
| Outcome | Prepared with right-sized system documentation |
Sources: the site's engagement index [12]; county from Census place files [6].
The record describes the method:
- sizing with NREL's PVWatts approach at about 5.2 peak sun-hours a day for south and central Georgia;
- array size held to the facility's historical consumption;
- battery storage screened;
- rural-area eligibility checked against the USDA map;
- the operator's small-business status checked under SBA size standards (13 CFR 121).
The record also lists the grant cap and cost-share terms of the program at the time. Those terms were superseded by the rescission and the 2026 rule, and none of them is used here.
The question the lender and USDA needed answered
A cold store is close to an ideal solar host on paper. Refrigeration runs day and night, summer loads peak when the sun is strongest, and the roof is large, flat and unshaded. The study's job is to turn that intuition into numbers a credit officer can sign, and to answer four questions.
- How big should the array be? Size it to the roof, and to the share of output the building will use itself rather than export at a low credit. The record's "right-sized" outcome is this question answered.
- What is a kilowatt-hour worth to this facility? On a commercial demand tariff, solar mostly avoids energy charges. It trims demand charges only when the array is producing at the moment of the monthly peak. A study that values every kilowatt-hour at the average bill rate overstates savings.
- Can the roof carry it? Insulated cold-storage roofs, attachment methods and remaining roof life decide whether the project is buildable without damaging the envelope that keeps the product cold.
- Who repays the loan? As with most solar on an operating business, the savings alone will not cover the debt service. The warehouse's cash flow does.
Under 7 CFR part 5001, a REAP guaranteed loan for a renewable energy system needs a technical report, and a feasibility study when the lender or the Agency deems one necessary [3]. Since 16 October 2026, a REAP grant for a renewable system needs a resource assessment, or a feasibility study in its place, and projects of $200,000 or more need two years of pro forma statements and an engineer's completion certificate [1]. The new rule also makes the payback calculation an eligibility test: a project whose simple payback exceeds its useful life cannot receive a grant [1].
What changed for this engagement type on 16 October 2026
| Topic | Rule effective 16 October 2026 [1] | What it means for cold-storage solar |
|---|---|---|
| Timing | Project Period must end 12 to 24 months before the grant application (12 to 36 months in the first window) | The loan builds the array; a grant can follow a year of metered output |
| Data | 12 months of actual production plus 12 months of pre-installation data | Interval data before and after installation becomes the evidence |
| Award size | At most 25% of cost and $500,000 for a renewable system | On a $1.07 million project the 25% limit binds first |
| Combined funding | Grant plus REAP guaranteed loan at most 75% of project cost | A 75% loan leaves no room for a later grant |
| Loan alone | A REAP guaranteed loan without a grant is capped at 75% of eligible costs, with at least 25% from the borrower (7 CFR 5001.406(d)) | The operator funds a quarter even without a grant [3] |
| Cropland | Ground-mount solar or wind on Cropland is ineligible for a grant | The roof is the right place; a field array next door would not qualify |
| Sizing | Ground-mount systems must show commensurate historical energy use; storage limited to 120% of prior annual use | Roof arrays escape the clause, not the economics |
| Foreign components | Solar systems with any component from a foreign adversary under 15 CFR 791.4 are ineligible | Country-of-origin list for modules, inverters, racking |
| Exclusions | Stand-alone storage, storage retrofits, EV chargers and retail energy dispensing | Batteries only as part of the original system, if at all |
| Economics | Simple payback must not exceed useful life | The thin margin on cost and tariff becomes an eligibility risk |
| Gambling | Any gambling income bars a grant; the loan keeps the 15% test in 5001.127(b) | A revenue screen of the operator and affiliates |
Project description (modelled)
The modelled facility is an existing 60,000-square-foot cooler and freezer warehouse near Tifton, serving south Georgia's agricultural and food distribution trade. It uses a modelled 2,100,000 kWh a year, an average load of about 240 kW that never falls far because the compressors run around the clock. The roof can take about 488 kW of modules once setbacks, rooftop units and access paths are deducted.
The project is a 400 kW (DC) rooftop array on ballasted or low-penetration racking engineered for the insulated roof, with string inverters, a production meter and monitoring. The timeline runs: interval-data analysis and design in the first quarter, structural review and interconnection application in parallel, installation over about five months, then permission to operate from the utility. Battery storage was screened and left out of the base case, for reasons set out below.
Facility and solar inputs
| Facility input (model) | Value | How a real study sets it |
|---|---|---|
| Building | 60,000 sq ft cooler and freezer warehouse | Site plan and lease |
| Annual electricity use | 2,100,000 kWh (35 kWh per sq ft) | Twelve months or more of bills and interval data |
| Average load | 240 kW | Interval data; refrigeration runs around the clock |
| Roof area usable for modules | 65% of the roof | Roof survey: setbacks, units, skylights |
| Module density | 12.5 W per sq ft of usable roof | Layout drawing |
| Largest roof array | 488 kW DC | Computed |
| Peak sun-hours | 5.2 | Figure cited in the engagement record for south and central Georgia |
| Performance ratio | 0.77 | Assumption; heat losses on a hot roof |
| Year-1 output per kW DC | 1,461 kWh | Computed |
| Avoided energy value / export credit | $0.085 / $0.035 per kWh | Assumption; the tariff and export rider set them |
| Demand-charge reduction | 10% of AC capacity at $14 per kW-month | Assumption; interval data shows coincident peaks |
| Installed cost | $2.30 per watt DC plus $80,000 fixed, plus 5% contingency | Installer bids; range $1.90 to $3.10 |
| O&M and inverter reserve | $24 per kW a year | O&M contract; inverter warranty |
| Degradation / useful life | 0.5% a year / 25 years | Module warranty |
Right-sizing the array
| Array (DC) | Year-1 output (kWh) | Share of annual load | Assumed share used on site | Installed cost (model) | Year-1 value | Simple payback | Fits the roof? | Payback within 25 years? |
|---|---|---|---|---|---|---|---|---|
| 250 kW | 365,365 | 17% | 98% | $687,750 | $34,191 | 20.1 years | Yes | Yes |
| 400 kW | 584,584 | 28% | 92% | $1,050,000 | $52,951 | 19.8 years | Yes | Yes |
| 480 kW | 701,501 | 33% | 87% | $1,243,200 | $61,788 | 20.1 years | Yes | Yes |
| 600 kW | 876,876 | 42% | 78% | $1,639,312 | $73,289 | 22.4 years | No: needs a carport or ground array | Yes |
Payback is almost flat between 250 kW and the full roof, so size is decided by fit, not by marginal economics. The 400 kW array produces 584,584 kWh in year one, 28% of the facility's use. The model assumes 92% of that output is used on site, with exports mainly on cool weekend middays when refrigeration load dips. Filling the roof at 480 kW adds output but pushes more of it into the export credit. Going to 600 kW means a carport structure, at a modelled premium of $0.90 a watt above roof cost, and lengthens payback to 22.4 years. A ground array on adjacent farmland would be ineligible for a grant if the land is Cropland [1]. "Right-sized" in the record means this trade-off was documented against the facility's own consumption, which is still the discipline a reviewer expects.
Monthly match of output and load
| Month | Facility use (kWh, model) | Solar output (kWh, model) | Solar as share of use |
|---|---|---|---|
| Jan | 151,200 | 35,075 | 23% |
| Feb | 142,800 | 39,752 | 28% |
| Mar | 159,600 | 51,443 | 32% |
| Apr | 168,000 | 57,289 | 34% |
| May | 184,800 | 61,381 | 33% |
| Jun | 197,400 | 59,628 | 30% |
| Jul | 205,800 | 60,212 | 29% |
| Aug | 205,800 | 57,874 | 28% |
| Sep | 189,000 | 51,443 | 27% |
| Oct | 172,200 | 46,767 | 27% |
| Nov | 159,600 | 34,490 | 22% |
| Dec | 163,800 | 29,229 | 18% |
The monthly profile shows why cold storage suits solar. Facility use peaks in July and August, when output is high, and solar covers a fairly steady 27% to 34% of use from February to October. Winter months fall to 18% to 23%. A real study replaces these shapes with 15-minute interval data and an hourly production model, then counts exports hour by hour rather than assuming a share.
Project cost and sources and uses (modelled)
No construction-cost database sits behind these figures. Each line is a labelled assumption with a range, and a real study replaces it with installer bids, the structural engineer's letter and the utility's interconnection estimate.
| Line item | Model amount | Range | Source in a real study |
|---|---|---|---|
| Modules, inverters, racking and installation (400 kW DC at $2.30 per watt) | $920,000 | $760,000 to $1,240,000 | Installer bids |
| Roof structural assessment, attachment design and any reinforcement | $35,000 | $15,000 to $120,000 | Structural engineer |
| Interconnection application, utility study and metering | $25,000 | $10,000 to $60,000 | Utility interconnection agreement |
| Design engineering, production model and permits | $15,000 | $8,000 to $25,000 | Engineer proposal |
| Monitoring system | $5,000 | $3,000 to $10,000 | Vendor quote |
| Contingency (5% of hard cost) | $50,000 | 5% to 10% | Lender and engineer judgment |
| Feasibility study (within the cost page band for REAP loans above $600,000) | $9,500 | $7,500 to $11,500 | Fixed-fee proposal |
| Loan closing and legal | $7,500 | $5,000 to $12,000 | Lender estimate |
| Total eligible project cost (model) | $1,067,000 |
| Cost outside the eligible budget | Model amount | Note |
|---|---|---|
| REAP guarantee fee, 1.0% of the guaranteed portion (Structure A) | $6,402 | FY2026 REAP rate [4] |
| Extra 0.50% if the guarantee issues before construction is complete | $3,201 | Only if the lender asks for it [4] |
| Interest during a five-month installation (half drawn on average) | $11,670 | From operating cash |
| Annual retention fee, year 1 (0.25% of guaranteed principal) | $1,600 | Usually passed to the borrower [4] |
| Roof replacement, if the roof has less life left than the array | Not modelled | A roof is not a REAP cost; replace first or move the array |
The engagement record cites Georgia commercial solar installation costs of $3.06 to $3.52 a watt in 2025. We have not verified that range, and it likely reflects smaller systems. The model uses $2.30 for a 400 kW roof array and tests the record's range in the sensitivity tables, where it decides the grant question.
Two ways to finance it
| Uses | Amount | Structure A: REAP loan only | Amount | Structure B: smaller loan, later grant | Amount |
|---|---|---|---|---|---|
| Solar installation and site costs | $1,050,000 | REAP guaranteed loan from the CDFI (75%) | $800,200 | REAP guaranteed loan | $533,500 |
| Feasibility study | $9,500 | Owner contribution (25%) | $266,800 | Owner cash at construction | $533,500 |
| Closing and legal | $7,500 | REAP grant | $0 | REAP grant applied for 12 to 24 months after completion | $266,700 |
| Owner cash returned if the grant is awarded | -$266,700 | ||||
| Total eligible project cost | $1,067,000 | Total | $1,067,000 | Total net of grant, if awarded | $1,067,000 |
| Grant test (rule effective 16 October 2026) | Model | Source |
|---|---|---|
| 25% of project cost | $266,750 | [1] |
| Renewable system maximum | $500,000 | [1] |
| Largest possible grant | $266,750 | Lower of the two |
| Grant plus loan cap (75% of project cost) | $800,250 | [1] |
| Grant room left if the loan is at 75% (Structure A) | $0 | [1] |
| Structure B: loan and grant | $533,500 + $266,700 = 75.0% | [1][3] |
| Earliest application | 12 to 24 months after the Project Period ends, with 12 months of production data | [1] |
| Competition | One national competition for requests over $20,000, in an annual window | [1] |
Structure A borrows the full 75% and plans no grant. Structure B borrows 50% and asks the operator for $533,500 at construction. It returns $266,700 only if a national, competitive grant is awarded at least a year after completion. The new rule runs one national competition for requests over $20,000 [1], and as of 7 November 2026 no funding notice under the rule had appeared. The CDFI should underwrite Structure B on its own debt service and test whether the operator can carry the bridge.
Market analysis
For a behind-the-meter array, the market is the facility's own demand for electricity and the tariff that prices it. The market section also confirms that the host business, the refrigerated warehouse, will be there to use the power for the life of the loan. That makes the cold-storage market part of the solar study.
The site and the rural test
| Measure | Value | Source |
|---|---|---|
| Tifton city population, April 2020 estimates base | 17,031 | Census Vintage 2025 subcounty estimates [6] |
| Tifton city population, July 2025 estimate | 17,595 | [6] |
| Tift County population, July 2025 estimate | 42,254 | Census Vintage 2025 county estimates [6] |
| Tift County OMB status (2023) | Non-metropolitan | OMB list1_2023 [7] |
| Georgia counties that are non-metropolitan | 85 of 159 | [7] |
| Georgia places above 50,000 in the April 2020 base | 20 | [6] |
| Nearest of those to Tifton | Albany (69,653); Valdosta (55,393) | [6] |
| Tift County manufacturing establishments with payroll, 2023 | 38 (1,690 employees) | County Business Patterns 2023 [5] |
| Of which food manufacturing | 8 (106 employees) | [5] |
Tifton is far below the 50,000-person line in 7 CFR 5001.3. The nearest places above it in the Census file are Albany (69,653) and Valdosta (55,393), each a separate city with its own urbanized area. Tift County is one of Georgia's 85 non-metropolitan counties out of 159. The Georgia hub lists Tifton among the south Georgia towns that screen as eligible [13]. The rural test uses the latest decennial census, so the study cites the 2020 Census count and attaches a dated USDA eligibility map determination for the parcel. The estimates base above is a proxy for that count.
Cold storage in Georgia and the Southeast
Georgia: refrigerated warehousing, supply and rural footprint
| Measure | Georgia | United States |
|---|---|---|
| Establishments, 2019 → 2023 | 58 → 72 (+24.1%) | 1,350 → 1,367 |
| Per 100,000 residents (2023) | 0.7 | 0.4 |
| Average employees per establishment | 93.2 | 48.4 |
| Average annual pay per employee | $57,904 | $56,901 |
| Share of establishments in non-metro counties | 0.0% (0 in 0 counties) | |
| Share of state population in non-metro counties (2025) | 16.6% |
Sources: U.S. Census Bureau, County Business Patterns 2019 and 2023 (state and county files; establishments with paid employees; employment and payroll suppressed in some cells); Vintage 2025 county population estimates; OMB metropolitan statistical area delineation, July 2023. Non-metro county is a proxy for USDA rural eligibility, not the eligibility test itself.
| State | Refrigerated warehouses 2019 | 2023 | Change | Per 100,000 residents | Employees 2023 | Employees per site | Average pay |
|---|---|---|---|---|---|---|---|
| Georgia | 58 | 72 | 24.1% | 0.65 | 6,712 | 93.2 | $57,904 |
| Florida | 82 | 74 | -9.8% | 0.32 | 5,119 | 69.2 | $49,380 |
| North Carolina | 32 | 21 | -34.4% | 0.19 | 812 | 38.7 | $53,828 |
| Alabama | 21 | 20 | -4.8% | 0.39 | 604 | 30.2 | $55,118 |
| South Carolina | 16 | 13 | -18.8% | 0.24 | 333 | 25.6 | $65,186 |
County Business Patterns 2019 and 2023, NAICS 493120; per-resident figures use Census Vintage 2025 state estimates for July 2023 [5][6].
Georgia added 14 refrigerated warehouses between 2019 and 2023 while every neighbour in the table lost sites. Its warehouses are large, averaging 93.2 employees against 48.4 nationally. Yet the county file places none of them in a non-metro county. A cold store in Tift County is therefore a rural exception in a metro-centred industry. That cuts both ways for the study. Competition for the same customers is thin locally, but the operator's market depends on agricultural and distribution flows the study must document with customer contracts and throughput history. Tift County's 38 manufacturers, 8 of them food manufacturers employing 106 people, are part of that local demand. Georgia's average refrigerated-warehouse wage of $57,904 is the benchmark for the operator's labour costs.
REAP activity in Georgia and its neighbours
| State | REAP grants | REAP grant dollars | Average grant | Dollars per resident | Rank: dollars / per resident (of 53) | REAP guaranteed loans | B&I guaranteed loans |
|---|---|---|---|---|---|---|---|
| Georgia | 221 | $36,934,061 | $167,122 | $3.45 | 9 / 30 | 0 | 13 |
| Florida | 67 | $19,015,422 | $283,812 | $0.88 | 20 / 49 | 1 | 24 |
| Alabama | 61 | $14,847,758 | $243,406 | $2.96 | 26 / 33 | 0 | 9 |
| South Carolina | 51 | $16,936,060 | $332,080 | $3.31 | 22 / 31 | 7 | 10 |
| North Carolina | 102 | $21,671,880 | $212,469 | $2.08 | 15 / 37 | 12 | 33 |
USAspending prime awards by place of performance, FY2024 to FY2026, for REAP (10.868) and B&I (10.768); dollars per resident use the population figures in the USAspending data [9]. These awards were made under the pre-2026 rules.
Georgia took 221 REAP grants worth $36,934,061 in three years, ninth among 53 states and territories by dollars. It recorded no REAP guaranteed loans. South Carolina (7) and North Carolina (12) used the loan program; Georgia did not. Under the 2026 rule the guaranteed loan becomes the construction money for solar, so a Georgia operator and its lender may be less familiar with it than with the grant. The Georgia hub reports that a cold-storage operator in the state has used REAP for a solar array [13], so the project type has precedent.
Solar and cold-storage financing benchmarks
Georgia: solar and other renewable power generation, supply and rural footprint
| Measure | Georgia | United States |
|---|---|---|
| Establishments, 2019 → 2023 | 6 → 26 (+333.3%) | 405 → 942 |
| Per 100,000 residents (2023) | 0.2 | 0.3 |
| Average employees per establishment | 2.8 | 12.3 |
| Average annual pay per employee | $213,230 | $138,631 |
| Share of establishments in non-metro counties | 62.5% (5 in 1 counties) | |
| Share of state population in non-metro counties (2025) | 16.6% |
Non-metro counties with the most solar and other renewable power generation (2023): Jeff Davis County (5).
Sources: U.S. Census Bureau, County Business Patterns 2019 and 2023 (state and county files; establishments with paid employees; employment and payroll suppressed in some cells); Vintage 2025 county population estimates; OMB metropolitan statistical area delineation, July 2023. Non-metro county is a proxy for USDA rural eligibility, not the eligibility test itself.
Georgia: SBA lending to solar and other renewable power generation
| Period | 7(a) loans | 7(a) dollars | 504 loans | 504 dollars |
|---|---|---|---|---|
| FY2010 to FY2016 | 0 | $0 | 0 | $0 |
| FY2017 to FY2019 | 1 | $350,000 | 0 | $0 |
| FY2020 to FY2023 | 4 | $6.2 million | 0 | $0 |
| FY2024 to FY2026 | 0 | $0 | 0 | $0 |
Average 7(a) loan, FY2020 to FY2026: $1.6 million. Georgia had fewer than 30 seasoned 7(a) loans for solar and other renewable power generation; nationally, 12.9% of 31 FY2010 to FY2016 loans were charged off, on average 51 months after approval.
Source: SBA 7(a) and 504 FOIA loan data, as of 30 Sep 2026 (project state; disbursed loans only; cancelled and undisbursed loans excluded).
Georgia's generation establishments are project companies and developers, not hosts like this warehouse. The SBA record is thin, and nationally 12.9% of seasoned solar 7(a) loans were charged off. The cold-storage pack records no SBA 7(a) or 504 loans to Georgia refrigerated warehouses from FY2010 to FY2026 [8]. The lesson for the CDFI is that neither benchmark describes a rooftop array on an operating cold store. The credit rests on the warehouse, and the study should present it that way.
Tariff and price evidence the study must source
The model's avoided energy value ($0.085 a kWh), export credit ($0.035) and demand-charge reduction are labelled assumptions. The engagement record cites several Georgia electricity figures that we have not verified:
- commercial rates up 18% from 2020 to 2024, to 14.14 cents a kWh;
- a further increase of about 3.5% in January 2025;
- a Plant Vogtle Unit 4 surcharge;
- an export buyback of about 7.2 cents a kWh.
The average commercial price is not the value of a solar kWh to a large refrigerated load on a demand tariff, which is why the model uses a lower energy-only figure and tests the record's numbers as sensitivities. The Georgia hub notes that the state's electric membership corporations serve most of its land area [13]. Whether this parcel is served by an investor-owned utility or an EMC sets the tariff, the export terms and the interconnection process, and the study names it.
Financial feasibility
Five-year pro forma at project level
| Line (Structure A) | Year 1 | Year 2 | Year 3 | Year 4 | Year 5 |
|---|---|---|---|---|---|
| Energy charges avoided and export credits (tariff +2.5% a year, 0.5% degradation) | $47,351 | $48,252 | $49,170 | $50,107 | $51,063 |
| Demand-charge reduction | $5,600 | $5,740 | $5,884 | $6,031 | $6,181 |
| Total energy value | $52,951 | $53,992 | $55,054 | $56,138 | $57,244 |
| O&M and inverter reserve | -$9,600 | -$9,600 | -$9,600 | -$9,600 | -$9,600 |
| Insurance increment | -$3,150 | -$3,150 | -$3,150 | -$3,150 | -$3,150 |
| Property tax increment (placeholder) | -$2,500 | -$2,500 | -$2,500 | -$2,500 | -$2,500 |
| Net project benefit (project NOI) | $37,701 | $38,742 | $39,804 | $40,888 | $41,994 |
| REAP loan debt service (7.00%, 20 years) | -$74,447 | -$74,447 | -$74,447 | -$74,447 | -$74,447 |
| Retention fee | -$1,600 | -$1,562 | -$1,522 | -$1,478 | -$1,431 |
| Project-level coverage | 0.50x | 0.51x | 0.52x | 0.54x | 0.55x |
The array does not pay its own debt service on the model: 0.50 times coverage in year one, rising to 0.55 times by year five as tariffs escalate. That is the normal result for commercial rooftop solar financed over 20 years at 7.00% with no tax credit and no grant at construction. The study says so directly and moves the credit question to the warehouse.
Global coverage: the warehouse carries the loan
| Year | Warehouse cash flow before project (model) | Project NOI | Cash available | Existing debt service (model) | New REAP debt service and fee | Coverage before | Global coverage after |
|---|---|---|---|---|---|---|---|
| Year 1 | $1,100,000 | $37,701 | $1,137,701 | $720,000 | $76,048 | 1.53x | 1.43x |
| Year 2 | $1,100,000 | $38,742 | $1,138,742 | $720,000 | $76,010 | 1.53x | 1.43x |
| Year 3 | $1,100,000 | $39,804 | $1,139,804 | $720,000 | $75,969 | 1.53x | 1.43x |
| Year 4 | $1,100,000 | $40,888 | $1,140,888 | $720,000 | $75,925 | 1.53x | 1.43x |
| Year 5 | $1,100,000 | $41,994 | $1,141,994 | $720,000 | $75,878 | 1.53x | 1.43x |
The warehouse's cash flow and existing debt service are modelled placeholders, since they are the operator's confidential figures. The method follows Part 5001's global credit evaluation [3]. Adding the REAP loan takes modelled coverage from 1.53 times to 1.43 times, and the solar savings recover a little each year.
Term choice
| Term at 7.00% | Annual debt service | Project-level coverage, year 1 | Global coverage, year 1 | Useful-life check (5001.402) |
|---|---|---|---|---|
| 15 years | $86,309 | 0.43x | 1.41x | Within module life |
| 20 years | $74,447 | 0.50x | 1.43x | Within module life |
| 25 years | $67,868 | 0.54x | 1.44x | Equal to module life: no margin |
A 25-year term would match the module warranty exactly and leave nothing for a shorter-than-expected life. Twenty years keeps a five-year margin and costs only 0.01 times of global coverage against the longer term. Under 5001.402 the term must stay within useful life with no balloon [3], and inverters, which wear out sooner than modules, need a funded replacement reserve, which the pro forma carries.
Payback: the grant eligibility test
| Payback test | Value | Source |
|---|---|---|
| Total project cost used in the test (installed cost) | $1,050,000 | Budget; study and closing excluded as non-system costs (confirm) |
| Year-1 dollar value of energy produced | $52,951 | Energy, export and demand value |
| Simple payback | 19.8 years | Cost / annual value [3] |
| Including study and closing costs | 20.2 years | Conservative reading |
| Useful life assumed | 25 years | Module warranty |
| Passes the grant eligibility test? | Yes, with a thin margin | Payback must not exceed useful life [1] |
Part 5001 defines simple payback as total project cost divided by the annual dollar value of energy produced or saved, excluding one-time benefits and tax credits [3]. On the model the array passes with about five years to spare. The margin is the issue. The break-even and cost tables below show how little it takes to lose it.
Break-even and stress tests
| Break-even test (year 1) | Value | Model base | Margin |
|---|---|---|---|
| Avoided energy value for project-level 1.00x | $0.156 per kWh | $0.085 | 84% |
| Avoided energy value at which payback equals the 25-year life | $0.065 per kWh | $0.085 | -24% |
| Output at which payback equals 25 years | 77% of modelled output | 100% | -23% |
| Installed cost at which payback equals 25 years | $2.95 per watt | $2.30 | 28% |
| Warehouse cash flow for global 1.00x | $758,346 | $1,100,000 | -31% |
| Warehouse cash flow for global 1.25x | $957,358 | $1,100,000 | -13% |
Coverage break-even is out of reach at the project level: the avoided value would need to reach $0.156 a kWh. The payback test is closer to the edge in the other direction. A 24% lower avoided value, 23% less output or a 28% higher installed cost each push payback past 25 years and disqualify the array from a grant, though not from the loan.
| Installed cost per watt (400 kW) | Installed cost | Simple payback | Passes the grant payback test? | REAP loan at 75% | Owner share | Global coverage, year 1 |
|---|---|---|---|---|---|---|
| $1.90 | $882,000 | 16.7 years | Yes | $674,200 | $224,800 | 1.45x |
| $2.30 | $1,050,000 | 19.8 years | Yes | $800,200 | $266,800 | 1.43x |
| $2.70 | $1,218,000 | 23.0 years | Yes | $926,200 | $308,800 | 1.41x |
| $3.06 (record-cited 2025 range) | $1,369,200 | 25.9 years | No | $1,039,600 | $346,600 | 1.39x |
| $3.52 (record-cited 2025 range) | $1,562,400 | 29.5 years | No | $1,184,600 | $394,800 | 1.37x |
This table is the main finding of the modelled study. At the record's cited 2025 installed costs of $3.06 to $3.52 a watt, a 400 kW array would pay back in 25.9 to 29.5 years and fail the 2026 grant test. The lender's case would survive, with global coverage of 1.37 to 1.39 times, but the grant would be lost. Competitive installer bids are therefore an eligibility matter, not just a cost matter.
| Avoided energy value per kWh | Output 10% below model | Output as modelled | Output 5% above |
|---|---|---|---|
| $0.065 | 0.31x; payback 27.2 yrs | 0.35x; payback 24.9 yrs | 0.38x; payback 23.9 yrs |
| $0.075 | 0.37x; payback 24.2 yrs | 0.43x; payback 22.1 yrs | 0.45x; payback 21.1 yrs |
| $0.085 | 0.43x; payback 21.8 yrs | 0.50x; payback 19.8 yrs | 0.53x; payback 19.0 yrs |
| $0.100 | 0.53x; payback 18.9 yrs | 0.60x; payback 17.2 yrs | 0.64x; payback 16.5 yrs |
| $0.1414 (record-cited commercial average) | 0.79x; payback 13.9 yrs | 0.89x; payback 12.6 yrs | 0.95x; payback 12.0 yrs |
If the facility's true avoided value is nearer the record's average commercial price of 14.14 cents, payback falls to 12.6 years and project coverage rises to 0.89 times. The study cannot assume that. It shows the tariff arithmetic line by line from interval data.
| Export credit per kWh | 400 kW (base) | 480 kW (full roof) | 600 kW (roof plus carport) |
|---|---|---|---|
| $0.025 | $52,484; payback 20.0 yrs | $60,876; payback 20.4 yrs | $71,360; payback 23.0 yrs |
| $0.035 | $52,951; payback 19.8 yrs | $61,788; payback 20.1 yrs | $73,289; payback 22.4 yrs |
| $0.072 | $54,682; payback 19.2 yrs | $65,162; payback 19.1 yrs | $80,427; payback 20.4 yrs |
The export credit barely matters at 400 kW, because 92% of output is used on site. It matters more as the array grows, which is the economic reason to size to the load.
| Fixed rate (20 years) | Budget as modelled | Overrun 10% | Overrun 20% |
|---|---|---|---|
| 6.00% | $70,395; owner $266,800; global 1.44x | $77,442; owner $293,400; global 1.43x | $84,479; owner $320,100; global 1.41x |
| 7.00% | $76,048; owner $266,800; global 1.43x | $83,660; owner $293,400; global 1.42x | $91,263; owner $320,100; global 1.40x |
| 8.00% | $81,919; owner $266,800; global 1.42x | $90,119; owner $293,400; global 1.40x | $98,309; owner $320,100; global 1.39x |
| 9.00% | $87,996; owner $266,800; global 1.41x | $96,804; owner $293,400; global 1.39x | $105,601; owner $320,100; global 1.38x |
| Delay case | Base year 1 | Permission to operate six months late |
|---|---|---|
| Months of output in year 1 | 12 | 6 |
| Project NOI, year 1 | $37,701 | $18,851 |
| Debt service and fee | $76,048 | $76,048 |
| Global coverage | 1.43x | 1.41x |
| REAP grant clock | Starts at the end of the Project Period | Starts six months later; 12 months of data still needed |
A late permission to operate costs half a year of savings, but global coverage barely moves (1.41 times) because the warehouse carries the debt. The bigger cost of delay falls on Structure B. The grant clock starts only when the Project Period ends, and twelve months of data are still needed, so a late interconnection pushes any grant back by at least as long.
Equity and loan-size tests
| Test | Requirement | Model result (Structure A) | Source |
|---|---|---|---|
| REAP guaranteed loan size | At least $5,000 and at most 75% of eligible project costs | $800,200 = 75.0% of $1,067,000 | 7 CFR 5001.406(d) [3] |
| Borrower contribution | At least 25% of eligible project costs, with or without a grant | $266,800 = 25.0% | 5001.406(d) [3] |
| REAP loans to one borrower | At most $25 million | $800,200 | 5001.406(d)(1) [3] |
| Grant plus loan | At most 75% of Project Cost | Structure B: $800,200 = 75.0% | 4280.122(b)(20) [1] |
| Equity certification | Lender certifies at closing against a closing balance sheet | Pro forma closing balance sheet in the study | 5001.306(a)(2) [3] |
The operator's 25% can come from cash or from documented contributions to eligible cost. For a CDFI, the closing balance sheet and the global coverage table are the two documents that carry the credit memo.
Technical and environmental feasibility
- Roof. Cold-storage roofs are insulated assemblies over a vapour barrier, and a poorly detailed penetration can let in moisture that freezes inside the panel. The study needs a structural engineer's letter on dead load, wind uplift and the attachment method, and the roofer's confirmation that the roof warranty survives. It also needs the roof's remaining life. If the roof will need replacing within the array's life, replace it first, because removing and reinstalling an array costs more than the roof work.
- Electrical. Interconnection at the facility's service, protection settings, and the utility's study and upgrade costs. The production meter must support the 12 months of data a grant application needs [1].
- Equipment. Commercially available modules and inverters with a country-of-origin record for every component [1]. The installer must be licensed in Georgia with at least three similar installations, and procurement must follow 2 CFR 200.318 to 200.327 for a grant [1].
- Fire and access. Rooftop pathways and setbacks required by the fire code reduce usable roof area. The model's 65% usable-roof assumption stands in for the layout drawing.
- Storage. The record says battery storage was analysed. Under the 2026 rule, storage must be part of an eligible system and is limited to 120% of prior annual use. Stand-alone storage and storage retrofits are excluded [1]. A battery's value here would come from cutting demand charges and riding through outages, which matter to a freezer. The modelled base case leaves it out until interval data show demand peaks a battery could reliably shave.
- Environmental. A roof array on an existing building disturbs no ground. REAP grants carry environmental review requirements [1], and the study places them in the schedule.
Management and governance feasibility
The study confirms who owns the building and the roof, and whether the operator leases them. REAP requires site ownership or a lease of at least 12 months before the Project Period [1], and a lease must outlast the loan or include the array. It names who monitors production and acts on alarms, who holds the O&M contract, and who keeps the interval data a grant application will need. It screens the operator and affiliates under 5001.127: federal judgments, delinquent federal debt and debarment for the loan, and for a grant also the 2026 rule's bar on any gambling income [3][1]. The record notes that the operator was confirmed as a small business under SBA size standards (13 CFR 121), which the 2026 rule uses for rural small business applicants [1]. Finally, the study notes how a CDFI lender fits the REAP guarantee, since the lender files the application and evaluates the study and technical report in writing [3].
Economic and community impact (modelled)
| Measure (modelled) | Value | Basis |
|---|---|---|
| Electricity generated on site, year 1 | 584,584 kWh | 400 kW at 1,461 kWh per kW |
| Share of facility use | 28% | Of 2,100,000 kWh |
| Energy cost avoided, year 1 | $52,951 | Pro forma |
| Peak installation labour | Regional electrical and roofing crews for about five months | Schedule |
| Permanent jobs added | None claimed | The project protects an existing employer's cost base |
The community case is a lower, more stable energy cost for a rural employer whose product spoils without power. REAP's 2026 scoring rewards energy per award dollar, the share of energy replaced and environmental benefits [1], and the modelled output above is the evidence base for those points if a grant is sought later.
Risks and mitigants
| Risk | Effect | Mitigant the study documents |
|---|---|---|
| Installed cost at the record's cited 2025 range | Payback exceeds 25 years; grant lost | Competitive bids; cost sensitivity; size to the roof |
| Avoided value lower than modelled | Thinner payback margin | Tariff analysis from interval data; rate sensitivity |
| Roof failure or leak | Product loss, removal cost | Structural letter, roofer sign-off, roof life check |
| Interconnection delay or upgrade cost | Lost output, later grant clock | Early application; utility study in the budget |
| Foreign-adversary component | Grant ineligible | Country-of-origin list before purchase |
| Grant not awarded | Bridge not repaid | Structure A as base case |
| Warehouse throughput falls | Global coverage weakens | Customer contracts, throughput history, global sensitivity |
| Lease shorter than the loan | Array stranded | Lease extension or landlord consent before closing |
| Battery added later | Storage retrofit excluded from a grant | Decide on storage at design |
| Gambling income in an affiliate | Grant bar | Screen all entities |
What the five Part 5001 components showed
| Component [10] | What it tests here | What the modelled study shows |
|---|---|---|
| Economic | Resource, tariff, contracts and cost against benefit | $52,951 a year of value on a $1,050,000 system; payback 19.8 years |
| Market | Whether the energy has a use and the host will last | 92% of output used on site; a cold store in a non-metro county with thin local competition |
| Technical | Roof, equipment, installer, interconnection | 400 kW fits the usable roof; roof and interconnection letters required |
| Financial | Repayment source, equity, sensitivity | Project coverage 0.50x; global 1.43x; owner 25% |
| Management | Site control, O&M, monitoring, reporting | Lease and roof control documented; data kept for a later grant |
Why studies like this are returned
- Savings valued at the average bill rate, not the tariff's energy and demand components.
- An array sized to roof area or to an aspirational load, with exports valued at retail.
- No structural letter, or no plan for a roof with less life left than the array.
- Production at nameplate, or a generic regional yield with no heat losses.
- Debt service shown as covered by the solar savings, or by a grant not yet awarded.
- A loan above 75% of eligible cost, or a grant planned on top of a 75% loan.
- No country-of-origin record.
- Battery storage added after the fact, which the 2026 rule excludes from a grant.
Program terms that applied
| Term | REAP guaranteed loan (FY2026) | B&I guaranteed loan under $5 million (FY2026) | Source |
|---|---|---|---|
| Guarantee | 80% | 85% | 91 FR 11272 [4] |
| Upfront guarantee fee | 1.0% of the guaranteed portion | 3.0% | [4] |
| Annual retention fee | 0.25% of guaranteed principal outstanding | 0.55% | [4] |
| Extra fee if the guarantee issues before construction is complete | 0.50% | 0.50% | [4] |
| Loan size | $5,000 to 75% of eligible project costs; $25 million per borrower | Up to $25 million per borrower | 5001.406(c), (d) [3] |
| Borrower money | At least 25% of eligible project costs | 10% (existing business) to 25% (new business with construction) | 5001.406(d); 5001.105(d) [3] |
| Term | Useful life, at most 40 years, no balloon | Same | 5001.402 [3] |
| Rate | Negotiated; fixed, variable or both | Same | 5001.401 [3] |
| Rural test | Outside cities over 50,000 and their contiguous urbanized areas (latest decennial census) | Same | 5001.3 [3] |
| Gambling screen | More than 15% of gross revenue from gambling bars the loan | Same | 5001.127(b) [3] |
| Feasibility study | When the lender or Agency deems it necessary; technical report required | Required over $1,000,000 to a new business | 5001.306(a)(3), 5001.307 [3] |
These FY2026 rates apply to loans obligated in FY2026 unless a later appropriation changes them [4]. The 2026 CFR still lists solar and other renewable energy systems as B&I uses where REAP cannot finance them or REAP funds are unavailable (5001.105(b)(21)) [3]. A USDA press release of 19 August 2025 announced a policy on solar on farmland, so ask the Rural Development state office in Athens before planning a B&I guarantee for a solar project. A REAP guaranteed loan, which has no cropland restriction in part 5001 [3], is the natural route for a rooftop array.
Lessons for sponsors of similar projects
- Size to the load and the roof, not to ambition. On a cold store the economics are flat across sensible sizes, and oversizing only adds exports and structure.
- Price the kilowatt-hour from the tariff. Energy charges, demand charges and export credits each need their own line. The average bill rate overstates the value.
- Get bids before you plan a grant. At the record's cited 2025 installed costs, the modelled array fails the 25-year payback test. A competitive bid can be the difference between eligible and ineligible.
- Fix the roof first. A roof with less life left than the array is the most expensive mistake on a cold store.
- Plan for 25% of your own money. The REAP loan stops at 75% of eligible cost, and a grant cannot be stacked on a 75% loan.
- Underwrite without the grant. It arrives, if at all, a year or more after completion through a national competition.
- Decide on storage at design. A battery added later is a storage retrofit, which the 2026 rule excludes from a grant.
- Bring the warehouse's numbers. The lender's decision rests on the operator's global cash flow, so give the analyst financial statements, debt schedules and customer contracts at kickoff.
Related pages
- USDA REAP engagements: the program page for this engagement type.
- Renewable energy feasibility study for USDA loans and the USDA REAP feasibility study page.
- Cold storage feasibility study for USDA loans: the national page for this asset.
- Georgia USDA feasibility studies, with the Georgia cold storage feasibility study and Georgia solar energy feasibility study pages.
- Case studies and the full engagement index.
- USDA feasibility study cost and timeline.
The site's cost page lists REAP studies at $4,900 to $8,500 and 4 to 5 weeks for a loan under $600,000, and $7,500 to $11,500 and 5 to 7 weeks above it, each as a firm fixed fee in a written proposal [11]. For a cold-storage, packing or processing operator planning rooftop solar in Georgia or elsewhere, an independent USDA feasibility study consultant can build the sizing, tariff and credit case together.
Request a scoped proposal for your REAP solar feasibility study at the contact page.
USDA Feasibility, Inc.