Agrivoltaics vs. Conventional Solar Farms: How Do Land Use and Costs Compare

Conventional solar normally fits more MW onto the same land and costs less per watt. NREL cites an average of about 5.75 acres/MWdc for utility-scale solar, while crop-oriented agrivoltaics often needs wider rows or taller structures. On 100 acres, a conventional layout using 6 acres/MWdc could fit about 16.7 MWdc, while an agrivoltaic layout using 8 acres/MWdc would fit about 12.5 MWdc. Agrivoltaics becomes more attractive when continued farming creates enough value to make up for the lower solar density.^[1]

That is more useful than asking which approach is simply “better.” Conventional PV uses the site mainly to produce electricity. Agrivoltaics intentionally gives up some light, space, or access so farming can continue.

A practical comparison comes down to four questions: how many MW fit, how much electricity they produce, how much the solar system costs, and how much agricultural value remains.

How the Land-Use Comparison Changes With the Question

NREL reports that utility-scale solar currently averages about 5.75 acres per MWdc, or roughly 0.17 MWdc per acre. It also notes that the real density changes with topography and technology.^[1]

For the examples below, that figure is rounded to 6 acres/MWdc for conventional PV.

Agrivoltaics does not have one standard land-density number. A sheep-grazing system may stay fairly close to normal solar spacing. A crop project built around tractors, harvesters, or high-clearance modules may need much more room.

To show the effect without pretending there is one universal design, use these planning cases:

The agrivoltaic numbers are example scenarios, not industry benchmarks. They simply represent progressively more space being kept open for farm access.

DOE describes the same basic tradeoff: wider solar rows allow more light to reach crops and make machinery access easier, but they reduce how much PV capacity fits on the land. Raising the panels can also improve access, but more steel and higher wind loads can push construction costs up.^[2]

Fixed-MW and Fixed-Acreage Projects Give Different Answers

First, compare projects that all need to deliver 20 MWdc.

The conventional case uses DOE's 2025 Q1 modeled market price for a representative 100 MWdc utility-scale PV-only system: $1.12/Wdc. DOE lists a minimum sustainable price of $1.07/Wdc for the same benchmark.^[3]

The $1.35/Wdc and $1.55/Wdc agrivoltaic figures are sensitivity assumptions. They represent premiums of about 21% and 38%:

($1.35 − $1.12) ÷ $1.12 ≈ 20.5%

($1.55 − $1.12) ÷ $1.12 ≈ 38.4%

At the same 20 MWdc capacity, the moderate agrivoltaic case needs 40 more acres and adds $4.6 million of assumed PV CAPEX. The wider design needs 80 more acres and adds $8.6 million.

Now keep the site fixed at 100 acres instead.

This produces a result that is easy to miss: agrivoltaics can cost more per watt but still have lower total PV CAPEX on a fixed parcel. The reason is simple. Fewer watts are installed.

In this example, the moderate agrivoltaic project costs about $1.8 million less in total solar CAPEX than the conventional 100-acre project, even though every installed watt is assumed to cost more.

So saying that agrivoltaics is simply “a more expensive solar farm” is too crude. The answer changes depending on whether MW or land area is being held constant.

How Much Electricity Is Lost When Solar Density Falls?

To keep MWdc and MWac capacity-factor definitions from getting mixed up, use specific yield instead.

For illustration, assume every layout produces 1,600 kWh per installed kWdc each year. This is not a site forecast. Real production depends on irradiation, temperature, tracker or fixed-tilt geometry, shading, inverter sizing, soiling, downtime, and curtailment.

Under these assumptions, the moderate layout generates about 6.7 GWh less electricity per year than the conventional layout.

The wide layout generates about 10.7 GWh less.

Better airflow, module temperature, or bifacial response can change yield per installed kW. But a small efficiency gain usually will not erase a large difference in installed MW.

When land is tight, electricity per acre often tells you more than module efficiency by itself.

What Does the Cost Difference Actually Mean?

DOE's $1.12/Wdc benchmark represents one utility-scale reference system. A real project's cost can move significantly with site preparation, foundations, labor, interconnection, equipment sourcing, permitting, and financing.^[3]

Agrivoltaics adds another set of site-specific requirements.

DOE specifically notes that raising PV panels can require additional steel and longer foundations because higher structures see greater wind loading.^[2]

That means a grazing-oriented design may carry only a small premium, while a high-clearance crop system built around large machinery may cost much more.

Operating cost also needs to be treated carefully. DOE's 2025 benchmark puts utility-scale PV O&M at about $20/kWdc-year, including cleaning, inspection, repair, component replacement, land lease, property tax, insurance, and management.^[3]

For a 20 MWdc project:

20,000 kW × $20 = $400,000/year

Do not just apply a random agrivoltaic percentage on top. Grazing may reduce mowing costs. Crop production may add coordination, cleaning, insurance, or equipment-protection costs.

IEA PVPS recommends full financial modeling for agrivoltaics, including CAPEX, O&M, land rent, electricity sales, agricultural revenue, financing, tax, curtailment, and farm-related operating limits. It also warns against making a final investment decision from a simple LCOE calculation alone.^[4]

Does Retained Farm Income Offset Lower Solar Output?

Go back to the 100-acre moderate agrivoltaic case.

At an illustrative electricity value of $50/MWh:

The $50/MWh value is only a sensitivity assumption, not a forecast power-purchase price.

The moderate agrivoltaic layout earns about $333,000 less electricity revenue per year in this example.

How productive would the farm need to be to make up that annual difference by itself?

If the agrivoltaic system retains 80% of the farm's original net agricultural contribution:

$333,000 ÷ 100 acres ÷ 80% ≈ $4,160/acre/year

So the original farm would need to generate roughly $4,160 per acre per year in net contribution for an 80% retained farm result to equal the annual electricity-revenue gap.

That is only a screening calculation.

The moderate agrivoltaic system also requires about $1.8 million less initial PV CAPEX on the fixed 100 acres because it installs only 12.5 MWdc instead of 16.7 MWdc.

A fair comparison cannot subtract the electricity loss, add crop revenue, and then ignore the difference in initial investment.

The useful comparison is lifetime cash flow:

Electricity revenue + retained agricultural net income − PV CAPEX − agricultural adaptation costs − O&M − financing − taxes − other project costs

Then calculate NPV or IRR using the same project life and discount rate.

Land Equivalent Ratio Measures Something Different From Financial Return

Conventional solar can produce more electricity per acre while agrivoltaics produces more combined output from the same land.

Land Equivalent Ratio, or LER, helps separate those two ideas.

In simplified form:

LER = relative solar output + relative agricultural output

In the 100-acre example, moderate agrivoltaics installs 12.5 MWdc versus 16.7 MWdc for conventional PV.

Assuming the same specific yield per installed kW:

Relative solar output = 12.5 ÷ 16.7 ≈ 0.75

If crop production remains at 80% of the open-field reference:

LER = 0.75 + 0.80 = 1.55

An LER of 1.55 means roughly 1.55 acres of separate reference land would be needed to reproduce the same combined solar and agricultural output generated on 1 acre of the agrivoltaic site, assuming the same reference yields.

That is a strong land-use result. It still does not tell you which project has the higher IRR.

A tonne of crops and a MWh of electricity have different prices, margins, and risks. Use LER for land productivity and financial modeling for investment performance.

Crop Yield, Water Use and Machinery Can Change the Economics

Do not assume crop production stays at 100% under the panels.

Shade changes available radiation, plant temperature, soil temperature, and evapotranspiration. Depending on the crop, climate, irrigation, and array geometry, the effect may be positive, neutral, or negative.

USDA Climate Hubs reports that a 2021 Oregon project found potatoes grown under panel shade produced an overall yield about 20% higher than potatoes in full sun. USDA also discusses possible improvements in water-use efficiency under some agrivoltaic conditions.^[5]

That result should not be copied directly into a model for wheat, corn, grapes, or a different climate.

Before local field data is available, test several crop scenarios:

Use the same approach for water.

If a 100-acre farm uses 2 acre-feet of irrigation water per acre each year:

100 × 2 = 200 acre-feet/year

If later measurements show a 15% reduction:

200 × 15% = 30 acre-feet/year

You can then value those 30 acre-feet using the site's real water and pumping costs. Until it is measured, the 15% remains only an assumption.

Machinery can have an even bigger effect on the solar design.

A tractor that is 2.5 m wide cannot realistically work in a 2.6 m gap. The operator still needs steering room, implement clearance, crop space, and safe distance from solar posts.

Sprayers and harvesters can need much more room than the tractor itself, and headlands need space for turning.

If those practical needs push row spacing from 6 m to 9 m, a large amount of PV capacity can disappear from the parcel.

That is why machinery width, operating height, turning radius, and headland dimensions should be known before the solar geometry is finalized.

PV Design Still Matters Once Agricultural Spacing Is Fixed

When farm access limits the number of module rows, higher module power density becomes more valuable because it can recover some of the lost DC capacity.

Tongwei's utility-scale module application range includes high-efficiency formats for large power stations. In an agrivoltaic project, however, module choice still has to work with tracker geometry, string length, row pitch, inverter design, and farm clearance. Wattage alone is not enough.

Bifacial modeling also needs extra care on farmland.

Dry soil, wet soil, young plants, mature crop canopy, and post-harvest ground all reflect light differently toward the rear of the module.

Tongwei's discussion of bifacial PV performance in agricultural environments explains how crop height, soil condition, and seasonal change can alter rear-side irradiance.

Using one optimistic rear-gain number for the entire year can therefore overstate production.

The agricultural use also changes the physical system. Tongwei's utility-scale portfolio includes a 300 MW Agriculture-Photovoltaic-Storage Integration project and a 100 MW Animal Husbandry & PV Integration project.

Both are dual-use applications, but their access and structural requirements are different. Livestock can often move around lower structures that would be impractical for crop machinery.

Land Competition and Interconnection Can Change the Answer

USDA Economic Research Service reviewed 3,177 large commercial solar projects built in rural U.S. areas from 2012 through 2020. More than 70% were on agricultural land: 43% on cropland and 28% on pasture-range land.^[6]

That helps explain why local land-use questions can become contentious even though renewable-energy infrastructure occupies only a small share of total U.S. farmland.

USDA separately reported that the combined rural footprint directly affected by utility-scale solar and wind was about 424,000 acres in 2020, compared with roughly 897 million acres of U.S. farmland—less than 0.05%.^[7]

That 424,000-acre figure covers solar and wind together. It should not be presented as solar-only land use.

For an individual project, the local situation matters much more than the national percentage.

A particular parcel may be valuable because it combines flat terrain, good solar resource, suitable soil, and affordable access to a substation. Moving elsewhere may completely change the grid economics.

Suppose the 100-acre parcel fits 16.7 MWdc conventionally and 12.5 MWdc with moderate agrivoltaic spacing.

If the grid connection can economically accept the full 16.7 MWdc, the missing 4.2 MWdc has real value.

If export is heavily constrained and the extra capacity cannot be used well, giving up some DC density for agriculture may cost much less than the nameplate difference suggests.

Keep interconnection, substation, and network-upgrade assumptions consistent when comparing the two layouts.

Which Numbers Should Be Compared Before Choosing?

A useful first comparison can fit in one project table.

DOE describes agrivoltaics as an active research area looking at farmer revenue, land productivity, ecosystems, and solar deployment costs.^[8]

That same approach makes sense for a commercial project. Seeing crops between solar rows does not automatically prove a land-use benefit. The agricultural activity needs measurable output.

When Does Each Approach Make More Sense?

Agrivoltaics does not have to beat conventional solar on every single metric.

It needs enough agricultural, land-use, or other value to make up for the areas where its solar performance or cost is weaker.

In the same way, a high LER does not make a project financially attractive if the retained agricultural output is worth very little and the lost solar density is expensive.

Conclusion

Conventional solar normally fits more MW per acre and costs less per watt. Using 6 acres/MWdc for conventional PV and 8 acres/MWdc for agrivoltaics, 100 acres would hold about 16.7 MWdc versus 12.5 MWdc. The agrivoltaic case needs retained farm income, land-use value, or other savings to compensate for lower solar output. Compare annual MWh, total CAPEX, agricultural net income, and NPV—not $/W by itself.