Solar Photovoltaic System Cost Benchmarks | Department of Energy

Nov 06, 2024

Each year, the U.S. Department of Energy (DOE) Solar Energy Technologies Office (SETO) and its national laboratory partners analyze cost data for U.S. solar photovoltaic (PV) systems to develop cost benchmarks. These benchmarks help measure progress towards goals for reducing solar electricity costs and guide SETO research and development programs. Read more below to find out how these cost benchmarks are modeled and download the data and cost modeling program.

Market analysts routinely monitor and report the average cost of PV systems and components, but more detail is needed to understand the impact of recent and future technology developments on cost. Consequently, benchmark systems in the utility-scale, commercial, and residential PV market sectors are evaluated each year. Each benchmark system is representative of what is currently being installed in the United States and is defined in sufficient detail to assess the impact of system size, module efficiency, overhead, and many other factors on cost.

Unlike most PV cost studies that report values solely in dollars per watt, SETO’s PV system cost benchmark reports values using intrinsic units for each component. For example, the cost of a mounting structure is given in dollars per square meter of modules supported by that structure. This measure is independent of how much power is produced by those modules, making it possible to assess the benefit of improving PV module efficiency (the structure’s cost per module area is divided by the module’s power output per module area to obtain the cost per watt for that structure when used with those modules). This approach is intended to allow any input parameter in the model to be varied by up to a factor of two (up or down) to assess its impact on cost.

All costs reported are represented two ways: Minimum Sustainable Price (MSP) and Modeled Market Price (MMP). MSP is the minimum price (with inflation adjustment) that a company can charge for its product or service in a balanced, competitive market and remain financially solvent for the long term, assuming that each of the company’s input costs also represent the MSP for that cost element. MMP is the actual price in the current market, which may differ from MSP as a result of temporary market distortions. MSP is the more useful metric for long-term planning, including R&D direction and predicting the future of the power grid. MMP is the more useful metric for short-term planning, including the impact of tax and trade policies.

Three DOE national laboratories—Lawrence Berkeley National Laboratory, National Renewable Energy Laboratory, and Sandia National Laboratories—collect cost data from PV industry stakeholders. Each stakeholder is contacted by only one lab to avoid overlap. The industry survey seeks to understand the cost structure for each stakeholder, including how their costs are affected by scale, overhead, and market distortions. Data collection focuses on transactions in the first quarter of the calendar year.

Each lab’s data is analyzed by that lab and submitted to SETO along with a weighting factor that represents the number of independent data sources utilized for each cost element.

In addition to the cost of installing each benchmark system, the cost for operation and maintenance is also analyzed. The total cost over the service life of the system is amortized to give a levelized cost per year.

In the PV System Cost Model (PVSCM), the owner’s overnight capital expense (cash cost) for an installed PV system is divided into eight categories, which are the same for the utility-scale, commercial, and residential PV market segments:

The first five categories are referred to as the hardware cost and the last three categories are referred to as the soft cost.

Each of the eight cost categories is divided into up to 12 cost elements. Each cost element is the sum of a fixed cost that is independent of size plus a variable cost that is proportional to size. The meaning of “size” depends on the category: the annual production rate of the manufacturing facility for Module, Inverter, and ESS; the rated capacity of the installed system for SBOS, EBOS, Fieldwork, and Officework; and the developer’s annual system installation capacity for Other. The variable cost is given in dollars per intrinsic unit, with the unit chosen that most directly scales with the size for that category.

The PVSCM system cost is the price paid by the system owner to the system developer. Any tax credit realized by the owner is excluded and must be considered separately. Tariffs paid on imported hardware are treated as temporary market distortions that increase MMP but not MSP. Subsidies for domestically produced hardware are also treated as temporary market distortions, which may decrease MMP but not MSP. Tariffs and subsidies are noted in the spreadsheet’s Comments column.

PVSCM is implemented using an Excel spreadsheet. It collects the cost elements for each category, then sums the categories to obtain the system cost, for both MSP and MMP. Unit conversion multipliers are listed on a separate sheet labeled Factors. An additional sheet is used to calculate the cost of operation and maintenance (O&M).

Download the PVSCM Excel Program and Cost Data

Figure 1 presents the UPV benchmark system cost components by cost category for both MSP and MMP, without ESS. These values represent weighted average figures based on the data collected by all three participating national laboratories. Details, including ESS and O&M, are available in the PVSCM model that can be downloaded above, using the UPV 2024Q1.txt parameter file.

Figure 1: Benchmark Utility-Scale PV System Costs (No ESS).

The representative utility-scale system (UPV) for 2024 has a rating of 100 MWdc (the sum of the system’s module ratings). Each module has an area (with frame) of 2.57 m2 and a rated power of 530 watts, corresponding to an efficiency of 20.6%. The bifacial modules were produced in Southeast Asia in a plant producing 1.5 GWdc per year, using crystalline silicon solar cells also produced in Southeast Asia. In 2024Q1, these modules were not subject to import tariffs.

About 189,000 of these modules are mounted on single-axis tracking structures that are assembled in the field and occupy a land area of 160 hectares. The torque tubes and fasteners are produced domestically, and their producers receive 45X tax credits that correspond to $4.60 per m2 of mounted modules, half of which is assumed to be passed along to the tracker installer in the form of reduced component pricing.

The dc cables are connected to 19 utility-scale central inverters, each rated at 4 MWac, giving the PV system a rated ac power output of 76 MWac, which corresponds to an inverter loading ratio of 1.32. The inverters are made in Europe in a plant that produces 250 of them each year. These inverters are not subject to import tariffs.

When supplied with an energy storage system (ESS), that ESS is comprised of 80 pad-mounted lithium-ion battery cabinets, each with an energy storage capacity of 3 MWh for a total of 240 MWh of storage. The ESS cabinet includes a bidirectional inverter rated at 750 kWac (4-hour discharge rate) for a total of 60 MWac. The ESS inverter is ac coupled with the PV inverter. The ESS system is assembled in the United States using domestic components except for the battery cells, which are imported from China and subject to 25% import tariff. The ESS producer receives a 45X tax credit of $10/kWh for battery modules. Half of this credit is assumed to be passed along to the project developer in the form of reduced ESS pricing.

O&M for the UPV system includes module cleaning, periodic inspection, replacement of components, lease on the land, property tax, insurance, and management. Because component replacements occur well into the future, the price of those replacements is based on the MSP of those components, rather than MMP.

Figure 2 presents the APV benchmark system cost components by cost category for both MSP and MMP, without ESS. These values represent weighted average figures based on the data collected by all three participating national laboratories. Details, including ESS and O&M, are available in the PVSCM model that can be downloaded above, using the APV 2024Q1.txt parameter file.

Figure 2: Benchmark Agrivoltaics System Costs (no ESS).

The representative commercial PV system for 2024 is an agrivoltaics system (APV) designed for land that is also used for grazing sheep. The system has a power rating of 3 MWdc (the sum of the system’s module ratings). Each module has an area (with frame) of 2.57 m2 and a rated power of 530 watts, corresponding to an efficiency of 20.6%. The bifacial modules were produced in Southeast Asia in a plant producing 1.5 GWdc per year, using crystalline silicon solar cells also produced in Southeast Asia. In 2024Q1, these modules were not subject to import tariffs.

About 5,700 of these modules are mounted on a south-facing latitude-tilt structure that is assembled in the field and occupies a land area of 5 hectares. The module rails and fasteners are imported from China and subject to 25% tariff.

The dc conductors are connected to 220 three-phase string inverters, each rated at 10 kWac, giving the PV system a rated ac power output of 2.2 MWac, which corresponds to an inverter loading ratio of 1.37. The inverters are made in China in a plant that produces 100,000 of them each year and are subject to 25% import tariff.

When supplied with an energy storage system (ESS), that ESS is comprised of 2 pad-mounted lithium-ion battery cabinets, each with an energy storage capacity of 3 MWh for a total of 6 MWh of storage. The ESS cabinet includes a bidirectional inverter rated at 750 kWac (4-hour discharge rate) for a total of 1.5 MWac. The ESS inverter is ac coupled with the PV inverter. The ESS system is assembled in the United States using domestic components except for the battery cells, which are imported from China and subject to 25% import tariff. The ESS producer receives a 45X tax credit of $10/kWh for battery modules. Half of this credit is assumed to be passed along to the project developer in the form of reduced ESS pricing.

O&M for the APV system includes module cleaning, periodic inspection, replacement of components, lease on the land, property tax, insurance, and management. Both the cost of the sheep and income from selling the sheep are considered out of scope and excluded from this model. Because component replacements occur well into the future, the price of those replacements is based on the MSP of those components, rather than MMP.

Figure 3 presents the RPV benchmark system cost components by cost category for both MSP and MMP, without ESS. These values represent weighted average figures based on the data collected by all three participating national laboratories. Details, including ESS and O&M, are available in the PVSCM model that can be downloaded above, using the RPV 2024Q1.txt parameter file.

Figure 3: Benchmark Residential System Costs (no ESS).

The representative residential PV system (RPV) for 2024 has a rating of 8 kWdc (the sum of the system’s module ratings). Each module has an area (with frame) of 1.9 m2 and a rated power of 400 watts, corresponding to an efficiency of 21.1%. The monofacial modules were assembled in the United States in a plant producing 1.5 GWdc per year, using n-type crystalline silicon solar cells produced in Southeast Asia. In 2024Q1, these cells were not subject to import tariffs. All other module components are imported from China, and only the glass is excluded from 25% import tariff. The module producer benefits from a 45X tax credit of $0.07/Wdc, half of which is assumed to be passed along to the system installer in the form of a reduced module cost.

Twenty of these modules are mounted on a fixed south-facing rooftop using domestic roof mounts. The aluminum rails and module clamps are imported from China and subject to 25% tariff.

Each module is paired with a microinverter rated at 330 Wac, giving the PV system a rated ac power output of 6.6 kWac, which corresponds to an inverter loading ratio of 1.22. The inverters are made in China in a plant that produces 3 million of them each year. These are subject to 25% import tariff.

When supplied with an energy storage system (ESS), that ESS is a lithium-ion battery cabinet having an energy storage capacity of 13.5 kWh. The ESS cabinet includes a bidirectional inverter rated at 5 kWac. The ESS inverter is ac coupled with the PV inverter. The ESS system is assembled in the United States using domestic components except for the battery cells, which are imported from China and subject to 25% import tariff. The ESS producer receives a 45X tax credit of $10/kWh for battery modules. Half of this credit is assumed to be passed along to the system installer in the form of reduced ESS pricing.

O&M for the RPV system includes the same cost elements that apply to UPV, except there is no cost to lease the land. Cleaning and inspection are assumed to be performed by the homeowner and a cost associated with their time is included in the model, although most homeowners do not directly acknowledge this cost. The cost of increased property tax and insurance coverage are also included in the model, although these costs are often not assessed to the property owner. Because component replacements occur well into the future, the price of those replacements is based on the MSP of those components, rather than MMP.

SETO tracks progress in the levelized cost of electricity (LCOE) towards long-term objectives and uses MSP rather than MMP for this purpose. Calculating LCOE for solar power requires four main inputs: system capital cost, system operating cost, solar resource, and a financial model. PVSCM provides the first two inputs for each benchmark system. For the remaining two inputs, SETO relies on the most recent release of NREL’s Annual Technology Baseline (ATB), using its midrange (class 5) solar resource and its long-term (R&D) financial model (30-year cost recovery). The ATB uses cost per ac watt for UPV, so the multiplier used in the ATB (1.34) is applied to the cost per dc watt when inserting UPV costs into the ATB. For PV with energy storage, the LCOE is increased by an additional 6% to account for energy losses in the storage system. Note that the ATB itself uses MMP values for calculating the current-year LCOE, whereas SETO’s cost benchmark uses MSP values.

The following tables summarize this year’s cost benchmarks and resulting LCOE values, for PV-only systems and for PV+ESS. All dollar values are inflation-adjusted to 2023 U.S. dollars (CPI-U=304.7).

These benchmark LCOE values do not reflect any system-level subsidies, which reduce the effective LCOE in proportion to the subsidy percentage. LCOE is lower than the value listed in these tables in locations with more annual sunshine (30% less in the desert southwest), and higher in regions with less annual sunshine (30% more in the Pacific northwest).

Figure 4 illustrates this year’s benchmark LCOE values for both PV and PV+ESS. For comparison, the corresponding LCOE value for each type of system in 2020 and 2023 are shown. Note the benchmark commercial system in 2023 was a community-solar system, whereas in 2024 it was an agrivoltaics system.

Figure 4. Levelized cost of electricity (LCOE) for representative systems installed in an average U.S. climate without deductions for federal, state, or local incentives. All costs have been inflation-adjusted to 2023 U.S. dollars.

DOE: Paul Basore (technology manager), Krysta Dummit, Christie EllisLawrence Berkeley National Laboratory: Peng Peng (principal investigator), Peter Benoliel, Thomas HendricksonNational Renewable Energy Laboratory: David Feldman (principal investigator), Vignesh Ramasamy, Michael Woodhouse, Jarett Zuboy, Jal Desai, Andy WalkerSandia National Laboratories: Jennifer Braid (principal investigator), Norman Jost, Emma Cooper