TJE Ventures

The Capital Cost

Installed cost per watt for utility-scale solar has fallen roughly 90% since 2010. As of 2026:

• Utility-scale ground mount (>10 MW): $0.80–$1.20/W installed, all-in
• Community solar (1–5 MW): $1.20–$1.80/W
• Parking canopy structures: $2.50–$4.50/W

A 100 MW utility-scale farm, which requires 500–1,000 acres depending on panel density and terrain, costs $80M–$120M to build. The wide range reflects terrain, labor market, interconnection complexity, and whether the developer owns or leases the land.

What that cost buys:

• Monocrystalline or bifacial panels (roughly 35–40% of total installed cost)
• Single-axis tracking systems, which follow the sun east to west and add 15–25% to energy output vs. fixed-tilt at a cost premium of $0.10–$0.15/W
• Central or string inverters (10–15% of cost)
• Substation, interconnection, and transmission tie-in (15–25% of cost — highly variable by location)
• Site preparation, fencing, roads, and balance-of-system hardware

The single most variable line item — and the one that most frequently blows up project economics — is grid interconnection. A farm sited 2 miles from a substation with available capacity has a very different cost profile than one sited 40 miles from the nearest adequate transmission node. Interconnection queues in the U.S. have grown dramatically: as of 2025, over 2,600 GW of proposed capacity is waiting in interconnection queues nationally, and wait times of 3–6 years are common. Projects that can connect quickly and cheaply have a structural advantage over those that cannot.

The Revenue Model

Solar farm revenue comes from three potential sources, rarely all three at once:

Power Purchase Agreements (PPAs)

Most utility-scale solar is built against a long-term PPA — a contract with a utility, corporate buyer, or municipality to purchase all or most of the farm’s output at a fixed price for 15–25 years. PPA rates for utility solar signed in 2024–2026 ranged from $0.028–$0.055/kWh in high-irradiance markets (Southwest, Texas, Southeast) to $0.055–$0.080/kWh in lower-irradiance or higher-cost markets (Northeast, Pacific Northwest).

These rates reflect the dramatic cost reductions in panel manufacturing since 2010. They also reflect a brutally competitive developer market. A PPA at $0.035/kWh leaves very little margin for error on capital cost, transmission cost, or capacity factor assumptions.

Merchant Power Sales

Farms without a PPA sell power at spot market rates — a strategy that works when power prices are high and is punishing when they are low or negative. Spot prices in markets with high solar penetration (California, Texas midday) regularly go negative during peak solar hours as more supply hits the grid than demand can absorb. A merchant solar farm in California’s CAISO market during midday in spring can be paid nothing, or paid to shut off, when the grid is oversupplied.

This is the “duck curve” problem in economic terms: the more solar is built, the less the marginal unit of solar power is worth at the exact time solar farms produce the most of it. Developers building merchant farms today are betting that battery storage, demand flexibility, or curtailment management will improve this dynamic. That is a reasonable bet over 20 years, but it is a bet, not a certainty.

Renewable Energy Credits (RECs) and Incentives

Solar farms generate one Renewable Energy Credit (REC) per megawatt-hour of production. REC prices vary enormously by state based on Renewable Portfolio Standard requirements — from under $1/MWh in oversupplied markets to $20–$40/MWh in compliance-driven markets like Massachusetts and New Jersey. The Investment Tax Credit (ITC) at 30% (through 2032 under current law) significantly improves project-level returns for tax-equity investors. The Inflation Reduction Act’s bonus credits for domestic content and energy community siting can push the effective credit rate to 40–50% for qualifying projects.

Short-Term Returns: The First 10 Years

A well-structured utility-scale project in a favorable market produces these approximate financials:

• 100 MW farm in the Desert Southwest
• Capacity factor: 28% (2,453 MWh/MW/year)
• Annual generation: 245,000 MWh
• PPA rate: $0.038/kWh
• Annual gross revenue: ~$9.3M
• Annual O&M (operations & maintenance): $1.0–$1.5M
• Land lease: $500,000–$1.5M depending on acreage and location
• Net annual revenue before debt service: ~$6.5–$7.5M
• Unlevered project IRR (no debt): 7–10%
• Simple payback on equity: 10–14 years

The 7–10% unlevered IRR is competitive with investment-grade bonds and significantly better than Treasury yields in the current rate environment — but it is not the exceptional return the headlines sometimes suggest. The financial case for utility solar is solid but not spectacular; it works because the cash flows are predictable and long-duration, which suits institutional investors (pension funds, infrastructure funds) who need exactly that.

Leverage improves equity returns but also amplifies risk. A project financed at 60–70% debt with a 7% construction loan and 5% long-term debt can achieve equity IRRs of 12–16% — attractive to developers — but now the project is sensitive to interest rate changes, refinancing risk at year 7–10, and any operational underperformance that threatens debt coverage ratios.

Long-Term Returns: Year 10 to 30

The long-term value of a solar farm is far more uncertain than the short-term cash flows suggest. Several factors create real risk beyond year 15:

• PPA expiration. Most PPAs run 15–20 years. What happens at expiration is entirely unknown at project inception. The farm may re-contract at a favorable rate, sell merchant power, or find that the market has moved so far against it that the project is economically stranded. New solar built in 2035–2040 will likely be cheaper than the existing farm, putting downward pressure on re-contracting rates.
• Panel degradation. At 0.5%/year degradation, a farm producing 245,000 MWh in year 1 produces approximately 221,000 MWh in year 20 and 209,000 MWh in year 25. Revenue falls proportionally. Projections that assume flat output over 25 years are wrong.
• Inverter replacement. Central inverters in utility-scale applications last 10–15 years. A 100 MW farm replacing its inverter bank at year 12–15 faces a $3M–$8M capital event that is often modeled too conservatively or excluded entirely from project pro formas.
• Decommissioning cost. A 100 MW farm contains 300,000–400,000 panels, miles of wiring, concrete foundations, and a substation. Decommissioning costs are estimated at $50,000–$120,000 per MW, or $5M–$12M for a 100 MW farm. Most project financings require a decommissioning bond or reserve, but the required amounts are frequently set below actual cost. When projects reach end of life and the developer is long gone, the gap becomes a problem for whoever holds the land.

Taking panel degradation, one inverter replacement cycle, and realistic decommissioning into account, a 25-year project-level IRR of 6–8% is a more honest estimate for a typical utility-scale project than the headline figures sometimes cited. That is still a viable investment return for long-duration infrastructure — it just requires accurate modeling to get there.

Where You Put It Matters Enormously

The choice of land type affects capital cost, revenue potential, public perception, environmental impact, and long-term land-use tradeoffs in ways that are often collapsed into a single line item in project underwriting. They deserve more careful treatment.

Productive Farmland

A significant and growing share of utility solar in the United States is being built on productive agricultural land — particularly in the Midwest and Southeast, where flat terrain, existing road access, and proximity to transmission infrastructure make the siting economics attractive. In some states, solar lease payments to farmers ($500–$2,000/acre/year) represent substantially more income than crop production, especially in years with low commodity prices.

The economic case for the landowner is often real. The broader case for society is more complicated:

• Food production displacement. The U.S. has roughly 900 million acres of agricultural land. Current utility solar installations occupy a small fraction of that — approximately 3–4 million acres. But the pipeline of proposed solar projects, if fully built, would convert an additional 15–20 million acres over the next 15 years. That is 2–3% of total U.S. agricultural land removed from food production. The impact is diffuse and long-term, which is why it rarely generates urgent concern — but it is real and cumulative.
• Soil compaction and land condition. Solar panels on agricultural land do not simply sit on top of it. The grading, compaction from equipment, and removal of ground cover during construction changes soil structure. Whether the land returns to full agricultural productivity after 25 years depends heavily on how carefully it was managed during the lease — and that is rarely contractually guaranteed in a way that is enforceable after a developer has moved on.
• Lease vs. ownership dynamics. Most solar-on-farmland is structured as a 25–35 year lease, not a sale. The farmer retains ownership but gives up use. The development entity often assigns the lease to a tax-equity investor or project fund. By year 10, the entity the farmer signed with may no longer exist, and the counterparty managing the decommissioning obligation at year 30 may be several ownership transfers removed from the original agreement.

Agrivoltaics: The Promising Middle Ground

Agrivoltaics — the practice of co-locating solar panels and agricultural production on the same land — addresses the productive-farmland tradeoff directly. Research from MIT, the University of Arizona, and several European programs shows that certain crops (lettuce, peppers, tomatoes, herbs) and grazing animals (sheep are the most common) can coexist productively with solar panels, sometimes with better outcomes than either use alone:

• Panels shade crops during peak heat, reducing water consumption by 20–50% and improving yields for heat-sensitive varieties
• Vegetation beneath panels reduces panel operating temperatures, marginally improving electrical output
• Sheep grazing between and under panels eliminates mowing costs and produces secondary revenue

Agrivoltaics is not a solution for commodity row crops (corn, soybeans, wheat) — those require full mechanized access that panel arrays prevent. But for specialty crops and grazing operations, it represents a genuine dual-use model that reduces the farmland displacement argument substantially. Adoption is growing but remains limited: the panel spacing and height required for agricultural access increases system cost by 15–25% vs. standard ground-mount, and most developers optimize for cost, not land stewardship.

Desert and Marginal Land

On paper, desert solar looks like the obvious answer: high solar irradiance, low land value, minimal displacement of food production. The Mojave, Sonoran, and Chihuahuan deserts receive 30–40% more solar irradiance than the Midwest and cover millions of acres of land with limited competing uses.

The reality introduces complications that are easy to underestimate from a satellite view:

• Transmission distance. Desert solar resources are frequently located far from population centers and existing high-voltage transmission lines. A farm sited for optimal solar resource in the Nevada desert may face 100–200 miles of new transmission line to reach a load center. At $1M–$3M per mile for high-voltage transmission, this cost can equal or exceed the farm’s own installed cost — and transmission lines require their own permitting, environmental review, and often a decade to build. The solar farm that looks cheap in the desert may only be cheap if someone else pays for the transmission.
• Water use. Panel washing in dusty desert environments is not optional — dust accumulation on panels in the Mojave can reduce output by 5–25% between cleaning cycles. Utility-scale farms use 20–100 acre-feet of water per year for panel washing. In arid regions where water is already the binding constraint on both agriculture and municipal supply, this is a real and often underdisclosed demand.
• Desert ecosystem disruption. “Unproductive desert” is a misnomer from an ecological standpoint. Mojave and Sonoran desert ecosystems support federally protected species including the desert tortoise, Agassiz’s desert tortoise, and dozens of plant species adapted to the specific microclimate created by desert soils. Several large-scale solar projects in California (Ivanpah, Desert Sunlight, Genesis) required extensive environmental review, species relocation programs costing tens of millions of dollars, and permanent mitigation set-asides. The characterization of desert land as “waste” land available for free appropriation has been consistently challenged by biologists and federal land managers.
• BLM permitting complexity. Much of the most solar-favorable desert land in the Southwest is federally managed Bureau of Land Management (BLM) land. BLM solar permitting involves environmental impact statements, tribal consultation, wildlife surveys, and multi-agency coordination. Timelines of 5–10 years from application to approval are not unusual for large-scale projects. The permitting cost alone can reach $5M–$15M for a utility-scale project on BLM land.

None of these factors make desert solar a bad idea. They do make it more expensive, more time-consuming, and more ecologically consequential than the “put it in the desert” shorthand implies. The highest-value desert solar sites — those closest to transmission with manageable environmental footprints — are already largely permitted or in development. What remains is more difficult, more remote, or more ecologically sensitive.

Parking Canopy Structures

Solar canopies over parking lots represent the cleanest land-use case for utility solar: no new land is consumed, no habitat is disturbed, and the structures provide secondary benefits (shade, EV charging) on top of electricity generation. The United States has an estimated 800 million parking spaces occupying roughly 6,000–12,000 square miles of paved surface. If 10% of that were covered with solar canopies, the theoretical generating capacity would exceed 100 GW.

The reason it accounts for a small fraction of installed solar capacity comes down to one factor: cost.

• Standard ground-mount utility solar: $0.80–$1.20/W installed
• Parking canopy solar: $2.50–$4.50/W installed

The cost premium reflects structural steel requirements (canopies must be engineered to support panel weight, wind loads, and snow loads over a large clear span), more complex electrical design (conduit runs, lighting integration, EV charging infrastructure), and site constraints that prevent the economies of scale available to open-field installations. On a levelized cost basis, parking canopy solar produces power at roughly 2–3 times the cost of utility ground-mount.

This makes canopy projects financially marginal on power economics alone. They typically require one or more of the following to pencil out:

• High local electricity rates ($0.20/kWh or above) reducing the effective payback period
• State incentives specifically targeting dual-use or canopy installations (New Jersey, Massachusetts, and California offer enhanced incentives)
• EV charging revenue from Level 2 or DC fast chargers integrated into the canopy structure
• Corporate sustainability commitments that value on-site renewable generation at above-market rates for reporting purposes
• Avoided costs from shading — vehicles parked under canopies require significantly less air conditioning after a hot day, which has measurable value to fleets and large employers

France enacted a law in 2023 requiring solar canopies over parking lots with more than 80 spaces, phased in over 5 years. Several U.S. states are considering similar mandates. Mandates solve the market failure — the private developer cannot capture the full public benefit of dual-use land — but they also transfer the cost premium to building owners and ultimately to consumers.

From a pure land-use perspective, canopy solar is the right answer. From a pure economics perspective, it is the most expensive answer. The honest framing is that it is worth pursuing aggressively in high-rate markets and where mandates or incentives bridge the cost gap, and it is a poor use of limited capital in low-rate markets without policy support.

Solar Canals

Covering irrigation canals with solar panels is the most compelling dual-use siting concept in renewable energy today — and, outside of India, the least deployed. The idea is straightforward: mount panels on a structural framework spanning an existing canal, generate electricity from the air above the water while simultaneously reducing evaporation from the surface below. No new land is consumed. The right-of-way is already owned. The water the canals carry is more valuable in arid regions than almost any crop the land might otherwise grow.

The concept moved from academic paper to hardware in 2023, when Project Nexus — a collaboration between UC Merced, the Turlock Irrigation District, and the California Department of Water Resources — completed the first U.S. solar canal installation: a 500 kW pilot over a section of the Turlock Irrigation District canal in California’s Central Valley. India’s Gujarat state had already demonstrated the model at utility scale, operating solar canal installations since 2012 with a decade of operational data.

The Water Benefit Is the Real Story

California’s Central Valley contains more than 4,000 miles of open irrigation canals operated by dozens of water districts. These canals are exposed to the same relentless sun that makes the region attractive for solar generation — and they lose enormous volumes of water to evaporation as a result. A 2021 study from UC Merced estimated that covering all of California’s irrigation canals with solar panels could:

• Save approximately 63 billion gallons of water per year — enough to irrigate 50,000 acres of farmland or supply water to 2 million people
• Generate approximately 13 GW of solar capacity — roughly one-third of California’s current total installed electricity generation

In a state where agricultural water rights sell for $300–$1,500 per acre-foot and the Colorado River compact is being renegotiated under drought conditions, water savings of that magnitude are not a secondary benefit — they are potentially the primary justification for the investment, with the electricity generation as the bonus. That inversion changes the financial calculus substantially: if a water district can credit saved water at its marginal value against project cost, solar canal economics can approach or match ground-mount solar even at higher construction costs.

Additional operational benefits documented in early deployments:

• Algae suppression. Shading canal water significantly reduces algae growth, which otherwise requires chemical treatment and clogs pumping infrastructure. Reduced treatment cost is a real operating savings for water agencies.
• Weed control on canal banks. Panels shade the embankments, reducing vegetation maintenance costs.
• Panel cooling. Water vapor from the canal cools the panels passively. Solar panels operate more efficiently at lower temperatures — canal-mounted panels run 3–5% cooler than equivalent ground-mount installations in the same climate, marginally improving annual output.
• Existing access infrastructure. Canal maintenance roads run the length of every installation, providing vehicle access for O&M crews without requiring new road construction.

The Cost Premium and Its Justification

Solar canals cost more than ground-mount installations. The structural engineering required to span open water — accounting for wind loads, panel weight, maintenance access, and the requirement that canal operations continue uninterrupted during installation and servicing — is more demanding than driving racking posts into a flat field. Estimated installed cost for canal-mounted solar:

• Narrow canals (10–15 feet wide): $1.80–$2.50/W
• Wide canals (30–60 feet wide): $2.50–$3.50/W

The Project Nexus pilot came in at the higher end of that range for a first-of-kind installation. As designs are standardized and contractors develop canal-specific installation experience, costs are expected to decline toward the lower end — the same trajectory that brought ground-mount solar from $8/W in 2010 to under $1/W today, driven by scale and repetition.

Maintenance above water introduces complications that ground-mount does not face: panel cleaning requires equipment that can operate over the canal without polluting it, electrical connections must be sealed against humidity and occasional flooding, and inspection protocols need to account for the water surface below. None of these are unsolved problems, but they add cost relative to a dry-land installation.

Scale and the Western U.S. Opportunity

California’s canal network is the most discussed, but the opportunity extends across the entire irrigated West. The Bureau of Reclamation operates more than 8,000 miles of canals and laterals in 17 western states. State and local irrigation districts add tens of thousands of miles more. The combined evaporation loss from this network runs into the hundreds of billions of gallons annually — water that could remain in rivers, reservoirs, and aquifers under a systematic solar canal program.

The primary structural barrier to deployment at scale is not technology or economics — it is institutional. Canal systems are operated by a patchwork of irrigation districts, water agencies, and federal operators with overlapping jurisdictions, complex water rights arrangements, and procurement processes designed for infrastructure projects measured in decades. A developer who wants to install solar across 50 miles of interconnected canals may be dealing with five different governing entities, each with different boards, different legal requirements, and different priorities. Aggregating those agreements into a bankable project is a transaction cost that ground-mount developers on a single private parcel do not face.

The states most likely to move first are those where water scarcity is acute enough to force creative solutions: California, Arizona, Nevada, and Colorado. California’s AB 1378 (2023) directed state water and energy agencies to develop a solar canal deployment plan, making it the first state to formally pursue the concept at a policy level beyond the pilot stage.

The Honest Limits

Solar canals are not a universal solution. Several real constraints limit their applicability:

• Canal width constrains structural design. Very narrow laterals (under 8 feet) are difficult to span economically. Very wide main canals (over 80 feet) require longer spans with proportionally higher structural cost. The sweet spot is mid-width canals of 15–50 feet, which cover a significant but not unlimited share of the total network.
• Shading affects aquatic ecosystems in live waterways. Canals that support fish passage or that feed into rivers and streams with temperature-sensitive species need careful analysis. Full-length shading of a waterway that acts as cold-water habitat is not environmentally neutral.
• Not all canals are near transmission infrastructure. A canal running through remote agricultural land may be miles from an adequate grid connection, reintroducing the interconnection cost problem that plagues remote ground-mount installations.
• Seasonal canal operations limit panel access. Many irrigation canals run seasonally, are drained for winter maintenance, and experience variable water levels that can affect structural stability and access. Installation and servicing schedules must be coordinated with canal operations.

These are engineering and logistical constraints, not fundamental barriers. The Project Nexus data — the first rigorous U.S. operational dataset — confirmed that the water savings estimates hold in practice, that the panels performed at or above projected output, and that canal operations were unaffected. The technology works. The challenge is scaling an institutional and commercial model to match the physical opportunity.

Factor	Productive Farmland	Desert / Marginal	Parking Canopy	Irrigation Canal
Installed cost ($/W)	$0.90–$1.20	$0.80–$1.10	$2.50–$4.50	$1.80–$3.50
Land cost / acquisition	Lease $500–$2,000/acre/yr	Low, but permitting is expensive	No new land required	Right-of-way already owned by water agency
Transmission access	Generally good (rural grid)	Often poor; high cost	Urban / suburban; existing grid	Variable; canals often run through remote farmland
Capacity factor	18–26%	24–32%	15–25%	22–30% (panel cooling adds 3–5%)
Environmental impact	Moderate; food displacement	High; habitat disruption	Minimal; dual-use	Minimal to positive; reduces evaporation and algae
Secondary benefits	Agrivoltaics possible	Few	Shade, EV charging, heat island	Water savings (63B gal/yr in CA), algae reduction, weed control
Project IRR (unlevered)	7–10%	8–12% (high irradiance)	4–7% (requires incentives)	6–10% (improves significantly when water savings are credited)
Overall land-use verdict	Acceptable with agrivoltaics; problematic otherwise	Best irradiance; worst environmental record	Best land-use; worst economics	Best dual-use case; limited by institutional complexity and canal width

Is Solar Farm Power Actually Clean?

The short answer is yes — significantly cleaner than fossil fuel power across its operational life. The longer answer is that “clean” is not binary, the manufacturing and end-of-life phases carry real environmental costs, and several impacts are systematically omitted from standard lifecycle analyses.

Lifecycle Carbon: The Honest Numbers

Lifecycle emissions for electricity generation (grams of CO₂ equivalent per kilowatt-hour, including manufacturing, construction, operation, and decommissioning):

• Coal: 820–1,050 g CO₂e/kWh
• Natural gas (combined cycle): 410–650 g CO₂e/kWh
• Utility solar PV: 20–50 g CO₂e/kWh
• Wind: 7–15 g CO₂e/kWh
• Nuclear: 4–12 g CO₂e/kWh

Solar is 15–40 times cleaner than coal and 8–20 times cleaner than gas on a per-kWh lifecycle basis. That margin is large enough that the directional conclusion — solar is better for the climate than fossil fuels — is robust even under pessimistic manufacturing assumptions.

The energy payback period — the time a panel must operate to generate the energy used to manufacture it — is approximately 1–3 years for monocrystalline panels in average sun conditions, and under 2 years in high-irradiance locations. For a panel with a 25-year operating life, the remaining 22–24 years of production are essentially carbon-free from an energy-debt perspective.

What Lifecycle Analysis Typically Misses

Standard lifecycle analyses capture manufacturing energy and operational emissions reasonably well. Several other environmental dimensions are handled less consistently:

Mining and Raw Material Extraction

Solar panels require silicon (abundant and relatively benign to extract), silver (scarce, energy-intensive to mine), aluminum (energy-intensive to smelt, though highly recyclable), and copper (significant mining footprint). A 100 MW solar farm contains approximately:

• 1,500–2,500 metric tons of aluminum (frames and racking)
• 4–8 metric tons of silver (cell contacts)
• 800–1,200 metric tons of copper (wiring and busbars)

Silver is the most constrained material. Solar currently accounts for roughly 10% of global silver demand, and that share is growing. Silver mining generates significant tailings waste and operates in regions with variable environmental standards. The lifecycle emissions models generally include the energy cost of silver production but not the localized habitat and water impacts of mining operations in, for example, Mexico or Peru, where much of the world’s silver originates.

Cadmium telluride (CdTe) panels, used by First Solar, substitute toxic cadmium for silver in the cell contacts. CdTe panels have a lower manufacturing carbon footprint than crystalline silicon but introduce cadmium into the end-of-life waste stream — a material that is hazardous at low concentrations in soil and groundwater.

Panel Disposal: The Coming Wave

This is the most underacknowledged problem in solar’s environmental story. The first large wave of utility-scale solar panels installed in the mid-2000s through 2010 will reach end of life between 2030 and 2035. The International Renewable Energy Agency estimates that cumulative solar panel waste will reach 78 million metric tons globally by 2050.

The current state of panel recycling:

• The United States has no federal mandate for solar panel recycling. Panels are classified as solid waste in most states and can be legally landfilled.
• The European Union requires manufacturer take-back under the WEEE (Waste Electrical and Electronic Equipment) directive. The U.S. does not have an equivalent.
• Commercial-scale panel recycling infrastructure in the U.S. is nascent. A handful of companies (We Recycle Solar, First Solar’s own program) operate recycling facilities, but capacity is a small fraction of the waste that will arrive over the next decade.
• The economics of panel recycling are currently negative in most cases: the recovered value of silicon, silver, and glass does not cover collection, transportation, and processing costs at current commodity prices. Recycling requires either a mandate or a subsidy to be economically viable at scale.

The practical outcome is that most of the 3–4 million tons of solar panels reaching end of life in the U.S. over the next 10–15 years will be landfilled. Panels contain lead solder and, in some chemistries, cadmium — materials that can leach into groundwater from conventional landfills. The environmental credit solar receives for displacing fossil fuel emissions needs to be held against a deferred waste disposal cost that the industry has not yet confronted at scale.

Grid Integration and Curtailment

Solar power has a fundamental mismatch with electricity demand: it peaks at midday and produces nothing at night, while electricity demand peaks in the morning and evening. As solar penetration increases, this mismatch creates two problems with environmental implications:

• Curtailment. When the grid cannot absorb all solar output, grid operators curtail (turn off) solar generation. California regularly curtails 5–15% of solar output during spring afternoons. Curtailed solar is wasted generation — the carbon cost of manufacturing the panel was paid, but no electricity was produced. As solar share grows without commensurate storage buildout, curtailment rates increase and the effective carbon intensity per useful kWh rises.
• Backup generation requirements. Because solar does not produce at night or during extended cloudy periods, every solar megawatt requires backup generating capacity — typically natural gas peaker plants — that must be available when solar is unavailable. The carbon cost of maintaining, starting, and cycling those gas plants is real and is not attributed to the solar capacity they back up. True system-level carbon accounting would assign a share of backup generation emissions to the intermittent source that necessitates them. Virtually no published lifecycle analysis does this.

The Net Environmental Verdict

Solar is substantially better for the climate than fossil fuel generation on a lifecycle basis, and that conclusion survives most reasonable adjustments for manufacturing, mining, and curtailment. The directional case is sound.

The honest qualifications are:

• Desert habitat disruption from large-scale Southwestern installations is significant and largely irreversible on human timescales
• Panel recycling infrastructure does not exist at the scale needed, and the industry is running a deferred liability on panel disposal that will land on someone in the 2030s and 2040s
• Silver scarcity is a real constraint on solar expansion at the scale required for grid decarbonization — technology transitions toward silver-reduced or silver-free cell architectures are underway but not complete
• System-level carbon accounting that includes backup generation requirements would raise solar’s effective emissions per reliable kWh, though it would still be well below fossil fuels

Solar is a net environmental positive. It is not a free environmental positive, and the industry’s tendency to present it as one does a disservice to the real-world tradeoffs that need to be managed.

Bottom Line

Is the ROI real?

Yes — for well-sited, well-structured projects. A utility-scale solar farm with a 20-year PPA, good solar resource, reasonable transmission access, and realistic decommissioning reserves produces an unlevered IRR of 7–10% with predictable, long-duration cash flows. That is a legitimate infrastructure return and explains why pension funds and sovereign wealth funds have become the dominant owners of utility solar assets.

The ROI degrades substantially under unfavorable conditions: distant transmission, expired or missing PPAs, merchant power in oversupplied markets, or sites with high decommissioning complexity. Projects marketed on optimistic capacity factors, omitted decommissioning costs, or expected REC prices that assume favorable policy are overstated.

Which land use is right?

In order of overall value (financial + environmental + societal):

• Irrigation canals are the highest-value siting opportunity that is currently underbuilt. No new land consumed, water savings that can independently justify the investment in arid regions, existing right-of-way, and a panel cooling benefit that improves output. The barrier is institutional, not technical or economic. In the water-scarce West, this should be the first priority, not an afterthought.
• Parking canopies are the right answer where incentives and electricity rates make the economics viable. No land consumed, significant secondary benefits, and the U.S. has an enormous underutilized inventory. The priority should be to build more of this, with policy support where needed to bridge the cost gap.
• Desert and marginal land makes sense when transmission access is reasonable and environmental review is genuinely rigorous. The best sites are largely committed; what remains requires careful selection. “Put it all in the desert” ignores real ecological costs and real transmission economics.
• Productive farmland is defensible only with agrivoltaic design that maintains meaningful agricultural use, contractual decommissioning obligations with real financial backing, and honest accounting of soil condition at lease end. Standard utility ground-mount on prime corn-belt farmland is a poor trade against the alternatives.

Is it good for the environment?

Overall, yes. The climate benefit of displacing fossil fuel generation is real and large. The environmental costs — habitat disruption, mining impacts, and deferred panel disposal — are real and need to be managed, not dismissed. The industry’s record on the latter is not strong: decommissioning obligations are underfunded, recycling infrastructure is inadequate, and desert habitat impacts from the first generation of large-scale Mojave projects were more severe than initial environmental reviews projected.

A solar industry that takes its environmental responsibilities seriously — on recycling, on habitat mitigation, on decommissioning bonds, and on honest lifecycle accounting — produces substantially better outcomes than one that treats “zero emissions during operation” as a license to ignore everything else.

Solar farms deliver genuine financial and environmental value — but only when the full ledger is kept honestly, not just the favorable half of it.