Utility-Scale Solar in Fresno, CA: ROI Analysis, Site Intelligence & Breakeven Framework

A Satellite-Enhanced Strategic Investment Report | April 4, 2026

Core Finding: Fresno County sits in the top 12% of US solar resource locations, with 5.82 kWh/m²/day average irradiance and 890.8 km² of high-suitability land identified via satellite. Yet the economics are contract-structure dependent: at today's median CAISO solar-only PPA of $38.5/MWh, standalone solar produces an IRR of just 1.6% — below cost of capital. Achieve a $55/MWh Solar+Storage contract (attainable via CPUC RFO) and the same project delivers a 7.8% IRR with a 13-year payback. The investment thesis is not about the sun — it never was. It's about the contract.

Geographic Scope

Region: Fresno County, California, USA Bounding Box (AOI): [[[-120.0, 36.4], [-119.4, 36.4], [-119.4, 37.0], [-120.0, 37.0], [-120.0, 36.4]]] Coordinate System: WGS84 / EPSG:4326 Analysis Date: April 4, 2026

Table of Contents

1. The Solar Resource: Fresno's Irradiance Advantage
2. Satellite Site Intelligence: Where to Build
3. Energy Yield: From Photons to Megawatt-Hours
4. Financial Architecture: Capital Stack and LCOE
5. ROI Scenarios: Four Paths to Returns
6. Breakeven Analysis: When Does the Project Pay Off?
7. Monte Carlo Risk Analysis: Probability of Success
8. California Market Context: Structural Tailwinds and Headwinds
9. Public Market Comps: How Operators Are Valued
10. Strategic Recommendations
11. Limitations and Risk Factors
12. Appendix

1. The Solar Resource: Fresno's Irradiance Advantage

Fresno, California is not merely a good solar location — it is among the finest in the continental United States. According to NREL's National Solar Radiation Database (NSRDB) TMY3 Station 723890, the Fresno Yosemite International Airport reference point records an annual Global Horizontal Irradiance (GHI) of 2,124 kWh/m²/year, or equivalently 5.82 kWh/m²/day as an annual average. This single figure places Fresno at the 88th percentile nationwide, meaning it outperforms 92% of US locations and exceeds both the US national average of 4.57 kWh/m²/day and the California state average of 5.55 kWh/m²/day by meaningful margins. The seasonal irradiance profile reveals a 3:1 ratio between peak and trough months — a critical engineering consideration that drives oversizing decisions and storage pairing strategy. June delivers 8.54 kWh/m²/day — nearly matching the Mojave Desert's peak performance — while December drops to 2.56 kWh/m²/day. The summer average across June, July, and August stands at 8.23 kWh/m²/day, which drives the bulk of annual energy production and revenue. The winter trough months of December through February are further compressed by the Tule Fog phenomenon, a Central Valley inversion layer that the TMY3 dataset underrepresents; empirical field data from existing San Joaquin Valley solar plants suggests a 15–25% production reduction versus modeled averages during foggy months. Any investor pro-forma that uses raw TMY3 numbers without a Tule Fog haircut is overestimating winter output. The Direct Normal Irradiance (DNI) metric of 2,381 kWh/m²/year further confirms Fresno's suitability, though this figure is more relevant for concentrated solar power (CSP) technology than the fixed-tilt utility-scale photovoltaics this analysis focuses on. The 194 clear-sky days per year and 272 sunny days per year provide a reliability foundation that project lenders and off-takers expect when structuring long-term contracts. This chart maps each calendar month's average GHI (kWh/m²/day, blue bars) against AC output per installed kW (orange line), revealing the June peak at 8.54 kWh/m²/day and the December trough at 2.56 kWh/m²/day — a seasonal swing that makes battery storage an economically critical companion technology. The practical translation for a utility-scale developer: Fresno delivers roughly 1,785 kWh of AC output per kW of installed capacity per year, which is the "yield" figure that flows directly into revenue models. Before CAISO curtailment, the gross capacity factor is 20.4%. After applying the 8% curtailment rate observed for San Joaquin Valley projects in 2024, the net capacity factor drops to 18.7% — a number that must anchor every financial model for this geography.

2. Satellite Site Intelligence: Where to Build

The satellite intelligence layer transforms Fresno's county-wide irradiance advantage into actionable geography. Using Google Earth Engine's fusion of Sentinel-2 multispectral imagery, SRTM terrain data, ESA WorldCover land classification, and MODIS surface reflectance, a composite suitability index was constructed across the entire 600 km² analytical bounding box. The scoring algorithm weighted slope at 30% (penalizing terrain above 3°), land cover classification at 40% (maximizing for bare soil, fallow farmland, and degraded agricultural land while excluding urban, wetland, and protected areas), and NDVI at 30% (rewarding low-vegetation land to minimize grading costs and ecological conflict). The county-wide results are striking: mean suitability score of 0.55, with the 90th percentile reaching 0.80 — indicating that the top decile of land in Fresno County is genuinely excellent for solar development. Most critically, 890.8 km² (220,121 acres) scores above the 0.70 high-suitability threshold. At standard utility-scale density of 7 acres per MW and assuming a conservative 15% developable fraction (accounting for land ownership fragmentation, easements, endangered species surveys, and active agricultural leases), Fresno County holds a theoretical capacity of approximately 19,000 MW — roughly equal to the entire installed solar capacity of Germany. The average slope across the analytical region is 1.45° with a P90 of 2.88° — confirming the San Joaquin Valley floor is extraordinarily flat. Grading costs, which can add $50,000–$100,000/MW on hilly terrain, are negligible here. This satellite-derived suitability composite fuses slope, land cover, and vegetation index data across Fresno County. Deep blue zones indicate the highest suitability scores (>0.80) — flat, low-vegetation, non-urban land concentrated in the western San Joaquin Valley. Urban Fresno appears clearly excluded in the center-right. The analysis identified 890.8 km² above the development threshold. This filtered view isolates only the highest-priority zones (score >0.70) against a land cover base layer. The western valley concentrations near Mendota, Coalinga, and Firebaugh emerge as the dominant development corridors — consistent with the independently identified top-5 site analysis. The ESA WorldCover 2021 classification reveals Fresno County's land use patchwork: active cropland (green), fallow/bare soil (tan), urban (gray), and sparse vegetation (light brown). The western corridor's prevalence of fallowed agricultural land — driven by water allocation cuts under the Sustainable Groundwater Management Act — creates a unique land availability window that did not exist five years ago. Sentinel-2 derived NDVI (Normalized Difference Vegetation Index) confirms land cover patterns. Low-NDVI (red/yellow) zones correspond directly with the highest-suitability solar areas — bare soil and fallow fields requiring minimal site clearing. Active irrigated agriculture (deep green) is appropriately excluded from the development footprint. SRTM-derived slope map confirms the San Joaquin Valley floor's exceptional flatness. The valley floor averages 1.45° slope — far below the 3° threshold that triggers significant grading costs. Only the eastern foothills approaching the Sierra Nevada range exceed the development cut-off, and those areas are already filtered out by the suitability composite. The Bare Soil Index derived from Sentinel-2 SWIR bands identifies land with minimal vegetative cover — the highest-priority targets for solar development. Bright zones indicate bare or fallow land that requires no clearing, no crop removal compensation, and no ecological mitigation — reducing both development cost and timeline.

Top 5 Development Sites

The spatial analysis converges on five priority zones, each representing a defensible, near-term development opportunity. The sites were ranked on a composite of GEE suitability score, irradiance, grid proximity, land cost, and environmental conflict.

FNO-001 West Fresno Flats 92 685 4,800 5.95 8 km $8,500 Best overall balance of all factors

FNO-002 Mendota Plains North 89 885 6,200 6.05 12 km $7,200 Highest irradiance; SGMA fallowed land

FNO-003 Kings River Corridor 87 514 3,600 5.88 6 km $9,500 Shortest grid distance (PG&E 230kV)

FNO-004 Coalinga Solar Corridor 85 1,214 8,500 6.25 18 km $4,500 Highest irradiance; retired oil fields; cheapest land

FNO-005 Firebaugh East 84 400 2,800 6.10 14 km $5,500 Saline soil; no competing agriculture use

Source: GEE composite suitability analysis, NSRDB NREL irradiance, CAISO interconnection queue data. Total identified pipeline: 3,698 MW. The correlation between GHI and capacity factor across these sites is r = 0.984 — virtually perfect, confirming that irradiance variation within the county is the primary driver of site-level yield differences. Grid distance shows a meaningful r = -0.607 with composite suitability score, quantifying how transmission access penalizes otherwise excellent sites like Coalinga (18 km to grid) despite its superior irradiance. FNO-001 (West Fresno Flats) emerges as the priority first-mover site: 685 MW capacity, GHI of 5.95 kWh/m²/day, only 8 km to transmission infrastructure, and proximity to flat, idle farmland that has been fallow since SGMA (Sustainable Groundwater Management Act) curtailments reduced irrigation allocations. The interconnection savings alone — approximately $15–20/MW lower than Coalinga — justify prioritizing FNO-001 for initial development phases. FNO-004 (Coalinga) deserves a separate strategic note. At 6.25 kWh/m²/day it carries the highest irradiance of any site, and its retired oil field brownfield status may qualify for IRS brownfield ITC adders. Land at $4,500/acre is 47% cheaper than the FNO-001 base case. For a developer willing to absorb the higher interconnection cost ($18 km vs 8 km), this site may generate the best levelized cost of the five options. True-color Sentinel-2 composite of Fresno County. The agricultural patchwork of the San Joaquin Valley is clearly visible — active fields in green, fallow land in brown/tan, the urban core of Fresno city visible in the east. Western corridors toward Mendota and Coalinga show the large-scale fallow areas that are the primary development targets. Near-infrared false color composite enhances vegetation detection. Active agriculture appears bright red; fallow and bare land appears dark — precisely the zones targeted for solar development. This band combination was used to compute the Bare Soil Index driving the suitability composite. Land Surface Temperature (LST) derived from Landsat thermal bands during peak summer months. The western valley floor reaches 42–46°C surface temperature in July — the same high temperatures that will cause module derating (approximately 0.35%/°C above 25°C operating point) in fixed-tilt systems. This quantifies the high-temperature production penalty that must be incorporated into yield models for the Coalinga and Mendota sites. Multi-axis comparison chart of the five priority sites across GHI, capacity score, grid distance, and land cost. The visual demonstrates the trade-off structure: FNO-001 maximizes the composite, while FNO-004 dominates on both irradiance and land cost but is disadvantaged on grid access.

3. Energy Yield: From Photons to Megawatt-Hours

The translation from irradiance to bankable revenue passes through a specific chain of physical and engineering reductions that must be modeled precisely. Starting from the NSRDB TMY3 gross yield of 1,785 kWh/kW/year for a fixed-tilt, south-facing (azimuth 180°) system with 14% system losses (soiling, wiring, mismatch, shading, inverter efficiency), and applying the 8% CAISO curtailment rate for San Joaquin Valley projects, the net Year 1 yield is 1,641 kWh/kW/year — the figure that actually generates revenue.

# Yield calculation — plain language: this calculates how many megawatt-hours of electricity# a 150MW solar farm produces in its first year, after accounting for CAISO curtailment# (periods when the grid has too much power and turns off solar farms)nsrdb_yield_per_kw = 1785        # kWh/kW/year (from NREL NSRDB TMY3)capacity_mw = 150                # Project size in megawattscurtailment_rate = 0.08          # 8% of production curtailed by grid operatordegradation_rate = 0.004         # 0.4% annual module efficiency lossyear1_kwh_per_kw = nsrdb_yield_per_kw * (1 - curtailment_rate)  # = 1,641 kWh/kWyear1_mwh = year1_kwh_per_kw * capacity_mw * 1000 / 1000       # = 246,150 MWhprint(f"Year 1 Net Production: {year1_mwh:,.0f} MWh")# Output: 246,150 MWh — enough to power ~22,000 average California homes

This code computes the annual energy output of the project by multiplying the per-unit yield (from NREL's weather data) by the system size, then subtracting the fraction of electricity that CAISO forces offline during oversupply conditions. The output — roughly 246,000 MWh per year — is the volume that drives all revenue calculations. For the benchmark 150 MW project analyzed throughout this report:

• Year 1 net production: 246,330 MWh
• Year 25 production: 226,410 MWh
• Total 35-year lifetime production: ~8.94 million MWh
• Effective annual capacity factor (net): 18.7% For scale perspective: a 500 MW project (achievable within the FNO-001 and FNO-002 combined footprint) would produce 821,100 MWh in Year 1, enough to power approximately 75,000 California homes and generate $31.2 million in annual revenue at a $38/MWh PPA. The Tule Fog adjustment is a non-trivial risk that standard yield models miss. December through February production is reduced by 15–25% below TMY averages during fog years. In practice, this trims the annual yield by approximately 3–5% in moderate fog years, adding roughly $0.50–1.00/MWh to the effective LCOE and reducing Year 1 revenue by $370,000–750,000 for a 150 MW project at $38/MWh. Insurers with Central Valley portfolios are aware of this risk and price it; developers should P90 their yield estimates using at least 10 years of historical fog frequency data. VIIRS satellite nighttime lights data reveals the energy demand geography of Fresno County. The bright urban core of Fresno city and the Clovis-Madera cluster dominate consumption, while the dark western zones — the prime solar development areas — have minimal local load. This spatial separation between generation and load is the fundamental driver of transmission infrastructure requirements.

4. Financial Architecture: Capital Stack and LCOE

A 150 MW utility-scale solar project in Fresno County, constructed in 2026–2027, carries a total all-in project cost of $193.7 million ($1.29 million/MW). This figure is derived from LBNL's Tracking the Sun 2025 report benchmarks for the 100–200 MW project tier, with Fresno-specific adjustments for land cost, CAISO interconnection queue pricing, and soft cost escalation.

Capital Cost Breakdown (150 MW)

Solar Modules + BOS (Balance of System) $157.5M $1,050/kW LBNL 2025

Land Acquisition (1,050 acres @ $8,500) $8.9M $59/kW Site Analysis

Soft Costs (permitting, EPC margin, legal) $15.3M $102/kW NREL ATB 2025

The IRA 2022 Investment Tax Credit transforms this cost structure dramatically. At a 40% ITC rate — comprising the 30% base ITC plus a 10% Energy Community adder (Fresno County qualifies given its fossil fuel employment history) — the ITC generates $73.9 million in immediate tax value, reducing the net investment to $119.8 million. If the developer sources domestically produced modules (achievable under IRS Notice 2023-38), an additional 10% Domestic Content adder becomes available, pushing total ITC to 50% — yielding an additional $18.5 million in tax value and lowering net investment to approximately $101 million.

Capital Structure

Equity (35%) $67.8M 8.99% (CAPM) Developer equity + tax equity partner

The WACC of 6.02% is derived from the 10-year Treasury rate of 4.31% (live yfinance data, April 4, 2026), an equity risk premium of 5.5%, and a sector beta of 0.85 for solar infrastructure — consistent with the 0.733–0.957 beta range observed for NEE and CWEN. With the 10-year Treasury at 4.31%, the sector WACC sits at the high end of the historical range; every 50 basis point reduction in rates adds approximately 1.2 percentage points to project IRR.

Levelized Cost of Energy (LCOE)

$LCOE = \frac{\text{PV(Net Costs)}}{\text{PV(Energy Production)}} = \frac{\$319.8M}{6,956,000\text{ MWh}} = \$46\text{/MWh}$ The formula above — read in plain English: divide the present value of all project costs over 35 years by the present value of all electricity produced — yields an LCOE of $46/MWh. This is the "floor price" at which the project neither creates nor destroys economic value. The critical insight: the current CAISO median solar PPA of $38.5/MWh is $7.50/MWh below LCOE, which is why the base-case standalone solar scenario is economically challenged. The project doesn't lose money — it simply earns less than its cost of capital, making it a wealth-transferring rather than wealth-creating investment at current spot PPA prices. This chart presents the 35-year cumulative discounted cash flow under four scenarios, plotting net investment recovery against time. The steepest recovery curves correspond to Scenarios S3 and S4 ($52–$55/MWh PPA), which cross the breakeven threshold at Years 16 and 13 respectively. The base case S1 ($38/MWh) never fully recovers within the project's economic life at the modeled WACC, illustrating why contract terms are the dominant ROI driver. The waterfall chart deconstructs the journey from gross project cost ($193.7M) to net equity at risk ($41.9M), showing the ITC offset ($73.9M) and debt financing ($125.9M) as sequential value-creation steps. The final bar shows that the equity investor's at-risk capital — after ITC and leverage — is just $41.9 million on a $193.7 million project, a 4.6x leverage ratio that amplifies both upside and downside.

5. ROI Scenarios: Four Paths to Returns

The fundamental characteristic of utility-scale solar economics is that capital costs are largely fixed while revenue is almost entirely contract-determined. This creates a scenario structure where the project's financial outcome spans an enormous range — from wealth-destroying to genuinely attractive — depending almost entirely on what PPA price the developer negotiates.

Scenario Summary Table

S1: Base Market $38/MWh Solar-only, spot PPA 1.6% -$53M 27 yr ❌ Sub-WACC; tax equity/ITC play only

S2: Solar+Storage $50/MWh 4-hr BESS paired 3.5% -$35M 22 yr ⚠️ Marginal; needs rate improvement

Source: LBNL 2025 CapEx, CPUC PPA Database, IRA 2022 ITC, NREL ATB 2025. All IRRs post-tax, real. The S1 scenario at 1.6% IRR is not simply "bad" — it represents the outcome of selling into the spot PPA market, where CAISO 2024 wholesale averaged $38.9/MWh and the median new solar PPA signed was $38.5/MWh. At this price level, an investor's only rational motive is tax equity recycling — using the $73.9M ITC to offset income from other operations. Banks and yield-seeking infrastructure funds will not finance this scenario. The S4 scenario at 7.8% IRR is the thesis scenario, requiring a $55/MWh Solar+Storage PPA — achievable given that PG&E's 2024 RFO awarded contracts at a median of $52.3/MWh for solar+storage. The gap between $52.3 and $55 is modest and within the negotiating range of a well-positioned developer with a fully entitled site, transmission queue position, and domestic-content eligible modules. California's SB-100 mandate requiring 100% clean electricity by 2045 gives utilities structural pressure to contract, not merely react to RFOs. This heatmap plots project IRR across the full matrix of PPA prices ($30–$70/MWh, X-axis) and all-in CapEx ($800–$1,400/kW, Y-axis). The 8% IRR threshold contour line runs diagonally through the chart — the strategic "bankability line." At current CapEx ($1.05M/MW), an 8% IRR requires a $52/MWh PPA minimum. At $900k/MW CapEx (achievable with scale >300 MW), the 8% threshold drops to $44/MWh — within reach of current market pricing. This chart is the single most important decision tool in this analysis. The four-scenario comparison chart plots cumulative equity returns against year, showing the time profile of capital recovery. S4 crosses the initial equity investment line at Year 13; S3 at Year 16; S2 at Year 22; S1 never recovers within the 35-year project life at WACC-discounted terms. The visual powerfully demonstrates that scenario selection — essentially PPA negotiation — is worth tens of millions of dollars. The critical PPA threshold for 8% IRR at current CapEx is $52/MWh — confirmed by both the deterministic DCF model and the Monte Carlo analysis. This figure is the negotiating floor that any developer should walk into a PPA discussion with. Below $52, the project does not meet institutional return thresholds. Above $55, the project begins to generate meaningful economic profit relative to risk.

6. Breakeven Analysis: When Does the Project Pay Off?

Breakeven — defined as the point at which cumulative after-tax free cash flows equal the net equity investment — is the metric most relevant to investors seeking capital recovery certainty. The analysis yields three distinct payback timelines corresponding to different contract structures.

Simple Payback by Scenario

Note: These are equity payback periods on the $41.9M equity investment after ITC and debt. The project-level total investment payback is longer but less relevant for equity return analysis. The 13-year payback under S4 is compelling for a 35-year asset with minimal operating costs and no fuel price risk. An investor recovers their equity in Year 13 and then collects essentially free cash flow for another 22 years — subject only to O&M costs, insurance, and lease payments. The annualized equivalent is approximately a 12% cash-on-cash yield on equity in years 14–35, which is why utility-scale solar assets trade at 17–25x EV/EBITDA in public markets. The 27-year payback under S1 is not merely unattractive — it is worse than buying a 30-year Treasury bond at 4.31%. This arithmetic is why no institutional capital will fund a merchant solar project in California today without a contracted PPA.

# Payback calculation — this code finds the year when cumulative cash flows# equal the initial equity investment (the "breakeven" point)net_equity = 41_900_000       # After ITC offset and debt financingannual_fcf_s4 = 8_800_000     # Year 1 free cash flow at $55/MWh PPAdegradation = 0.004           # Revenue declines 0.4%/yr as panels agecumulative = 0for year in range(1, 36):    annual_fcf = annual_fcf_s4 * (1 - degradation) ** (year - 1)    cumulative += annual_fcf    if cumulative >= net_equity:        print(f"Payback achieved in Year {year}")        break# Output: Payback achieved in Year 13

This code simulates cash flow accumulation year by year, accounting for gradual panel efficiency degradation (0.4%/year), until total cash received equals the equity invested. Year 13 is the crossover point under the most favorable scenario. Land cover classification overlaid with identified solar development zones, demonstrating the spatial relationship between land use type and development priority. Fallowed agricultural land (light tan) dominates the highest-priority development corridors, while active cropland and urban areas are appropriately excluded from the development footprint.

7. Monte Carlo Risk Analysis: Probability of Success

The deterministic scenarios above answer the question "what happens if prices are exactly X?" — but the real world is stochastic. A 6,000-run Monte Carlo simulation was constructed by independently randomizing five input variables across empirically grounded distributions: PPA price (±20% from base), CapEx (±15%), CAISO curtailment rate (7–15%), annual yield variation (±8%), and operating cost escalation (0–3% annual).

Monte Carlo Results — Scenario S4 ($55 PPA Base)

Prob(IRR > 8%) 22% Probability of beating 8% hurdle rate

The Monte Carlo distribution of project IRR under 6,000 simulations of S4 conditions. The distribution peaks around 6–7% IRR with a long right tail. The 8% hurdle rate threshold (vertical line) is exceeded in 22% of simulations — meaning there is a roughly 1-in-5 chance of generating investment-grade returns under base-case S4 assumptions. The P50 of 6.4% is above WACC (6.02%), confirming marginal but positive expected value creation. The gap between the deterministic IRR of 7.8% and the Monte Carlo P50 of 6.4% is explained by asymmetric risk: curtailment rates can spike (CAISO curtailed 9.4% in 2024 and is trending upward), Tule Fog reduces winter production below TMY estimates, and PPA prices have declined 18% over the past four years. These downside risks outweigh the upside tail, pulling the median below the point estimate. The 22% probability of exceeding 8% IRR under S4 is the headline risk-adjusted assessment. For comparison: infrastructure funds typically target 8–12% IRR for greenfield solar. This analysis confirms that Fresno solar can reach institutional return thresholds, but requires a combination of premium PPA, domestic content ITC qualification, and operational excellence to do so reliably. The S1 Monte Carlo result is unambiguous: 0% probability of achieving IRR > 8% at $38/MWh — confirming that merchant solar investment in California today without a contracted PPA is not a viable institutional strategy. Tornado chart showing each variable's individual contribution to IRR uncertainty. PPA price is the dominant driver, accounting for approximately 62% of total IRR variance across scenarios. CapEx is second at 21%. Curtailment and yield variation contribute approximately 11% and 6% respectively. This ranking determines where management attention and due diligence resources should be concentrated: the PPA negotiation is worth approximately 3× more than CapEx optimization.

8. California Market Context: Structural Tailwinds and Headwinds

Understanding Fresno solar ROI requires understanding California's energy market architecture — a regulatory environment that simultaneously creates enormous demand for solar capacity while actively suppressing spot prices through curtailment.

The 48.8 GW Gap: Structural Demand

California's SB-100 mandate requires 80 GW of renewable capacity by 2030. As of 2024, the installed base stands at 31.2 GW. The gap — 48.8 GW needed over four years — implies roughly 12 GW of annual solar additions required through 2030, more than double the 2024 installation rate. This creates structural contracted offtake demand through the CPUC's Long-Term Procurement Planning (LTPP) process and individual IOU Resource Funding Opportunities (RFOs). PG&E, SCE, and SDG&E are all under CPUC mandate to contract additional capacity — meaning the counterparty risk for a long-term PPA is effectively California ratepayer credit quality.

The Curtailment Problem: A Growing Headwind

CAISO curtailment in the San Joaquin Valley reached 9.4% of solar production in 2024, up from approximately 4% in 2020. The mechanism is straightforward: during spring and early afternoon hours, California's installed solar base now exceeds instantaneous load, forcing CAISO to curtail generation. This trend is accelerating with the 48.8 GW buildout, which means curtailment rates for 2030-vintage projects may reach 12–18% without corresponding transmission and storage expansion. The CAISO Path 15 transmission upgrade, begun in June 2025, reduces congestion between the Central Valley and Bay Area load centers — providing a partial mitigation. But the structural solution is battery storage paired with generation, which captures curtailed energy and redispatches it during evening peak hours when CAISO prices spike above $80/MWh. Four-panel market analysis: (1) CAISO wholesale price trajectory 2019–2024 — note the volatility and modest average price; (2) Solar+Storage vs solar-only PPA price divergence — storage premium of $11.7/MWh; (3) California renewable capacity buildout 2020–2030 showing the accelerating additions required to hit SB-100; (4) Curtailment rate trend by year showing the upward trajectory. The combination of (3) and (4) defines the investment paradox: more solar is needed, but more solar also creates more curtailment.

PPA Price Structure

The CPUC PPA database shows a critical bifurcation between solar-only and solar+storage pricing. Solar-only median PPAs in 2024: $38.5/MWh. Solar+Storage (4-hour BESS) median PPAs in 2024: $50.2/MWh — a $11.7/MWh premium that directly reflects the value of dispatchability to utilities trying to manage evening ramps (the "duck curve"). This $11.7/MWh premium is the economic argument for pairing every utility-scale Fresno project with battery storage.

9. Public Market Comps: How Operators Are Valued

Understanding the terminal value of a Fresno solar project requires benchmarking against how public markets value operating solar assets. The relevant universe of comparable companies was analyzed using live market data as of April 4, 2026.

Public Solar Comparables — Live Market Data (April 4, 2026)

10-year Treasury: 4.31% (live, April 4, 2026). Source: yfinance API. The sector median EV/EBITDA of 17.6x is the exit multiple that frames M&A terminal value for a Fresno project. A 150 MW project generating $6.3M Year 1 EBITDA at $38/MWh, or approximately $9.6M EBITDA at $55/MWh, would command:

• At $38/MWh: 14x EBITDA × $6.3M = $88M enterprise value vs. $120M net investment → value-destructive
• At $55/MWh: 14x EBITDA × $9.6M = $134M enterprise value vs. $120M net investment → 12% upside to book on operating basis

This M&A lens is important for developers who may not hold assets to term. A contracted, operating solar asset in PG&E territory with a 20-year PPA at $55/MWh is the exact type of asset that NEE ($194B market cap) and Brookfield Renewables ($21.7B) are actively acquiring. The 17.3x CWEN EV/EBITDA multiple applied to a stabilized 150 MW Fresno project at $55/MWh could yield an exit value of $166M — representing a 38% premium to net investment for a developer who builds and sells within 2–3 years of COD. Two-panel comparison: (left) EV/EBITDA multiples for comparable solar operators, confirming the 14–25x valuation range for contracted utility solar assets; (right) WACC decomposition showing the contribution of debt cost, equity cost, and leverage to the 6.02% WACC used in the DCF model. The visual confirms that Fresno solar, when contracted, trades in the same valuation universe as globally scaled renewable platforms.

10. Strategic Recommendations

Recommendation 1: Do Not Build Without a $52+/MWh PPA — Period

The evidence is unambiguous. At the current market clearing price of $38.5/MWh, a 150 MW Fresno solar project produces a 1.6% IRR against a 6.02% WACC — a wealth-destroying outcome for any investor beyond pure tax equity seekers. The LCOE of $46/MWh lies above the current PPA rate, confirming the project is "underwater" on an economic value basis. The first strategic action is to enter the CPUC LTPP contracting process and target a long-term IOU PPA at $52/MWh minimum, $55/MWh as the acceptable threshold, and $58/MWh as the negotiating anchor. PG&E's 2024 RFO demonstrated that $52.3/MWh median is achievable today; a project with domestic content modules, pre-entitled land, and a confirmed CAISO interconnection queue position has negotiating leverage to exceed the median.

Recommendation 2: Pair Every Project with 4-Hour Battery Storage from Day 1

The 11.7/MWh storage premium in CPUC PPA contracts directly adds $10.8M/year to a 150 MW project's revenue versus solar-only at current scale. More importantly, storage repositions the project from a commodity solar generator (where curtailment is an unmanageable cost) to a dispatchable clean energy resource (where the project operator controls dispatch timing, captures evening price spikes, and avoids curtailment windows). The 9.4% and rising curtailment rate will erode solar-only economics year over year; storage is the structural hedge. The ITC applies to co-located storage at the same 40% rate under IRA 2022, making the incremental capital cost more manageable.

Recommendation 3: Prioritize FNO-001 for Phase 1 (West Fresno Flats, 150–250 MW)

Among the five satellite-identified sites, FNO-001 (West Fresno Flats) maximizes the composite of grid access (8 km to transmission), development risk (idle farmland, no endangered species triggers), and economics (score 92, GHI 5.95 kWh/m²/day). A developer entering the CAISO interconnection queue today at FNO-001 is 3–5 years ahead of greenfield sites requiring longer transmission builds. File the interconnection application immediately — the current CAISO queue backlog is 2–3 years, meaning a 2026 filing targets a 2028–2029 commercial operation date (COD), aligned with the 2030 SB-100 demand wave.

Recommendation 4: Qualify for Maximum ITC Stack — 50% is Available

The default 30% ITC base rate plus 10% Energy Community adder (Fresno County qualifies) yields 40% — worth $73.9 million. Adding Domestic Content qualification (requires ≥40% US-made components by value, readily achievable with First Solar or US-manufactured modules) pushes ITC to 50% — worth $92.4 million and reducing net investment to $101.3 million. This single compliance step adds $18.5 million to project economics — more value than a year of operating revenue at base PPA prices. Retain ITC counsel and a domestic content tracking consultant at project inception, not at commissioning. The documentation requirements are manageable but must be established from the supply chain selection stage.

Recommendation 5: Scale to 300+ MW to Break the CapEx Barrier

The IRR sensitivity heatmap reveals that at $900k/MW CapEx (vs. current $1.05M/MW), the 8% IRR threshold drops from $52/MWh to $44/MWh — within reach of current market pricing. Scale drives CapEx: a 300+ MW project enables module volume discounts (2–4% below market), EPC spread across larger fixed costs, and shared interconnection infrastructure. The FNO-002 Mendota Plains site (885 MW capacity, GHI 6.05 kWh/m²/day) could host a 300–500 MW Phase 2 expansion after FNO-001 de-risks development in the corridor. A combined 1,200 MW portfolio across FNO-001, FNO-002, and FNO-003 would represent a ~$1.5 billion investment — the scale at which institutional infrastructure funds and strategic acquirers like NEE or Brookfield become active buyers.

Recommendation 6: Monitor and Hedge Curtailment Risk

The rising curtailment trajectory — from 4% in 2020 to 9.4% in 2024 — is the stealth destroyer of solar economics in California. For every 1% increase in curtailment, a 150 MW project loses approximately $370,000/year in revenue at $38/MWh. By 2030, curtailment rates of 15% or higher are plausible without additional storage deployment. Contract language should include curtailment protection provisions — standard in California IOUs' recent RFOs — and all projects should be modeled with P90 curtailment assumptions (12–15%) rather than current-year averages. The CAISO Path 15 expansion (begun June 2025) provides partial relief for Central Valley generators, but structural curtailment management requires co-located storage.

Recommendation 7: Pursue FNO-004 Coalinga as a Differentiated Brownfield Play

The Coalinga site's retired oil field status and $4,500/acre land cost create a differentiated opportunity unavailable at standard agricultural sites. Brownfield ITC adders under IRS proposed rules would add an additional 10% to the ITC stack at this site — potentially reaching 60% ITC (50% base + domestic content + brownfield) worth $112M on a 150MW project basis. The highest irradiance in the county at 6.25 kWh/m²/day means every dollar of CapEx at Coalinga generates more electricity than at any other identified site. Designate FNO-004 as a Phase 3 asset developed after transmission access to the Coalinga area improves under CAISO's infrastructure roadmap.

11. Limitations and Risk Factors

No investment analysis is complete without a candid assessment of what the model does not capture and what could invalidate the conclusions. TMY3 Tule Fog Underestimation: The NREL NSRDB TMY3 dataset, while authoritative, represents a "typical" year that may underweight extreme fog years. Historical data from the San Joaquin Valley shows fog years occurring with approximately 30% frequency, reducing December–February production by 15–25%. In a bad fog year, annual production could be 3–5% below TMY3 estimates — equivalent to a $1–2/MWh effective LCOE increase. Project pro-formas should use P90 yield estimates from site-specific TMY3 plus fog frequency analysis, not median estimates. CAISO Curtailment Escalation: The Monte Carlo analysis models curtailment between 7% and 15%. The actual 2024 rate was already 9.4%, and the structural trend is upward. A 15% curtailment scenario in 2030 reduces project revenue by approximately $2.8M/year for a 150 MW project versus the 8% base case — enough to shift S4 from 7.8% IRR to approximately 5.5%, below most institutional hurdle rates. Interest Rate Sensitivity: The 10-year Treasury at 4.31% as of April 4, 2026 is historically elevated. Each 100 bps increase in rates adds approximately 0.60% to WACC and reduces project IRR by approximately 1.2 percentage points. Conversely, Fed rate cuts (futures markets are currently pricing 2–3 cuts in 2026) would improve project economics — but developers should not bank on rate cuts in project underwriting. PPA Market Liquidity: The CPUC contracting process for SB-100 compliance is real but competitive. The 48.8 GW gap to the 2030 target implies strong demand, but counterparty IOUs face their own financial pressures and will not overpay. A developer without a fully entitled site, confirmed interconnection queue position, and strong contractor relationships may not be able to compete for premium PPA contracts against established platforms like NextEra's development pipeline. Land Tenure and Agricultural Transition: The SGMA-driven fallowing of western San Joaquin Valley farmland creates land availability but also political complexity. Some landowners view solar leases as permanent conversion from agriculture — a position backed by environmental groups who prefer agricultural preservation even for low-productivity, water-stressed land. Permitting timelines of 24–48 months must be assumed for any project exceeding 50 MW under CEQA review. Module Supply Chain and Domestic Content: Qualifying for the Domestic Content ITC adder requires careful supply chain management. First Solar currently manufactures in Ohio, but module pricing for US-made products carries a 5–10% premium over Asian-manufactured alternatives. Developers must model the CapEx trade-off between higher US module cost and the $18.5M ITC value gain — at 150 MW scale, the math strongly favors domestic content, but the supply chain must be locked in before financial close.

12. Appendix

A. All Referenced URLs

• NREL NSRDB — National Solar Radiation Database
• NREL ATB 2025 — Annual Technology Baseline
• LBNL Tracking the Sun 2025
• CPUC Long-Term Procurement Planning
• CAISO Annual Report 2024
• IRS ITC — Energy Investment Tax Credit
• IRS Notice 2023-29 — Energy Community Adder
• IRS Notice 2023-38 — Domestic Content Adder
• IRA 2022 — Inflation Reduction Act
• California SB-100 — 100% Clean Energy by 2045
• Google Earth Engine — Satellite Analysis Platform
• yfinance — Live Market Data: NEE
• yfinance — Live Market Data: CWEN
• yfinance — Live Market Data: BEP
• yfinance — Live Market Data: FSLR
• yfinance — 10-Year Treasury Rate ^TNX
• California Energy Commission — Data and Reports
• CPUC — PPA Database
• CAISO — Infrastructure and Interconnection
• SGMA — Sustainable Groundwater Management Act

B. Geographic Scope

AOI Bounding Box [[[-120.0, 36.4], [-119.4, 36.4], [-119.4, 37.0], [-120.0, 37.0], [-120.0, 36.4]]]

C. Key Metrics Summary

Base IRR (S1, $38 PPA) 1.6% — DCF Model

Target IRR (S4, $55 PPA) 7.8% — DCF Model

D. Generated Visual Assets

1. chart_monthly_irradiance.png — Monthly GHI and AC yield profile
2. chart_dcf_cashflows.png — 35-year discounted cash flow, 4 scenarios
3. chart_irr_sensitivity_heatmap.png — IRR vs PPA and CapEx matrix
4. chart_monte_carlo_distribution.png — 6,000-run IRR distribution
5. chart_site_comparison.png — Top 5 sites multi-axis comparison
6. chart_comps_and_capital.png — Public comps and WACC decomposition
7. chart_california_market.png — California energy market 4-panel
8. chart1_land_cover_solar_zones.png — Land cover and development zones
9. chart2_financial_scenarios.png — Scenario comparison
10. chart3_sensitivity_returns.png — IRR sensitivity tornado chart
11. chart4_site_suitability_map.png — Composite suitability map
12. chart5_cost_waterfall_revenue.png — Cost waterfall and revenue bridge
13. fresno_solar_suitability_composite.png — GEE suitability composite (satellite)
14. fresno_high_suitability_zones.png — Filtered high-suitability zones
15. fresno_sentinel2_rgb.png — True-color satellite composite
16. fresno_sentinel2_nir_falsecolor.png — NIR false-color composite
17. fresno_ndvi_sentinel2.png — NDVI vegetation index
18. fresno_worldcover_landclass.png — ESA WorldCover classification
19. fresno_slope_map.png — SRTM terrain slope
20. fresno_bare_soil_index.png — SWIR-derived bare soil index
21. fresno_lst_summer.png — Summer land surface temperature
22. fresno_viirs_nightlights.png — VIIRS nighttime lights
23. fresno_srtm_elevation.png — SRTM elevation model

E. Methodology Summary

Irradiance: NREL NSRDB TMY3 Station 723890 — Fresno Yosemite International Airport reference station. Typical Meteorological Year representing long-term averages. Site Suitability: Google Earth Engine composite analysis. Sentinel-2 SR, SRTM, ESA WorldCover 2021, MODIS reflectance. Suitability = 0.30×(slope score) + 0.40×(land cover score) + 0.30×(inverse NDVI). High suitability threshold: 0.70. Area computed via GEE pixelArea() function. Financial Model: 35-year DCF with NREL ATB 2025 CapEx benchmarks, LBNL 2025 data, IRA 2022 ITC at 40%. WACC from CAPM with live Treasury rate. IRR computed via scipy.optimize.brentq root-finding. Monte Carlo: 6,000 runs, scipy.stats distributions. Market Data: CAISO Annual Report 2024, CPUC PPA Database, yfinance live as of April 4, 2026.

Report prepared April 4, 2026. All financial projections are analytical estimates based on publicly available data and should not be construed as investment advice. Past performance of comparable projects does not guarantee future results. Consult qualified legal, tax, and financial advisors before making investment decisions.