A solar farm is not financed on how much sun falls on the site. It is financed on how convincingly the contracts allocate risk when the plant produces less than the model said it would. That makes the EPC contractor a financing variable, not a procurement line item.
A solar farm is bankable when an Independent Engineer can confirm that its design, equipment, production forecast, and contracts hold up under third-party scrutiny, and when a single EPC wrap puts one creditworthy counterparty behind the schedule and the performance number. Lenders back that with liquidated damages, performance guarantees tested to ASTM E2848, retainage of 5 to 10 percent, performance bonds, and parent guarantees. The two risks that most often break a project are interconnection timing, where only 13 percent of capacity requested between 2000 and 2020 had reached commercial operation by the end of 2025, and a production model built on assumptions the measured data does not support.
What bankability actually means
Bankability is the condition in which a project's technical, contractual, and revenue assumptions survive third-party scrutiny well enough that a lender will lend against them on a non-recourse basis. It is not a quality rating. It is a question about who absorbs each risk if the plant underperforms.
The instrument that answers it is the Independent Engineer report. Commissioned by banks, developers, IPPs, and tax equity providers, an IE report is a due diligence review to help project financiers identify and mitigate risk. Its scope typically covers:
- Third-party review of electrical, structural, and civil design
- Equipment and supplier review, including modules, inverters, and racking, plus purchase orders
- Independent solar resource evaluation and kWh production forecasting, benchmarked against the developer's model
- Contract review across the PPA, the EPC agreement, the O&M contract, the interconnection agreement, net metering, and the site lease
- Site and regulatory review: topography, shading, zoning, code compliance, seismic, flood, and geotechnical conditions
Notice that the EPC agreement sits inside the IE's contract review. Your choice of contractor is not a procurement decision that happens after financing. It is an input to whether financing closes at all.
Lenders then wrap the whole thing in structure. Per Stoel Rives, construction lenders impose tight covenants restricting all activity other than development, require detailed progress reporting to the lenders and the independent engineer, take collateral assignment of equity interests, real property, project contracts, permits, and cash, and require consents or direct agreements from every material counterparty, typically including a bankruptcy replacement clause obligating that counterparty to enter a new agreement directly with the lender if the project company becomes insolvent. Distributions are gated by a debt service coverage ratio test, with 1.25 to 1.00 cited as an illustrative threshold. Miss it and cash is trapped rather than distributed.
The EPC wrap, and why lenders insist on it
A full-wrap or turnkey EPC agreement assigns substantially all project tasks to one entity, in contrast with design-bid-build using separate agreements. The bankability driver, in Akin Gump's phrasing, is single point of responsibility resting with the EPC contractor.
The reason is not administrative tidiness. Split contracts create interface gaps where no single party owns the schedule or the performance number. When the plant comes in 4 percent below the guaranteed capacity, a wrapped project has one counterparty to look to. An unwrapped project has an argument between a designer, an installer, and an equipment supplier, none of whom guaranteed the outcome.
The completion milestones are where careful reading pays. They are deal-specific and reputable firms define them differently:
| Milestone | Meaning |
|---|---|
| Mechanical Completion | Mechanically and electrically complete and capable of safe operation but for minor defects and omissions, and not yet interconnected |
| Substantial Completion | Commissioned, tested, and fully interconnected. Almost always a guaranteed date the contractor must meet |
| Final Completion | All punch list items complete |
Increasingly, owners and financiers require the contractor to guarantee the mechanical completion date as well, with associated delay liquidated damages, sometimes refundable if Substantial Completion is still achieved on time.
Liquidated damages and performance guarantees
Liquidated damages are how a schedule promise becomes an enforceable number. Typical market structures, per PwC and Wilson Sonsini:
- Delay liquidated damages capped at 10 to 15 percent of contract price
- Performance liquidated damages capped at 10 to 15 percent
- Combined delay plus performance cap of 20 to 25 percent
- Overall liability commonly capped at 100 percent of contract price
For a real executed example, the SunPower EPC contract template published by the Connecticut Green Bank sets delay LDs at $100.00 per MW of final project size per day, running from the required Substantial Completion Date, capped at 15 percent of contract sum, with aggregate liability capped at the contract sum except for gross negligence, fraud, or willful misconduct.
Performance guarantees are tested, not asserted. The governing standards are ASTM E2848, which regresses measured AC power against plane-of-array irradiance, ambient temperature, and wind speed at agreed reporting conditions, and ASTM E2939, which determines those reporting conditions and expected capacity.
Real numbers from a US utility procurement document, the Entergy Arkansas Solar BOT Scope Book, give a sense of the levels actually negotiated:
| Item | Level |
|---|---|
| Guaranteed Capacity Ratio | 97.0 percent |
| Minimum Guaranteed Capacity Ratio | 95 percent |
| Guaranteed Availability | 99.7 percent |
| Capacity test data requirement | At least 750 cumulative minutes across 3 or more separate days within a 15 consecutive day window |
| Availability test | Minimum 5 consecutive days |
PwC reports that the minimum performance ratio threshold for commercial operation is typically 95 to 98 percent of the guaranteed PR, with a defects liability period of 12 to 24 months after performance testing and acceptance.
Performance LDs are sized as net present value buy-downs rather than punishments. Wilson Sonsini's worked example on a 20 MW plant: a 1 percent PR shortfall costs roughly $50,000 per year, about $400,000 NPV, against a typical LD near $1 million. A 5 percent shortfall costs roughly $250,000 per year, about $2 million NPV, against a typical LD near $4 million.
Contractor credit: bonds, retainage, and the parent behind the signature
A liquidated damages clause is only worth the balance sheet standing behind it. That is why bankable EPC packages carry credit support:
- Parent company guarantee. Where the contracting entity is a thinly capitalized subsidiary, the parent typically guarantees performance. Note that this runs both ways. Contractors demand the same from owners, since the owner is usually a special purpose entity whose only asset is the project.
- Performance bond or bank guarantee. PwC reports typical sizing at 10 to 20 percent of contract price, varying with contractor identity, technology risk, and payment schedule timing.
- Retainage. Akin Gump puts it at typically 5 to 10 percent of each invoiced amount. PwC reports the same range, with a final payment holdback of 5 to 10 percent of contract price until performance tests complete, and notes retention is often replaced by a retention guarantee bond rather than holding both cash and bond. Stoel Rives describes holdback released at Substantial Completion, subject to a further punch-list holdback.
If a contractor fails mid-project, the lender protections above are precisely the response. The lender's collateral assignment of the EPC contract and its consent agreement with a bankruptcy replacement clause exist so the lender can step in and continue the contract. What no structure fully repairs is the schedule. Construction draws suspend pending IE sign-off, the project slips against the PPA outside date, and in 2026 it slips against a federal tax deadline as well.
The production gap nobody puts in the model
This is where diligence separates from optimism. Solar plants systematically produce less than their base cases.
- kWh Analytics' Solar Risk Assessment 2025 reports 8.6 percent nationwide underperformance below P50, drawn from more than 34,000 system-months of data between 2015 and 2023.
- Lawrence Berkeley National Laboratory's Utility-Scale Solar 2025 Edition reports a median year-one performance index of 93 percent across 342 plants of 2022 vintage, and notes the figure has been declining in recent vintages. Measured annual degradation is 1.6 percent, considerably steeper than the 0.5 percent often assumed.
- Availability assumptions are routinely overstated. The industry has historically modeled 99 percent lifetime availability. kWh Analytics measured mean availability of 96.4 to 97.2 percent across 165 systems and recommends modeling at 97 percent.
- Soiling costs at least 3 to 4 percent of annual PV production globally per IEA-PVPS Task 13, and in dry agricultural areas monthly soiling rates can range from 0.3 to 3.6 percent in winter up to 2.2 to 10.9 percent in summer.
Regional context matters enormously for a Northeast project. LBNL puts national average capacity factor near 25 percent, with the Northeast at roughly 17 percent, the lowest region in the country. A model built on national averages will overstate a Massachusetts plant materially before a single other assumption is tested.
The diligence question. Ask what performance index, degradation rate, and availability figure the pro forma uses. If the answers are 100 percent, 0.5 percent, and 99 percent, the model is describing a plant that does not exist in the measured data. A 2 point availability gap on a Northeast plant at 17 percent capacity factor is a permanent revenue haircut against the lender's base case.
Interconnection is the schedule
For utility-scale and larger community solar, interconnection is not a step in the project. It is the project.
LBNL's Queued Up 2026 Edition, covering data through the end of 2025, reports:
- More than 2,060 GW of generation and storage actively seeking interconnection, across roughly 8,200 projects
- Solar at 773 GW active, down 19 percent year over year, while natural gas rose 86 percent to 253 GW
- Median duration from interconnection request to commercial operation of more than 5 years for projects built in 2025
- Of capacity that requested interconnection between 2000 and 2020, only 13 percent reached commercial operation by the end of 2025, while 75 percent was withdrawn
- 549 GW already holds interconnection agreements without having reached commercial operation
That last figure is the one to sit with. An executed interconnection agreement is not a built project.
In New England specifically, ISO-NE moved from serial to cluster studies, with projects studied serially only prior to April 4, 2025. The Transitional Cluster Study launched October 20, 2025 with 26 interconnection requests totaling roughly 8 GW summer-rated, of which only about 350 MW is solar. Most requests are in Massachusetts. Modo Energy's July 2026 outlook puts the active ISO-NE generator queue at 127 projects and 14,541 MW, with only 4,395 MW forecast to reach commercial operation by 2029, and reports historical build-through rates by technology in which solar comes in around 50 percent, against 86 percent for gas and 78 percent for batteries. The next cluster request window runs October 5 to November 19, 2026.
None of this applies to a 500 kW rooftop or a small ground mount on a distribution circuit, which never enters the ISO-NE queue. Conflating the two is the most common error in solar marketing content. But for anything at solar farm scale, the queue is the critical path and the schedule risk is structural rather than contractor-specific.
Massachusetts revenue in 2026: SMART 3.0
Massachusetts remains one of the better solar revenue environments in the country, and the program is currently open.
SMART 3.0 is the active program. DOER began accepting Program Year 2026 applications on January 1, 2026, running through December 31, 2026, with 600 MW-AC of capacity for capacity-capped units. On July 8, 2026 the Department of Public Utilities approved company-specific SMART 3.0 tariffs, and DOER indicated it will begin issuing Final Statements of Qualification.
| System size (AC) | Base rate, $/kWh |
|---|---|
| Above 25 up to 250 kW | 0.2807 |
| Above 250 up to 500 kW | 0.2430 |
| Above 500 up to 1,000 kW | 0.2317 |
| Above 1,000 up to 5,000 kW | 0.1790 |
Two structural changes matter for underwriting. First, SMART moved away from the fixed declining block design toward a structure in which DOER annually adjusts rates, capacity allocations, and program features. That is more responsive but less predictable across a multi-year development timeline. Second, siting rules tightened: ground-mounted projects above 250 kW on land that is not previously developed are ineligible where they overlap BioMap Core Habitat or exceed 10 percent overlap with the highest forest-carbon areas, and eligible greenfield projects pay a land mitigation fee based on weighted environmental criteria and undergo environmental monitoring.
Adders remain the difference between a marginal project and a good one. They span location-based categories including brownfield, building-mounted, canopy, landfill, agricultural, floating, and raised racking, off-taker categories including community shared solar, low-income property, and public entity, and other categories including pollinator habitat and an energy storage multiplier.
The federal clock after OBBBA
This is the most time-sensitive section of this article, and the area most likely to have changed by the time you read it. Confirm everything below with tax counsel.
- Wind and solar facilities beginning construction after July 4, 2026 must be placed in service by December 31, 2027 to claim 45Y or 48E.
- Facilities that began construction before December 31, 2025 have a continuity safe harbor placed-in-service deadline of December 31, 2029.
- Facilities beginning construction between January 1 and July 4, 2026 have a placed-in-service deadline of December 31, 2030.
- IRS Notice 2025-42 eliminated the 5 percent cost safe harbor for all wind and for solar facilities exceeding 1.5 MW-AC nameplate capacity, beginning construction on or after September 2, 2025. Those projects must qualify under the Physical Work Test only. Facilities at or below 1.5 MW-AC may still use the 5 percent safe harbor.
- The 1.5 MW threshold carries an aggregation rule: facilities with integrated operations are measured collectively where they share the same owner, the same taxable year of placement in service, and the same interconnection point. That rule matters directly for phased community solar portfolios.
- FEOC material assistance rules apply to facilities beginning construction after December 31, 2025. IRS Notice 2026-15, released February 13, 2026, set the calculation as (A minus B) divided by A, where A is the direct cost of all equipment and B is the direct cost of prohibited foreign entity equipment, with three compliance routes: supplier certifications, the domestic content adjusted cost percentage tables from Notice 2025-08, or actual cost data. Steel and iron construction materials, main power transformers, and polysilicon are excluded from the calculation. For 48E qualified facilities beginning construction in 2026, the minimum non-PFE share is 40 percent, rising to 60 percent by 2030.
- The energy community bonus list was updated by IRS Notice 2026-39, released June 10, 2026.
The operational point for a developer is that "begin construction" is now a design decision, not a paperwork decision. The Physical Work Test turns on physical work of a significant nature judged by nature rather than cost, and installation of racking or structures to affix PV panels qualifies while planning, permitting, environmental studies, and site clearing do not. Which means the EPC contractor's ability to mobilize real work on a date certain has become a tax-credit-preservation function.
Land, O&M, and decommissioning
Land
LBNL's empirical study of 736 US plants found 2019 median array-area densities of 2.8 acres per MW-DC for fixed tilt and 4.2 acres per MW-DC for single-axis tracking. Those are array footprint figures, not total leased site area. NYSERDA's landowner guidance offers a practical gross site rule of thumb at roughly 6 acres per MW.
On lease structure, Penn State Extension reports an initial term of 25 years, aligned to module warranties, with typical 5 and 10 year extension options exercisable by the developer, preceded by a 2 to 5 year option or development agreement, and rent escalators commonly between 1 and 2.5 percent per year.
On per-acre rates, be careful with numbers you read online. There is no methodologically sound Northeast-specific published dataset. The best available survey evidence is the Purdue and CME Group Ag Economy Barometer, which found in May 2024 that more than 55 percent of respondents who had received solar lease offers were offered $1,000 per acre or more and 27 percent above $1,250, but that is Midwest data. Penn State Extension declines to publish a rate at all, noting the amount offered varies a lot. Treat any confident Massachusetts per-acre figure from a marketing page with suspicion.
Operations and maintenance
LBNL's Utility-Scale Solar 2025 Edition, drawing on 1,759 PV projects totaling 111 GW-AC, reports median empirical O&M falling from about $40 per kW-AC per year in 2012 to about $11 per kW-AC per year in 2024. NREL's Annual Technology Baseline 2024 models a fully loaded fixed O&M of $24 per kW-AC per year for a 100 MW-DC single-axis tracking system.
That roughly 2x gap is real and worth understanding rather than averaging away. LBNL captures reported spend, which may exclude some owner costs, property tax, insurance, and major-component reserves. NREL models a fully loaded figure. If your pro forma uses the LBNL number, confirm what it is not covering.
Decommissioning in Massachusetts
DOER's original model solar zoning bylaw required physical removal no more than 150 days after discontinued operations and did not mandate surety. DOER released an updated model bylaw in October 2025 that does include a financial surety requirement of 125 percent of a fully inclusive estimate of removal costs, with updated cost estimates required after 10 years and every 5 years thereafter. Adoption remains voluntary at the municipal level, and municipal jurisdiction extends to systems under 25 MW, with larger facilities falling to the Energy Facilities Siting Board.
For a sense of scale, a filed decommissioning plan for a 1.3 MW-DC project in Sturbridge, Massachusetts put net present value decommissioning cost at $124,500, roughly $96 per kW-DC, or $204,255 over 25 years at 2 percent inflation, with no salvage credit applied.
The investor and landowner checklist
- Full EPC wrap with a single point of responsibility, not a stack of separate agreements
- Written definitions of Mechanical, Substantial, and Final Completion, read rather than assumed
- Guaranteed Substantial Completion date with delay LDs, and ideally a guaranteed Mechanical Completion date
- Performance guarantee tested to ASTM E2848 and E2939, with capacity ratio and availability levels stated
- LD caps and overall liability cap disclosed, with the sublimits itemized
- Credit support: parent company guarantee, performance bond, and retainage of 5 to 10 percent
- Pro forma stress-tested at 93 percent year-one performance index, 1.6 percent degradation, and 97 percent availability
- Northeast capacity factor used, not the national average
- Interconnection position documented: queue, cluster, study status, and identified network upgrades
- SMART 3.0 application status and which adders the project qualifies for
- Documented begin-construction position under Notice 2025-42, including the 1.5 MW aggregation analysis
- FEOC material assistance analysis with supplier certifications retained
- Site lease term, escalator, extension options, and restoration language reviewed
- O&M scope and cost, benchmarked against both the reported and the fully loaded figures
- Decommissioning estimate prepared by a qualified engineer, with the municipal surety requirement confirmed
Why the EPC choice is a financing decision
Every item on that checklist eventually runs through one question: does the contractor who signed the wrap have the engineering, the crews, and the balance sheet to make the guarantees real?
Ferrius Energy handles engineering, procurement, and construction in house, under one contract, with no subcontractor chain to negotiate through when a schedule or a performance number is in question. Our project development team takes solar farms from site sourcing and feasibility through permitting, EPC construction, and long-term operations and maintenance, which means the same company that guaranteed the production number is the one monitoring it in year seven.
If you are a landowner evaluating a lease, an investor reviewing a development package, or a developer choosing an EPC for a bankable project in Massachusetts, New Hampshire, Connecticut, Rhode Island, Maine, Vermont, Florida, or Texas, we are glad to review the technical and contractual position with you before you commit.
Key takeaway. Bankability is not a property of the site. It is a property of the contract stack and the counterparty behind it. Stress the pro forma at a 93 percent year-one performance index, 1.6 percent degradation, and 97 percent availability, confirm the interconnection position in writing, and make sure the entity guaranteeing your production number is one that can still honor it in year ten.
Sources
- Partner Engineering and Science, What Is an Independent Engineering Report. partneresi.com
- Stoel Rives, The Law of Solar: Project Finance for Solar Projects. stoel.com
- Stoel Rives, The Law of Solar: Design, Engineering, Construction and Installation Agreements. stoel.com
- Akin Gump, A Legal and Commercial Checklist for Bankable Solar EPC Contracts. akingump.com
- PwC, EPC Contracts in the Solar Sector. pwc.com
- Wilson Sonsini, Hot Issues in PV Performance and Solar Project Finance. wsgr.com
- Connecticut Green Bank, SunPower EPC Contract Template. ctgreenbank.com
- Entergy Arkansas, Solar BOT Scope Book Appendix 7, Performance Test Procedure. entergyarkansas.com
- kWh Analytics, Solar Risk Assessment 2025. kwhanalytics.com
- Lawrence Berkeley National Laboratory, Utility-Scale Solar 2025 Edition. emp.lbl.gov
- IEA-PVPS Task 13, Soiling Losses. iea-pvps.org
- Lawrence Berkeley National Laboratory, Queued Up 2026 Edition. emp.lbl.gov
- Modo Energy, ISO-NE interconnection queue outlook July 2026. modoenergy.com
- Mass.gov, SMART 3.0 Program Details. mass.gov
- Gibson Dunn, Clean Energy Tax Provisions in the One Big Beautiful Bill Act. gibsondunn.com
- Holland & Knight, Beginning of Construction for Solar and Wind Facilities. hklaw.com
- Norton Rose Fulbright, New FEOC Guidance: Notice 2026-15. projectfinance.law
- Lawrence Berkeley National Laboratory, Land Requirements for Utility-Scale PV. emp.lbl.gov
- Penn State Extension, Pennsylvania Landowners Guide to Utility-Scale Solar Leasing. extension.psu.edu
- NREL Annual Technology Baseline 2024, Utility-Scale PV. nrel.gov
- DarrowEverett, Massachusetts Zoning Laws Solar Updates. darroweverett.com
- Town of Sturbridge, 200 Rt. 15 Sturbridge PV LLC Decommissioning Plan. sturbridge.gov
Published 2026-08-04 by Ferrius Energy LLC, a commercial solar EPC headquartered in Saugus, Massachusetts, serving MA, NH, CT, RI, ME, VT, FL, and TX. Codes, incentive programs, and federal tax rules change. Figures cited above were current at the date of publication. This article is general information, not legal, tax, or engineering advice for a specific project. Confirm current requirements with your own counsel, tax adviser, and authority having jurisdiction.