The Basics of Project Finance
Hello everyone,
There’s not a lot of information explaining project finance, especially at a rudimentary level, so I thought I’d write something up.
Project finance is a form of non-recourse financing, meaning essentially all of the collateral sits on the project being constructed rather than the broader sponsor. This is done through the creation of a special purpose vehicle (SPV). Let’s say it’s a solar farm project called Blue Solar. The sponsor may create an SPV called Blue Solar LLC. Since it’s a non-recourse financing, a bank generally cannot go after the assets of the sponsor, only those of Blue Solar LLC and the contracts tied to the project.
Prior to the project, like any other acquisition or development, the sponsor and advisors will determine the optimal mix of debt and equity to fund the project. Unlike companies with actual operating history, many project finance transactions are greenfield projects, meaning they are built 100% from scratch. As a result, there are no historical income statements or cash flows to underwrite. Instead, the revenues banks and sponsors use in evaluating the project are projections derived from contractual agreements, market assumptions, or bottoms-up operating models.
Before construction begins, there may already be some kind of revenue agreement depending on the project type. Energy projects often have power purchase agreements (PPAs) with another party called an offtaker, usually a utility or large tech/industrial firm, where the offtaker agrees to purchase power from the project over a long-term contract. Infrastructure projects may instead have tolling, leasing, concession, or availability-based agreements.
It is worth noting that project financings tend to have revenues that are more predictable than standard businesses since assets like toll roads, airports, or utility-scale solar farms operating under long-term agreements generally have more visible cash flows than a typical corporate business. As a result, debt-to-equity ratios are often more aggressive and debt pricing is generally lower than what you would typically see in standard corporate financings.
Once the optimal debt-to-equity ratio is determined, the actual structuring can occur. Let’s say Blue Solar is a $1 billion solar farm financed with 20% equity and 80% debt.
Two common ways equity can be injected into a project financing are fully upfront or pro rata. Fully upfront is more lender-friendly and, in the case of Blue Solar, the full $200 million of equity would be injected before meaningful debt funding occurs. Under pro rata equity funding, both debt and equity fund the project proportionately as costs occur. For instance, if there are $100 million in project costs incurred, $80 million would be funded with debt and the sponsor would inject $20 million in line with the target capital structure.
In a project finance deal, debt structures are very bespoke and are generally sized around a project’s cash flow available for debt service (CFADS). CFADS is broadly calculated as:
EBITDA – Cash Taxes +/- Change in Working Capital – Maintenance Capex
CFADS is used to calculate the debt service coverage ratio (DSCR), which is calculated as:
CFADS / Debt Service
For instance, if Blue Solar had CFADS of $120 million and debt service of $100 million in Year 1, then the DSCR would be 1.2x in Year 1.
The various tranches of a project finance transaction are then structured and sized around minimum coverage requirements. For instance, let’s say a bank financing Blue Solar is issuing a $200 million construction loan with a minimum DSCR requirement of 1.3x during the operating period. The repayment schedule and total debt sizing would need to be structured such that projected DSCRs do not fall below 1.3x over the amortization period under the lender’s base case assumptions.
DSCRs are also used in the process of “debt sculpting,” where the repayment schedule of each tranche is tailored around the project’s projected cash flows in order to maintain a target minimum DSCR throughout the life of the loan. Unlike a traditional corporate loan with a fixed repayment schedule, project finance debt amortization is often specifically shaped around the expected cash generation profile of the underlying asset.
When structuring a project finance transaction, lenders and sponsors utilize scenario analysis for different revenue and repayment scenarios, but it differs from the standard base/best/worst case framework often used in corporate financings. Instead, project financings frequently use a probabilistic downside analysis, or P-level analysis, in which probabilities are assigned to different operating outcomes.
This analysis is generally presented as P50/P90/P99. A P50 case means there is a 50% probability actual production or cash flow will exceed that level. A P90 case means there is a 90% probability actual production or cash flow will exceed that level, and so on. Higher P-levels therefore represent more conservative operating cases.
In renewable energy projects, these P-levels are typically tied to actual energy generation. For instance, Blue Solar may have:
• P50 energy generation of 500 MWh (megawatt-hours)
• P90 generation of 490 MWh
• P99 generation of 450 MWh
P50 can generally be thought of as the base case, while P99 represents an extreme downside case.
One quick way to assess a project’s operating riskiness is the spread between P50 and P99 generation. A tighter spread implies more predictable generation and lower variability, while a wider spread suggests greater uncertainty in project performance.
Projected generation is then used to determine revenue either through contracted prices under a power purchase agreement (PPA) or through prevailing market electricity prices in merchant exposure scenarios. Renewable deals typically have two types of revenue: contracted revenue and merchant revenue. Contracted revenue comes from agreed energy purchases under a PPA, while merchant revenue is based on prevailing market electricity prices. For example, if Blue Solar has 80% contracted revenue and 20% merchant exposure, then 80% of generation is sold at the fixed PPA price and 20% is sold at market prices.
Infrastructure projects are very similar in their use of P-level analysis; however, rather than being tied to energy generation, the scenarios are typically tied to utilization metrics such as traffic volume, passenger counts, throughput, or usage rates. For instance, in a toll road project, the P50 case may represent expected annual vehicle traffic, while the P90 or P99 cases reflect downside traffic scenarios. These assumptions are then used to project toll revenue and determine the project’s ability to service debt under different operating conditions.
In infrastructure projects, revenue is also typically driven by either explicit contractual frameworks or quasi-contracted demand structures, similar in concept to a power purchase agreement (PPA) in renewables. Instead of PPAs, infrastructure assets often rely on tolling agreements, concession agreements, availability payments, or ticket-based pricing mechanisms depending on the asset class.
For example, a toll road or bridge may operate under a concession agreement where the sponsor is granted the right to operate the asset and collect tolls over a fixed concession period. In some cases, toll rates are regulated or pre-agreed within a contract framework, which reduces pricing risk and makes cash flows more predictable. In other structures, demand risk is fully absorbed by the project company, meaning revenues fluctuate directly with traffic volumes.
Similarly, airports, rail systems, and ports may generate revenue through usage-based charges (tickets, landing fees, container handling fees) and contractual arrangements with governments or anchor users. Some infrastructure assets also use availability-based payment structures, where the government or counterparty pays a fixed fee as long as the asset meets performance or service-level requirements, effectively shifting demand risk away from the project.
Because of these contractual and semi-contracted structures, infrastructure cash flows tend to be more stable than typical corporate operating businesses, which is why lenders are often willing to apply higher leverage and rely heavily on P-level downside cases when sizing debt and structuring amortization profiles.
Finally, once the debt structure is set and cash flows are modeled through the P-level scenarios, the same framework is used to determine equity returns and overall project IRRs.
From the sponsor’s perspective, the key output of the model is not just DSCR compliance, but the equity cash flows after debt service. After CFADS is allocated to debt repayment based on the sculpted amortization schedule, the remaining cash flows are distributed to equity investors. These distributions typically occur after construction is complete and the project reaches operations, though in some structures partial distributions may occur earlier depending on cash sweep mechanics and reserve requirements.
Equity returns are then evaluated using standard IRR and multiple-based metrics. The equity IRR is calculated based on the timing and magnitude of all equity contributions versus all post-debt-service distributions back to equity holders. Because project finance structures are highly levered and cash flows are relatively stable under contracted or semi-contracted frameworks, equity IRRs are highly sensitive to leverage levels and the conservatism of the underlying P-level case used in sizing debt.
In practice, sponsors typically run equity returns across:
• P50 case (base case)
• P90 case (downside case)
• Stress cases (lower prices, higher capex, construction delays)
This allows sponsors to evaluate downside protection and upside leverage effects, while lenders remain focused on ensuring debt service is covered even under conservative scenarios.
In short, project finance is a balancing act between maximizing leverage while maintaining DSCR discipline, and generating acceptable equity returns under conservative operating assumptions.
Really useful, thanks for this
too basic, pump GPT and Claude and post a cvasi-whitepaper on Tax Equity... will give SB...
Plz add tax equity, carbon credits and merchant risk modeling. Otherwise, thanks for your contribution.
Can PF bankers move to “traditional” advisory roles. With that in mind is it limited to energy/ infra or coukd you move to unrelated sectors?
good stuff!!
Multiple comments have asked about tax equity and credits so I thought I'd write about that too.
Tax Equity
Renewable projects are unique from other projects in that they carry monetizable attributes beyond just capacity and usage. These attributes are the ITC/PTC tax credits, Renewable Energy Certificates (RECs), and MACRS (Modified Accelerated Cost Recovery System) depreciation along with bonus depreciation. The problem is that while renewable projects produce a wide variety of tax benefits, sponsors typically hold them in SPV (Special Purpose Vehicle) structures with little to no taxable income, so the sponsor itself is generally incapable of monetizing those benefits. To solve this, corporations with large tax bills can be structured into the project financing as "tax equity investors," meaning investors in the equity of the project who take on the tax benefits the sponsor cannot use. Traditionally these were the more financially sophisticated institutions such as banks and insurance companies, but more recent policy changes, covered later in the section on the IRA and OBBBA, have broadened the set of parties able to absorb these benefits to a wider range of large corporations. There are three tax equity structures used to move tax benefits to a party that can monetize them: the partnership flip, sale-leaseback, and inverted lease.
The Partnership Flip
The partnership flip is the most popular tax equity structure, especially for the PTC (and for the ITC where the sponsor wants to retain ownership and upside). The sponsor and the tax equity investor form an SPV taxed as a partnership with membership split into two classes: Class A and Class B. Under Class A, the Tax Equity Investor (TEI) contributes the majority of equity to the project (can be 35-60% of capital cost, sized to the present value of the tax benefits it will receive). In the pre-flip period, Class A is allocated ~99% of taxable income, loss, and tax credits and a defined share of cash.
Class B, the Sponsor, holds the operational and development interest, usually retains a larger share of distributable cash even pre-flip, and is allocated ~1% of tax items pre-flip. The "flip" is an event that occurs when the TEI achieves its targeted after-tax IRR (a yield-based flip) or a fixed date (a time-based flip, which is riskier for the TEI). Upon the flip, Class A drops to ~5% of tax items and cash, and Class B (sponsor) rises to ~95%. The sponsor typically holds a fair-market value purchase option to buy out the TEI's residual interest after the flip.
The key to why this structure works is that cash and tax benefits are allocated on different schedules. The sponsor keeps most of the operating cash to fund its return and service any back-leverage, while the TEI takes most of the tax items, which is all it actually wants. That split is the whole point of the flip.
Sale-Leaseback
Used almost exclusively for the ITC (and historically common in solar and in storage). The developer sells the placed-in-service project to the TEI (the lessor), which then leases it back to the developer/sponsor (the lessee) under a long-term lease. The TEI as owner claims the ITC and the depreciation; the sponsor operates the asset, pays rent, and retains the upside via a purchase option at lease end. Constraints: Must be executed within 3 months (90 days) of placed-in-service to keep ITC eligibility with the new owner. Provides 100% financing of fair market value to the developer (it monetizes the whole asset, not just the tax benefits), attractive when the developer wants maximum upfront cash. The lessee can elect to not reduce basis by passing through, but more importantly the structure cleanly assigns both credit and depreciation to the lessor. Must respect true-lease characterization (genuine residual value to lessor, no bargain purchase option) to avoid being recharacterized as a financing, and must avoid §470 loss-deferral and tax-exempt-use traps.
Inverted Lease (Lease Pass-Through)
The most technical of the three, used when the parties want to split the ITC from the depreciation. The owner (often sponsor-controlled, the lessor) leases the project to a master tenant (often TEI-controlled, the lessee) and elects to pass the ITC through to the lessee under the §50(d) lease pass-through election. Result: the lessee (TEI) gets the ITC, and the lessor (sponsor) keeps the depreciation. It lets a developer that can use depreciation but cannot efficiently use the credit (or vice versa) split the two attributes between the best-suited holders, and it requires less TEI capital than a sale-leaseback because the TEI isn't buying the whole asset, only the credit stream. The trade-off: the lessee must pick up §50(d) income (a phantom income inclusion roughly equal to 50% of the ITC, recognized over the recovery period), and the structure is operationally and legally intricate (master lease, true-lease tests, recapture coordination). It saw heavy use in solar and historic-tax-credit deals; in renewables it's a niche but valuable tool.
Renewable Energy Certificates
Energy produced by nuclear power plants, renewables, and gas-fired plants is physically indistinguishable once part of the grid. As a result, in 1999 the Texas state government passed Senate Bill 7, which established the framework for renewable energy credits, a digital certificate associated with MWhs of metered renewable energy output, with 1 REC = 1 MWh of metered output. RECs are not given as a fixed amount upon construction of a renewable project, but rather as a continuous stream tied to generation over the course of the project's life. Within Senate Bill 7 the Texas state government created one of the earliest and most influential RPS (Renewable Portfolio Standard) programs in the country, which is a requirement for LSEs (Load-Serving Entities), meaning utilities and retail energy sellers, to procure a minimum amount of their energy from renewables (as measured by the usage/retirement of RECs) or face penalties. RECs are transferable, meaning they can be purchased and sold. There are two types of markets for RECs: voluntary and compliance markets. Compliance markets are markets where there is a specific RPS, so LSEs may purchase RECs to avoid penalties. Voluntary markets are markets where purchases aren't driven by a compliance mandate, but RECs are still bought and sold here. The reasoning behind purchasing RECs in voluntary markets generally revolves around meeting ESG/emission goals set by boards and stakeholders. RECs exist as an additional source of monetization within renewable projects. RECs are included in PPAs (usually PPAs with corporate offtakers, since utilities are less concerned with ESG branding, although they may still need them if operating within a compliance market) in one of two ways: bundled or unbundled. Under a bundled PPA the offtaker purchases both RECs and energy, so within the model revenue is captured with one all-in PPA price (energy + RECs): Energy revenue (t) = net MWh delivered (t) × all-in PPA price (t). If a PPA is unbundled, the PPA sells brown energy and monetizes the RECs separately. This can be confusing, as it begs the question of why the energy is considered brown energy if it's generated by a renewable plant. As mentioned earlier, all energy within a grid is physically indistinguishable, so when energy is stripped of the associated RECs it becomes brown energy. Modeling revenue with an unbundled PPA results in separate lines for the RECs and the generated energy, with the energy revenue calculated as Energy revenue (t) = net MWh delivered (t) × power price (t) and the REC revenue calculated as REC revenue (t) = RECs generated (t) × REC price (t), where RECs generated (t) = net MWh (t) [1 REC = 1 MWh].
It's worth noting that bundled vs. unbundled is really a contracting concept. It's just whether a PPA staples the energy and the RECs into one price or splits them apart, and it sits alongside a separate, risk-based concept: contracted vs. merchant. A contracted (PPA) revenue stream is bankable, since the price is fixed and lenders will size debt against it under a DSCR (Debt Service Coverage Ratio) test. A merchant stream is uncontracted, sold into the wholesale market at the prevailing price, and because that price is volatile it gets heavily haircut or excluded from the bankable case and instead lives in the equity/sponsor upside. This is a big part of why projects chase PPAs in the first place. Framed this way, merchant revenue is structurally always unbundled. Bundling needs a PPA to do the combining, and a merchant project has no PPA, so the energy and the RECs simply default to being sold separately into their own markets. A fully merchant plant therefore earns from two venues at once and is modeled with the same two-line structure as an unbundled PPA above, except neither price is contracted. The energy sells into the RTO/ISO at a market power curve rather than a fixed price, and the RECs sell into the REC market at a separate REC curve. It's the same single generation forecast priced twice, because merchant energy and merchant RECs clear in two physically different markets. So all of these cases come back to the same two questions, how many prices are attached to the megawatt-hour and whether those prices are locked in, with the one recurring modeling error being the temptation to add a separate REC line on top of a bundled PPA price that already includes the REC value.
Carbon Credits
Since the topic comes up alongside RECs, it's worth drawing the distinction clearly, because the two get conflated constantly. A carbon credit (or offset) is a certificate representing one metric tonne of CO2 equivalent that was reduced, avoided, or removed from the atmosphere. So where a REC is measured in MWhs and certifies that clean energy was produced, a carbon credit is measured in tonnes and certifies that an emission was prevented or pulled out. Different unit, different thing being proven. And neither is a tax credit, so carbon credits have nothing to do with the ITC/PTC or the tax equity structures above. Like RECs, they trade in two market types: compliance markets, which are cap-and-trade systems where regulated emitters must hold allowances for what they emit (the EU ETS, California, RGGI), and voluntary markets, where companies buy offsets to meet net-zero or ESG claims without being legally required to.
The reason carbon credits don't get much attention in renewable project finance is that a typical grid-connected wind, solar, or BESS project doesn't generate them. Its environmental benefit is already captured as a REC, and the two systems are deliberately kept separate to avoid double counting the same megawatt-hour. Carbon credits instead come from very specific project types whose entire product is the avoided or removed tonne: capturing methane off a landfill or a dairy, reforestation and improved forest management, direct air capture, replacing dirty cookstoves, and destroying high-potency industrial gases. None of those is a power plant. So for the purposes of this guide, carbon credits sit in a separate asset class with their own developers, registries (Verra, Gold Standard), and buyers, and they aren't part of the wind, solar, and storage revenue stack.
Federal Tax Credits: ITC and PTC
There are two federal tax credits a renewable project can earn, and a project generally elects one or the other, not both. The choice comes down to the project's own economics.
The Investment Tax Credit (ITC) is a one-time credit equal to a percentage of the project's eligible basis, which is roughly the depreciable cost of the energy property, and it's earned when the project is placed in service. Because it's based on what you spent rather than what you produce, the ITC rewards deploying capital and is indifferent to how the project actually performs. It's claimed under §48, or the technology-neutral §48E for projects placed in service after 2024. The ITC tends to win for high-capex or lower-capacity-factor projects, and it's the practical route for standalone storage, which is also ITC-eligible.
The Production Tax Credit (PTC) is a per-kWh credit earned on actual generation over the first 10 years of operation, and it's inflation-adjusted. Because it scales with output, the PTC rewards running the machine and penalizes underperformance. It's claimed under §45, or the technology-neutral §45Y. The PTC tends to win for high-capacity-factor resources like wind and strong-site solar, where there's a lot of generation to earn the credit on. So the election is an economics decision made per project. A high capacity factor and low capex per MW points toward the PTC, while a low capacity factor or eligibility for large ITC adders points toward the ITC.
The credit rates follow a two-tier structure: a low base rate and a bonus rate worth five times the base, which a project only earns if it meets Prevailing Wage and Apprenticeship (PWA) requirements. For the ITC that's a 6% base rate versus 30% with PWA. For the PTC it's roughly 0.3 cents per kWh base versus 1.5 cents with PWA in statutory terms, which inflation-adjusts to around 2.75 cents in recent years. PWA means paying laborers the prevailing wage (the Davis-Bacon wage determinations) on construction and for the repair and alteration period afterward, and using a required percentage of qualified apprentice labor hours. Projects under 1 MW are exempt and get the bonus rate automatically. The reason this matters so much is that missing PWA drops the project all the way to the base floor, roughly an 80% haircut to the credit, so PWA compliance is a core diligence and indemnity item rather than a box-check.
On top of the 30% ITC, a project can stack three adders:
Domestic content (+10 percentage points) for using enough US-produced steel and iron and a rising required percentage of US-manufactured components.
Energy community (+10 percentage points) for siting the project in a brownfield, an area with historical fossil-fuel employment, or a census tract where a coal mine or coal plant has closed.
Low-income community bonus (+10 or +20 percentage points) under §48E(h), which is allocation-based and capacity-limited, so unlike the other two it isn't automatic and has to be applied for and awarded.
Stacked, a project can reach a 50% ITC, or higher if it secures the low-income allocation. On a project with $100mm of eligible basis, a 50% ITC is $50mm of dollar-for-dollar tax value before depreciation is even counted, which is exactly the kind of benefit a sponsor with no tax bill needs a tax equity investor to monetize.
A few mechanics of the ITC are worth flagging here because they drive the structures above. The depreciable basis is reduced by 50% of the ITC, so a project taking a 30% ITC only depreciates 85% of its basis. The ITC also vests over five years at 20% per year, and if the asset is sold or the TEI's interest is reduced during that window, a portion of the credit is recaptured, which is why flip timing and change-of-control restrictions are written so carefully. Finally, at-risk and passive activity rules limit who can actually use these credits, which is the underlying reason the buyers have always been large corporations with active tax liability.
Depreciation
Depreciation is the other major tax benefit a renewable project produces, and it tends to be roughly half of the total tax value alongside the credit itself, so it's not a footnote.
Most renewable energy property qualifies for 5-year MACRS (Modified Accelerated Cost Recovery System), a 200% declining balance schedule. What that means in practice is the owner gets to deduct the cost of the asset over five years for tax purposes, even though the equipment physically lasts 25 to 35 years. Because the schedule is declining balance, the deductions are front-loaded into the early years, which pulls the tax shield forward and makes it more valuable in present-value terms.
The amount you depreciate is the project's eligible basis, with one adjustment tied back to the ITC. If the project takes the ITC, the depreciable basis is reduced by 50% of the credit, so a project taking a 30% ITC depreciates 85% of its basis. If the project takes the PTC instead, there is no basis reduction, so it depreciates the full basis.
On top of MACRS sits bonus depreciation, which lets the owner immediately expense a percentage of the depreciable basis in year one rather than spreading it across the five-year schedule. That percentage currently sits at 100%, meaning the entire depreciable basis can be written off in the first year. So a project taking a 30% ITC, with its basis reduced to 85%, can deduct that full 85% of basis in year one.
The reason this matters so much is the size and timing of the resulting tax shield. A deduction is worth the deduction amount times the tax rate, so at the 21% corporate rate, $85mm of year-one depreciation is roughly $17.85mm of tax savings. Stack that on top of the credit and the first-year tax benefit can actually exceed the equity check written into the project. That is precisely the mismatch tax equity exists to solve: a sponsor with no tax bill cannot use a benefit that large in year one, so it brings in an investor who can.
The IRA and the One Big Beautiful Bill
Everything in the credit and depreciation sections above describes the rules in their current form, but that form is recent, and it's worth knowing where it came from and where it's heading, because two laws drive almost all of it.
The Inflation Reduction Act of 2022 (the IRA) is what created most of the credit structure described above. It extended and expanded the credits, and for projects placed in service after 2024 it replaced the old technology-specific §45 and §48 with the technology-neutral §45Y and §48E, meaning eligibility now turns on whether a facility is zero-emissions rather than whether it appears on a specific list. The two-tier base-versus-PWA rate and the domestic content, energy community, and low-income adders all came from the IRA. It also made standalone storage ITC-eligible, which is a big reason the battery buildout took off.
The most important thing the IRA did for monetization, though, was create two new ways to get the tax benefits out of a project without a traditional tax equity partnership. The first is transferability under §6418, which lets a project simply sell its credits to an unrelated third party for cash. The cash isn't taxable to the seller or deductible to the buyer, and the credit can only be transferred once. This is the change referenced earlier that broadened the pool of parties who can monetize: before transferability you needed a structured partnership and a narrow set of banks and insurers with the sophistication to do these deals, but now any corporation with a tax bill can just buy credits for cash, usually at something like 90 to 95 cents on the dollar. The catch is that you can't transfer depreciation, only the credit, so a pure transfer leaves the MACRS benefit stranded. That's exactly why tax equity didn't disappear and why hybrid structures exist, running a partnership to place the depreciation while transferring the credit out for cash. The second new channel is direct pay under §6417, which lets tax-exempt and government owners, meaning nonprofits, municipalities, tribes, and co-ops, take the credit as a cash refund directly from the IRS, since they have no tax bill to offset in the first place. For-profit sponsors generally can't use direct pay except for a few specific credits.
The One Big Beautiful Bill Act of 2025 (OBBBA), signed in July 2025, then recalibrated all of this, and it cuts in two directions. On the restrictive side, it sharply accelerated the phase-out of the wind and solar credits specifically. A wind or solar project now has to be placed in service by the end of 2027 to qualify, unless it begins construction within twelve months of enactment, meaning by around July 2026, in which case it falls under the normal continuity rules instead of the hard deadline. The practical effect is a race to start construction on wind and solar projects before mid-2026. Other zero-emissions technologies, including storage, geothermal, nuclear, and hydro, keep the longer original phase-out timeline, so the bill deliberately treats dispatchable and storage resources more favorably than wind and solar. OBBBA also layered on foreign-entity restrictions, often referred to as the FEOC (Foreign Entity of Concern) or prohibited foreign entity rules, which can deny credits if a project has prohibited foreign ownership or influence, or if too much of its components and materials come from prohibited foreign sources, which in practice mostly means Chinese supply chains. These apply to construction beginning after 2025 and have made supply-chain provenance a core diligence item rather than an afterthought.
On the other side, and crucially for everything above, OBBBA preserved the monetization machinery. Transferability under §6418 survived, so the credit-transfer market the industry had just rebuilt around stays intact, and the bill also made 100% bonus depreciation permanent, which is the reason the depreciation section could say the bonus rate currently sits at 100%. So the net picture is that the IRA expanded the credits and opened up transferability and direct pay, while OBBBA compressed the wind and solar window and added supply-chain conditions but kept the plumbing that lets projects actually monetize. For deals being done today that means a scramble to begin construction on wind and solar before mid-2026, a real premium on clean non-Chinese supply chains, a relative tailwind for storage, geothermal, and nuclear, and credit transfer sitting alongside tax equity as a default way to get the benefits to a buyer.
One caveat worth stating plainly: the OBBBA beginning-of-construction and foreign-entity rules are still being worked out in Treasury and IRS guidance, so the precise thresholds and dates are the kind of thing to confirm against the live rules before relying on them in an actual deal.
Merchant Risk Modeling
Since merchant revenue keeps coming up, it's worth walking through how it actually gets handled in a model, because it's the piece that drives how much debt a project can carry. Contracted revenue under a PPA is easy: the price is fixed, so you plug it in and it's bankable. Merchant revenue is the hard part, because both the price and the volume are uncertain, and lenders won't lend against uncertainty the way they lend against a contract.
On the price side, merchant energy is modeled off a forward power price curve, usually one bought from a third-party market advisor. That curve isn't a flat number. It has to be adjusted for basis, the locational difference between the trading hub and the project's actual interconnection node, and it has to be shaped, meaning weighted to capture the price at the hours the project actually produces rather than a flat average. This matters most for solar, which produces in the middle of the day, exactly when a lot of other solar is also producing and pushing the midday price down. That effect is called cannibalization, and it's why a solar project's captured price is usually lower than the average market price.
On the volume side, the same uncertainty shows up through the production estimate. Projects are modeled across production scenarios, typically a P50 (the expected case, with a 50% chance of being exceeded) and a P90 (a conservative case, with a 90% chance of being exceeded), and the P90 is used to stress the downside.
The reason all of this matters is debt sizing. Lenders heavily haircut or fully exclude uncontracted merchant revenue from the bankable case, sizing debt to a conservative price and production scenario, so a fully merchant project supports far less leverage than a contracted one. This is why sponsors often bolt on a hedge or a price floor (such as a financial swap or a revenue put) to convert volatile merchant revenue into something that looks contracted enough for lenders to credit. The merchant upside still exists, but it lives in the equity case rather than the debt case, which ties back to why projects use PPAs in the first place.
Back-Leverage
Back-leverage is the debt in a tax equity deal, but it's structured in a way that trips people up the first time they see it, so it's worth explaining what it actually is before getting into how it's sized.
In a normal project financing, the debt sits at the project level and is secured by the project itself, meaning the lender has a lien on the assets and can foreclose on them if the borrower defaults. In a tax equity deal you can't do that, because the tax equity investor is already sitting at the asset level and will not allow a lender to take a security interest in the project. There are two reasons. First, the TEI put in the bulk of the capital to get the tax benefits and needs to be the senior party at the asset level, not sitting behind a lender. Second, and more importantly, a lien on the project assets is dangerous: if a lender ever foreclosed and the project changed hands during the five-year ITC recapture window, it could trigger recapture and claw back a chunk of the credit the entire deal was built around. So traditional asset-level debt simply doesn't fit.
The solution is to push the debt up a level. Instead of lending against the project, the lender lends to the sponsor's holding company that sits above the partnership, and takes security over the sponsor's equity interest and the cash that flows up to it, rather than over the turbines and panels themselves. That's what back-leverage means: the debt is levered against the sponsor's equity position behind the project, not against the project's assets. If the sponsor defaults, the lender takes over the sponsor's equity interest and steps into its shoes in the deal, but it never touches the assets directly and never disturbs the tax equity investor's position.
The reason this matters for everything that follows is that back-leverage is structurally junior to the tax equity at the asset level. The TEI gets its allocations and its share of cash first, and only the cash that's left and distributed up to the sponsor is available to service the back-leverage. So the loan isn't sized against the whole project's cash flow, it's sized against the sponsor's slice of it, which is a big reason it behaves differently from ordinary project debt and supports less leverage than an asset-level loan would. It also means the back-leverage lender and the tax equity investor have to agree up front, in a forbearance or intercreditor arrangement, on what the lender can and can't do on a default, since the lender's remedies have to stay clear of anything that would harm the TEI or trigger recapture.
Modeling and Sizing
Everything in this guide eventually feeds one of two engines in the model, and the whole exercise is figuring out how big each piece of the capital stack can be. The operating engine builds the cash flows from energy, capacity, and RECs, and that sizes the debt. The tax engine builds the credits and depreciation, and that sizes the tax equity. The two never really touch each other directly, but they meet in the sources and uses that fund construction and again in the sponsor's after-tax return at the end. Once you see the model that way, everything above slots into place.
Start with the operating side. CFADS, meaning cash flow available for debt service, is just project revenue minus operating expenses, and it's the cash the project actually has to pay lenders with. Revenue is the energy, the capacity, and the RECs, so this is where the RECs and the contracting choices from earlier actually bite. What matters for sizing is not how much CFADS the project produces in the expected case, but how much of it a lender is willing to count, and that depends entirely on whether the revenue is contracted or merchant. Contracted revenue under a PPA is bankable, so it gets credited in full. Merchant revenue is volatile, so it gets heavily haircut or excluded. The same logic applies to the RECs specifically: if they're bundled into a PPA they're already inside the contracted price and the lender credits them, but if they're sold merchant they mostly drop out of the bankable case and show up as equity upside instead. So the contracting decisions don't just change the revenue lines, they change how much of CFADS counts toward debt.
That bankable CFADS is what sizes the back-leverage debt. The debt is sized so that CFADS covers debt service by a target cushion, the DSCR, across the life of the loan, so the lender essentially solves for the largest loan whose service the conservative-case CFADS can still cover at that coverage ratio. The more contracted and stable the revenue, the higher the CFADS the lender will credit, and the more debt the project can carry. The more merchant it is, the thinner the bankable CFADS and the smaller the loan, which is exactly why a contracted project supports more leverage than a merchant one and why sponsors bolt on hedges or floors to make merchant revenue look contracted enough to count. And remember this is back-leverage, so it's sized against the cash that actually flows up to the sponsor in distributions, not the whole project's cash flow, which is another reason a tax equity deal carries less debt than a simple asset-level financing would.
Now the tax side, which sizes the tax equity rather than the debt. The credits and depreciation don't service debt and they aren't CFADS, they're the benefits the tax equity investor is buying, and the size of the tax equity check is set at the present value of those benefits discounted at the TEI's target after-tax yield. That single sentence is the whole reason depreciation matters so much in the model. The bigger and faster the tax benefits, the bigger that present value, the bigger the tax equity contribution, and the less of everything else you need to fund the project. MACRS and 100% bonus depreciation front-load almost the entire depreciation shield into year one, which maximizes that present value, while the 50% ITC basis reduction trims it slightly. The ITC adds a one-time benefit earned at placed-in-service, and the PTC adds a ten-year stream earned on production, which is why a PTC deal's tax equity is partly modeled off the same generation forecast that drives revenue. The credit size itself is driven by everything in the credit section, the base versus PWA rate and the adders, so a project that secures PWA plus domestic content plus an energy community is generating a far larger benefit, which in turn supports a far larger tax equity check.
Here's where the timing problem comes in, and where the TEBL gets created. The tax equity is committed, but it doesn't fund when you need it. The investor's contribution is typically paid in at placed-in-service, because that's when the ITC is earned, and in a PTC deal it funds partly upfront and partly over the first several years as production comes in, which is the pay-go dynamic. But construction has to be paid for as it happens, so you have a large committed equity check sitting on the sidelines during the exact window you need cash to build. The tax equity bridge loan (TEBL) fills that gap. It's a construction-period facility advanced against the committed tax equity, it funds the build alongside the construction loan and the sponsor's equity, and it gets repaid in one shot when the tax equity investor funds at placed-in-service. If the deal monetizes its credits by selling them under a transfer rather than through a partnership, the same timing gap exists on the credit-sale cash, and a transfer bridge does the same job against the expected proceeds, repaid when the buyer pays. A deal can have a TEBL, a transfer bridge, or both, and the presence of both is the signature of a hybrid deal that runs a partnership for the depreciation while transferring the credits out for cash.
The TEBL is sized off the committed tax equity, but not at the full amount. The lender applies an advance rate that haircuts the commitment for two risks: the credit quality of the tax equity investor, meaning whether they will actually be good for the money, and the conditions precedent to their funding, meaning whether the project will actually hit the milestones that obligate them to fund. So the chain runs all the way back through the tax engine: the size of the tax benefits sets the size of the tax equity, and the size of the tax equity, after the advance-rate haircut, sets the size of the bridge that can carry the construction period. A bigger, cleaner tax equity commitment supports a bigger bridge and takes pressure off the construction loan and sponsor equity.
All of this resolves in the sources and uses. During construction the build is funded by the construction loan, the TEBL or transfer bridge, and sponsor equity. Then at placed-in-service the picture turns over: the tax equity funds and repays the TEBL, any credit sale closes and repays the transfer bridge, and the long-term back-leverage takes out the construction loan, leaving the standing stack of tax equity, back-leverage, and sponsor equity. This is where the two engines finally meet. The operating engine sized the debt through CFADS, the tax engine sized the tax equity through the target yield, and together they determine how much sponsor equity is actually needed to fill whatever gap is left. Whatever return that sponsor equity earns, after the TEI takes its share and flips out, is the after-tax IRR the whole model exists to solve for. From there every sensitivity traces back to the same handful of inputs: PWA and the adders move the credit and therefore the tax equity, the contracting mix moves bankable CFADS and therefore the debt, and the production case, P50 versus P90, moves both at once.
Animi alias cumque architecto illo. Illum nisi qui voluptatem iusto quibusdam. Quasi sit aliquid cum est qui ex.
Omnis libero dolores accusantium nihil velit. Velit molestiae nisi rerum nam nemo animi qui. Officiis sed doloremque necessitatibus repudiandae necessitatibus repudiandae. Qui commodi consequuntur doloribus voluptatem.
Ut et consequuntur eligendi odio atque. Sapiente dolor molestias excepturi autem quod cum. Illo facilis nihil illum laudantium animi consequatur omnis voluptas.
Quis atque autem adipisci culpa est. Hic voluptatibus et officia odit error tempore non. Cupiditate voluptas aliquid saepe est eius placeat ea.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...