Buy, Lease or PPA: How to Pay for Your Company's Solar Project

The system is the same in all three cases: the same modules on the same roof, producing the same energy. What changes is the title over that asset, and every other difference flows from there.

In an outright purchase the company pays for the project with its own cash and the system enters fixed assets from day one. It is the simplest structure to understand and the one that ties up the most capital: the whole outlay happens before the first kilowatt-hour of savings arrives.

Under financing —a bank loan, a financial lease or a development bank facility— the company still ends up owning the system, but pays in instalments. The project stops being a question of available cash and becomes a comparison between two numbers: the monthly instalment and the monthly saving on the bill.

Under a third-party contract —what the market calls a PPA, «solar as a service» or energy supply— the company buys nothing. A third party installs, operates and maintains the system at its own cost, and the company pays for the energy it consumes or for the service, over a long term. There is no upfront investment and there is no asset either.

Everything else —tax, balance sheet, who replaces a failed inverter, what happens if the warehouse is sold— follows from that single variable. That is why a serious comparison starts there and not with the price.

This is the point that distorts comparisons the most, because Colombia's incentives for non-conventional energy are substantial and whoever does not own the asset does not touch them.

Ley 1715 de 2014, in the articles amended by Ley 2099 de 2021, offers four benefits to whoever invests in non-conventional generation: deducting 50% of the total investment from income tax over a period of no more than fifteen years, starting the tax year after the system enters operation; depreciating equipment, machinery and civil works at an annual rate of up to 33.33%; acquiring the project's equipment and services free of VAT; and importing machinery not produced domestically free of duty.

There is one cap that decides more projects than it seems: the amount deducted in a year cannot exceed 50% of the taxpayer's net taxable income, determined before subtracting that deduction. A company with large taxable income exhausts the benefit quickly; one with thin profit, with losses, or under a special regime carries it for years or simply never uses it. When that is the case, the right structure may be exactly the opposite of what the installer recommends: let the incentive go to whoever can actually use it, and buy the energy instead.

The other requirement is procedural and allows no shortcuts: all four benefits require the UPME to evaluate and certify the project. It is not an afterthought or a form obtained once the invoices are in; the VAT exclusion and the duty exemption are requested before the purchase or the import. A project schedule that leaves no room for the UPME is a schedule that will lose benefits along the way.

Buying with own funds wins when three conditions hold at the same time, and it is worth being candid: they do not always hold.

The first is having enough taxable income to use the deduction within a reasonable period. The Ley 1715 benefit is the difference between a good project and an excellent one, and only a taxpayer collects it.

The second is that the opportunity cost of capital is lower than the project's return. Tying up cash in a solar plant makes sense if that cash does not earn more inside the business; in a growing company where every peso put into inventory or productive capacity earns more, the same project can be a poor capital allocation even if the system itself is profitable.

The third is having someone to run the operation for twenty years. Ownership means taking on maintenance, monitoring, mid-life inverter replacement and warranty management. None of that is complicated, but it has to be assigned: an owned system with no internal owner degrades quietly, and nobody notices until the bill stops falling.

In exchange, an outright purchase delivers the best long-term economics in almost every scenario, because it pays nobody's financial margin and because, once the investment is recovered, the energy is practically free for the rest of the system's useful life.

Financing the purchase keeps ownership and defers the outlay, which is exactly what a company with a profitable project and tight cash needs. But three details rarely make it into the sales deck.

The first is good news: the special income tax deduction is available through a financial lease. Article 1.2.1.18.72 of Decreto 1625 de 2016 provides for it, so the lessee can take the benefit even though the asset is formally held by the finance company.

The second is a condition to write into the contract and into the board minutes: the benefit is tied to exercising the purchase option. Article 1.2.1.18.77 of the same decree requires that fact to be verified, and not exercising it forces the deductions taken to be repaid. A lease you plan to hand back at the end is not the right structure if the incentive was the point.

The third is bad news worth knowing before the financial model is built: the DIAN has held that the VAT exclusion does not apply to acquisitions financed through a financial lease, on the grounds that tax benefits are interpreted restrictively (Oficio 906434 de 2022). It is administrative doctrine rather than statute, and its scope can be discussed with an adviser; but if the installer's model netted off that VAT and also proposed a lease, the model is showing a saving that may not exist.

There is a fourth, accounting point. The argument that a lease keeps the project «off balance sheet» stopped being true: since 1 January 2019 Group 1 companies apply NIIF 16, brought into the Colombian framework by Decreto 2170 de 2017, which requires the lessee to recognise a right-of-use asset and a lease liability. A leverage covenant being negotiated with a bank today will see that liability. That is not a problem in itself; it is a fact to bring into the conversation beforehand, not after.

In this structure the company does not buy a system: it buys energy, or a service, from whoever did buy it. The third party invests, installs, operates and maintains, and invoices over a long term —ten, fifteen or twenty years are common— at a price per kilowatt-hour, usually with an annual escalation formula.

The appeal is obvious: no upfront investment, technical risk transferred, and savings from month one if the contracted price sits below the tariff. So is the cost: the company does not own the asset, does not take the incentives, and is tied to a long contract whose exit clause is usually the most expensive part of the document.

The scheme is recognised in regulation. The CREG has stated that a self-generating user may or may not own the self-generation assets, and that the relationship between that user and the third party who installs and operates them is bilateral and not subject to regulation. That second part deserves a slow read: it means no rule protects the company against a badly drafted contract. What is not written in the contract does not exist.

There is one thing the contract cannot transfer. The party that applies Resolución CREG 174 de 2021 before the grid operator and commercialises the exported surplus is the self-generating user —that is, the company that consumes— not the third party. The contract should state in plain words who files the application, who answers if the grid operator objects, and who owns the revenue or the energy credit that the surplus generates.

Since 2024 a variant has emerged that changes the map for anyone without a suitable roof, or unwilling to touch it. Decreto 1403 de 2024 set out the policy for remote self-generation —producing in one place and consuming in another— and expressly established that the generation assets may be owned by the person or company or by third parties, and that a third party may operate them (art. 2.2.3.2.4.3, item 5). The CREG developed that framework with Resolución CREG 101 099 de 2026, published in March 2026, which subjects remote self-generation to connection and operating rules comparable to those of a conventional plant. It is a new regime: if you are offered a remote PPA, ask how that resolution is complied with and who bears the associated charges.

One fact about the company conditions the whole comparison, and many proposals never even ask for it: whether it is a regulated or an unregulated user.

The threshold was set by Resolución CREG 131 de 1998 and has applied since 1 January 2000: an unregulated user is one whose power demand exceeds 0.1 MW or whose monthly consumption exceeds 55 MWh at a single delivery point. Above that threshold the company negotiates the price of energy freely with retailers; below it, it pays the regulated tariff its retailer applies.

Why it matters here: the price a PPA is compared against is not the same in both cases. An unregulated user already negotiates its energy and probably has a contract in force, with a term and a price that must be put on the table before signing another one for fifteen years. A regulated user compares against a tariff that moves with inflation and with the market, not against a negotiated price.

If your company sits close to the threshold, the solar project may push it from one side to the other by reducing the energy drawn from the grid. It is a second-order consequence almost nobody models, and one to check with the retailer before sizing.

The most expensive mistake in this process is putting a purchase price, a lease instalment and a price per kilowatt-hour into one table and picking the prettiest number. They are neither the same unit nor cover the same thing.

The only valid comparison is after-tax cash flow, year by year, across the full life of the system —twenty or twenty-five years— under identical output assumptions for all three. If the installer cannot produce that table, they are selling equipment, not a project.

The first is treating the incentive as a given. Many financial models net off the full deduction in year one, without checking the company's taxable income or the 50% cap, and without having started the UPME process. A project that is only profitable under that assumption is not a profitable project.

The second is choosing the structure before knowing the load. The plant's load curve —how much energy is consumed and at what hours— determines how much of the generation is self-consumed, which is the part that avoids the full tariff. Without twelve months of bills analysed, any financial structure rests on a guess.

The third is signing a PPA without reading the termination clauses. That is the document where the provider protects its investment, and rightly so: it put up the capital. But a company that closes a plant, relocates or changes hands with a fifteen-year PPA in place finds out late that leaving costs more than it saved.

The fourth is forgetting that the roof ages too. If the roof needs work within five years, all three structures get complicated: under ownership it is an unbudgeted cost, and under a third-party contract it is a negotiation with someone whose asset sits on top of your warehouse. It is solved by inspecting the roof before signing, not after.

The fifth is keeping the finance and tax team out of it until the end. The difference between the three structures is decided in income tax, in NIIF 16 and in the ratios the company uses to negotiate its debt. The installer cannot answer for that, and should not be asked to.