How Colombian Solar Sells Power: The Spot Market, Bilateral Contracts, and the First 15-Year Auction
Executive summary
On July 29 and 30, 2026, Colombia's Ministry of Mines and Energy ran the country's first long-term contracting auction under Decree 1091 of 2025 and awarded 995 MW of solar capacity plus 100 MW of battery storage. The contracts run 15 years, begin delivering on January 1, 2030, and cleared at COP 315.87 per kWh, about 9.9 US cents. The ministry itself pointed out that the price is less than 30% of what the spot market can reach during a drought.
For a capital allocator, the auction is most useful as the third leg of a revenue structure that most diligence materials describe loosely or skip. A Colombian solar asset earns revenue through some combination of three routes: the spot market (the bolsa de energia), bilateral contracts, and now auctioned 15-year contracts backed by regulated demand. Each route has its own price, its own counterparty, and its own way of going wrong. Whether a project can carry debt depends more on its contract mix than on its irradiation study.
Below I go through the three routes, what the July auction awarded, and the questions the results should add to a diligence list.
A market priced by water
XM, the system and market operator, administers Colombia's wholesale electricity market. Roughly two-thirds of the country's generation is hydroelectric, and that one fact drives how power is priced. When reservoirs are full, spot prices are low. When an El Nino cycle cuts rainfall, prices climb: during the 2023 to 2024 El Nino, hourly prices on the bolsa ran above COP 1,000 per kWh for stretches of early 2024, several times their level in a normal wet year. When the ministry says the auction cleared below 30% of drought-period spot prices, it is publicly acknowledging that same spread.
A foreign investor needs this context before reading any Colombian revenue model. Spot price risk in Colombia comes mainly from hydrology, far more than from solar supply or demand. A megawatt hour sold without a contract is a position on rainfall, and it pays best in exactly the years nobody can plan for.
Route one: the spot market
Every generator in the wholesale market settles through the bolsa. Output not covered by a contract sells at the hourly spot price, whatever it happens to be. For a solar project this is the default, where a plant ends up when it has capacity and no offtake.
Spot revenue is real, and in drought years it is large. Lenders still will not finance against it. A lender sizing debt against Colombian generation revenue will not underwrite a price series that swings with reservoir levels, so merchant exposure serves as equity upside or as a bridge between contracts and never as the base of a capital structure. If a revenue model needs drought-level prices to clear its hurdle, it is really a weather forecast.
Route two: bilateral contracts
The bilateral contract is the market's main instrument. Colombian demand splits into two customer classes. Regulated users, meaning households and smaller businesses, buy through retailers (comercializadores) that procure energy under competitive rules set by CREG, the energy regulator. Non-regulated users, roughly those with demand above 100 kW, negotiate supply directly, and that is where corporate PPAs sit.
For project finance, the problem has been tenor. Colombian bilateral contracts have historically run short, commonly one to five years, because retailers buying for regulated demand had little reason to lock in longer. Long-tenor deals exist in the non-regulated segment (Scatec signed a 15-year PPA for a 130 MW plant), but a developer has to negotiate each one as an exception.
Set a three-year contract against a ten-year amortization schedule and the mismatch is obvious. The project will reprice its revenue three times before the debt retires, and someone carries the risk that recontracting happens at a lower price or not at all. The question I raised in the piece on Colombia's grid connection queue comes from exactly this situation: if the contracted tenor is shorter than the debt, who holds the merchant tail?
Route three: what the July auction awarded
Decree 1091 of 2025 authorized energy contracts of up to 15 years, and the Ministry of Mines and Energy convened the first mechanism under it in January 2026. Rules issued in April 2026 opened participation to standalone battery storage, extending the direction set by Resolution 40178. The awards came in two rounds. On July 29, 270 MW of solar and roughly 100 MW of batteries cleared, with Celsia, the power arm of Grupo Argos, taking the first firm battery contracts awarded in Colombia and Bosques Solares de los Llanos among the solar winners. On July 30, a further 725 MW of solar cleared.
In total, 995 MW of solar and 100 MW of storage were awarded, all on 15-year contracts beginning January 1, 2030, at a clearing price of COP 315.87 per kWh. The ministry estimates the awarded volume covers about 3% of regulated electricity demand.
I read the result two ways, and both hold. The first is what it signals. The Colombian market now has a COP-denominated, 15-year contract backed by regulated demand. Project finance has been asking for that tenor since the 2019 auction round, and with nearly 1.1 GW cleared at a price the ministry is willing to defend publicly, the route is now a working option.
The second is the volume. Three percent of regulated demand is small, and 1.1 GW awarded compares with an approved pipeline of roughly 13.5 GW. Most of the pipeline will still have to sell power through bilateral contracts or run merchant. The auction changes the diligence conversation for the projects that won, and the structural question stays open for everyone else.
The 2030 start date
This detail has had less attention than it deserves. Deliveries begin on January 1, 2030, roughly three and a half years after award. A winning project has to be financed and built against a contract that pays nothing until then. A project that reaches commercial operation in 2028 sells into spot or short-tenor contracts in the meantime. A project that slips past its milestones faces the guarantees it posted to bid.
An award also does not come with a connection point. UPME assigns transmission capacity through a separate process with its own queue and its own ways to fail. A 15-year contract held by a project without a firm connection is a schedule of obligations attached to an asset that cannot yet deliver, which is why offtake diligence and connection diligence have to be read together.
Counterparty and currency
A contract is worth what its counterparty can pay. The auctioned contracts serve regulated demand through comercializadores whose credit quality varies widely by region: Air-e, the distributor serving much of the Caribbean coast, was placed under government intervention in September 2024. A 15-year contract means 15 years of exposure to whoever is on the other side, and Colombian distributor credit needs real underwriting.
All three routes pay in pesos. Spot, bilateral, and auctioned revenue are all COP-denominated, so currency sits underneath the whole revenue stack whichever route a project takes. Currency risk and the Colombian solar opportunity covers how a USD investor should structure around that, and the auction does not change any of it.
What this adds to a diligence list
What share of expected output is contracted, at what tenor, at what price, and with whom? A single blended "contracted percentage" hides the information you need.
Does contracted revenue alone cover debt service? If the answer requires spot revenue, ask what hydrology assumption is embedded in the spot price used.
Where contracts are shorter than the loan, who bears recontracting risk, and at what assumed renewal price? A model that rolls a three-year contract forward at today's price for a decade has made a large assumption without saying so.
For auction winners: what guarantees were posted, what are the milestone dates, and what does the project earn between commercial operation and January 2030? Is the connection assignment from UPME in hand, and does its timeline match the contract's?
Who is the offtaker, and what is its payment record? For regulated-demand contracts, distributor credit is the counterparty question.
These sit alongside the cost-side questions covered in what drives Colombian solar returns. Revenue route and cost structure are two halves of the same underwriting.
What to watch
I am watching three things through the rest of 2026 and into 2027. First, whether a second auction round is convened, and at what volume. One round covering 3% of regulated demand sets a precedent, and it takes repeated rounds to make a market. Second, whether clearing prices hold as more of the 13.5 GW pipeline competes in the next round. Third, whether the awarded projects secure their connection assignments on a schedule that fits a January 2030 start, since lenders will judge the auction by whether its first cohort delivers.
The government that took office in August 2026 inherits a mechanism convened in January and awarded in July. Contracts signed under it are commitments, but the new administration decides whether to run round two, and that decision is the clearest signal to watch.
Nuentero is software for the full solar project lifecycle, from planning and underwriting through capitalization and execution, in Colombia and the United States. Contact john.crye@nuentero.com or visit app.nuentero.com.
Disclaimer. This article is for general informational and educational purposes only. It does not constitute investment, legal, tax, or financial advice, and it does not take into account the objectives or circumstances of any particular person. Nothing here is an offer to sell or a solicitation of an offer to buy any security or interest in any fund, and no such offer or solicitation will be made except through definitive offering documents (such as a private placement memorandum) to qualified investors in jurisdictions where permitted. Any examples are illustrative of how solar project economics work and are not projections, forecasts, or guarantees of performance. Past or modeled performance is not indicative of future results, and no return is promised or guaranteed. Investments of this kind involve significant risk, including currency risk, regulatory risk, and possible loss of capital. Readers should consult their own professional advisers before making any investment decision.