Brazil's record auction prices assume a rate cut that hasn't happened
In January 2025, Brazil's A-6 auction cleared 1,040 MW of wind at BRL 98.89/MWh and 530 MW of solar PV at BRL 84.39/MWh, roughly $23.90 and $20.39 per megawatt-hour, per ANEEL and CCEE results. That same month, Brazil's policy rate stood at 14.75% and was headed to 15%, the highest Selic since May 2006, per the Central Bank. Those two numbers were set in the same economy, in the same quarter, and almost nobody discussed them together.
The headlines treated the auction as a cost story. Panel prices down, turbine prices down, developers competing hard for scarce slots in a capacity-constrained grid. All of that is real. It is also not the number that decides whether the project gets built on the terms it was bid.
The rate the bid actually has to clear
A 20-year regulated PPA sold through CCEE is indexed to IPCA, so the nominal tariff rises with inflation over the contract's life. That protects the real value of the revenue line. It does nothing about the rate at which a lender discounts it today. Project debt in reais, even investment-grade infrastructure debt, prices off a curve anchored to Selic, and Selic has spent 2025 at its highest level in two decades.
Run that arithmetic and a $20/MWh solar bid has very little room left for debt service once construction cost, O&M, curtailment risk and a double-digit cost of debt are all stacked against it. Someone bidding that price is underwriting one of three things: a construction cost curve that keeps falling faster than it already has, access to concessional BNDES financing most new entrants don't qualify for, or dollar-denominated debt hedged against a real-indexed tariff, which works beautifully until the hedge itself gets repriced.
A record-low auction price is not evidence that building got cheaper. It is evidence that someone is financing the gap, and the gap doesn't disappear just because nobody names it in the press release.
The indexation argument only goes so far
To be fair to the bidders, IPCA indexation is a genuine structural advantage Brazil's auction system has over Peru's or Chile's flat-dollar contracts, and it is the reason this isn't as reckless as the headline number alone suggests. A tariff that grows with inflation over two decades is worth meaningfully more than its first-year print. But indexation protects the numerator. It says nothing about the denominator, and the denominator is where a 15% policy rate shows up in the model whether or not the term sheet mentions it.
Wind did slightly better than solar in the same auction, clearing close to its own best-in-class LCOE estimate rather than well under it, which tells you the discipline varies by technology and probably by sponsor. Not every bid at this auction is mispriced. Enough of them likely are that the average is worth scrutinizing rather than celebrating.
What we think follows
Lenders underwriting this cohort should be asking for the cost-of-debt assumption behind the bid, not just the IPCA escalation schedule. A developer who can show investment-grade, BNDES-eligible, or properly hedged financing at these prices has a real project. A developer who modeled Selic drifting back toward single digits within a few years to make the DSCR work has a forecast, dressed up as a financial close.
There is a reasonable counter here too. Brazil's central bank has cut aggressively before, and a multi-year rate-cutting cycle inside a 20-year contract term is not an exotic bet, it is closer to the base case most Brazilian macro desks are already running. If Selic normalizes toward 8 to 9% by the back half of this decade, several of these projects refinance into comfortable territory on schedule. The risk is what happens to the ones that don't survive the years before that happens.
Our own pipeline sits in Peru, not Brazil, where the last renewable auction cleared 20-year solar PPAs nearer $27.36/MWh, per Osinergmin, a number that already draws its own skepticism in a market without Brazil's benchmark rate or its indexation cushion. We'd rather sign a contract priced as if debt has a cost attached to it, which is part of why Olmos Wind, Frontera, Aurelion and Solar II are underwritten against fixed Ready-to-Build dates instead of a bid that assumes the macro backdrop improves on schedule.
The open question
Does Selic actually ease enough over the next few years to make this cohort's debt sizing work out the way the auction price implies, the way Brazilian rate-cutting cycles have gone before? Or does a chunk of this generation reach construction only to restructure, sell down to a buyer with a lower cost of capital, or quietly renegotiate before COD? I don't think the auction result tells you which. I think it tells you which projects were careful enough to have an answer ready either way.
See the full active portfolio or read more Insights.