Brazil's battery capacity auctions: how low can bids go?
Brazil's first battery capacity auctions (LRCAP) take place on 2 and 4 December 2026. They favour projects with three advantages: a connection point that receives a ranking bonus, typically in the Northeast, cheaper financing, such as development-bank debt, and the buying power to secure cheaper equipment.
Our headline floor, the ranked break-even bid at which a well-placed project earns a 12% real return on equity, is about R$0.92m per MW a year. That is roughly 58% of the open auction's R$1.58m price cap.
The auction is pay as bid: each winner is paid its own offer, not a common clearing price, so bidders want to stay above break-even. At a crowded connection point, however, a low opening bid (sealed offer) may be needed to survive the first round. Bidders submit these offers without seeing rivals' bids or the auction's current price. A low break-even gives a bidder more room to make that offer without sacrificing its required return. Projects that fail the grid-capacity test cannot return.
The decision is therefore how close to break-even to bid before knowing how much capacity will be bought.
This is the third article in our LRCAP 2026 series. The first sized Brazil's storage need, and the second set out the rules bidders sign up to. This one prices them.
To discuss this with the author, contact santiago@modoenergy.com.
Key takeaways
- Imported equipment, regional tax relief and debt from Brazil's development bank (BNDES) bring our headline project's break-even to R$1.02m per MW a year. It ranks at R$0.92m on a bonus bus.
- A 10% capex saving lowers annual break-even by about R$0.08m per MW. That is twice the effect of a two-percentage-point lower equity return.
- Local-equipment projects have their best chance on 2 December, when they compete only with each other. Against comparable imports on 4 December, cheap BNDES financing offsets equipment prices only about 8% to 12% higher.
- Rivals on the same bonus bus receive identical ranking benefits. Bidders cannot rely on that bonus to preserve a margin above break-even.
The lowest bids combine a bonus bus, cheap capital and buying power
At certain connection points, mostly in the Northeast, bids are ranked at 90% of their value; we call them bonus buses. Winners there are still paid their full bid (Part 2). In the Northeast, the ranking bonus and income-tax relief from SUDENE, the regional development agency, more than offset higher grid charges in our model. Together, they put the ranked break-even of a project at the median bonus bus about R$0.07m per MW a year below comparable projects in the South, Southeast and Centre-West.
Where infrastructure sharing is permitted, reusing substations, lines and civil works can give existing wind or solar site owners a cost advantage over comparable greenfield projects. The battery must still stay standalone, charging only from the grid and keeping its own meters and grid contracts (Part 2). Cells and inverters must be new, and the project still needs sufficient grid capacity.
The bonus lets a Northeast project bid up to 11% more than a rival elsewhere and still rank level. Against a typical Southeast bus with lower grid charges, the edge is about 7%. It gives no advantage over neighbours receiving the same adjustment, though. Where competition for the same grid capacity is intense, bidders are likely to give up most of that advantage through lower offers. On our reading, the bonus decides where batteries are built more than what they earn.
Italy's MACSE auction also weighted bids in the ranking only. In its 2025 round, bids cleared about 65% below the cap. Enel, a large developer with access to cheap capital, secured roughly two-thirds of the volume.
If Brazil follows the same pattern, the bonus will steer where batteries are built. Bidders with cheaper capital and purchasing scale, which can include large utilities, may be best placed to win them.
Lower 2027 equipment costs and regional tax relief cut modelled break-even by about 10%, to R$1.02m
Because owners do not keep trading gains or losses, bidders must set the fixed capacity payment high enough to cover their costs and required return over the 15-year contract. We define break-even, or floor, as the payment that earns a 12% real return on equity in our model. The system operator (ONS) controls dispatch, and the trading result goes to the account that funds the contracts (Part 2).
Our reference case, using the planning agency EPE's cost and financing assumptions, requires R$1.13m per MW a year. New consumption taxes are charged on top, and later tax changes adjust the payment. Grid charges are not reimbursed, although a fully dispatched battery pays them only on injection. They can also change during the contract, while the payment tracks only inflation.
A 10% capex saving cuts break-even twice as much as a two-point cut in the equity-return target
Each R$500/kW saved (about a tenth of our headline project's capex) lowers annual break-even by about R$0.08m per MW. On our site and financing assumptions, a bid below R$1.0m per MW a year, before the ranking adjustment, needs capex below about R$4,770/kW.
Our model starts from EPE's R$5,500/kW at December 2024 prices and splits it 75/25 between imported equipment and local works, in line with the cost breakdown in Modo's 2026 capex survey. It carries both to mid-2027 procurement: equipment with Chinese system prices, the exchange rate and import taxes, and local works with inflation. That gives about R$4,920/kW, roughly a tenth lower.
Capital matters as well. A two-percentage-point change in the required return on equity moves the break-even bid by about R$0.04m/MW/yr. Each percentage-point increase in the market debt rate adds only about R$0.01m, because a third of the debt comes from BNDES at a fixed rate.
Small differences in financing terms and equipment prices can therefore decide which project stays profitable at a competitive bid.
The contract requires the full four hours for all 15 years. ONS decides how hard the battery cycles, up to 366 full cycles a year. Our model meets that with extra capacity at commissioning plus later additions, known as augmentation. This costs about R$0.06m per MW a year less than installing all the required capacity upfront.
Our model excludes post-contract revenue; at current Northeast spreads, standalone trading would not cover O&M and grid charges.
With the median bonus-bus grid charge, SUDENE relief and market debt supplemented by BNDES lending on local works, our headline project breaks even at R$1.02m per MW a year. Its Northeast bonus-bus location brings the ranked bid to R$0.92m.
Local equipment competes best on 2 December; on 4 December, only within 10% of imports
The 2 December auction is reserved for qualifying local-content systems, while the 4 December auction is open to all (Part 2). The development bank’s Fundo Clima lends at 6.5% plus a spread, but only for national equipment, materials and services. Lending is capped at R$1bn per group a year, enough for about 300 MW.
A project with a qualifying local battery pack can finance all of its financeable items from Fundo Clima. On our reading of BNDES's terms, an imported-equipment project can borrow only for its local works.
The local-content route changes how much of a premium cheap debt can absorb. Route 1 is the only BNDES route that does not require a local battery pack. A Route 1 system can finance only 80% of its financeable items from Fundo Clima, which offsets a price premium of about 8%. With a local pack, all financeable items qualify, and the threshold rises to about 12%.
No public price exists for either form. A BNDES director, Nelson Barbosa, has put the local premium at 20 to 40%, the band shaded on the chart. If the premium falls within that range, our model puts local-content systems above comparable imported systems in the open auction, despite their cheaper debt.
Barbosa argues that cheap financing closes the gap. Our comparison prices the imported project's finance as inflation-linked market debt in reais; against costlier finance, cheap BNDES debt would offset a larger premium.
If local-equipment premiums are as large as quoted, the separate auction on 2 December is the stronger opportunity in our model. Matching otherwise comparable imports on 4 December needs a premium within the range above.
Projects entering the open auction must also account for grid capacity already taken by first-day winners, leaving less grid capacity available at their registered connection point.
The sealed bid decides most of the price, because the grid test comes first
Whether winners are paid more than their break-even depends on competition for grid capacity and on bidding after the sealed round.
In the sealed round, bids are ordered by their ranked price. Projects pass the grid test only if their full registered export capacity fits within the remaining capacity at their substation and across the wider network. Those that do not fit are excluded permanently, whatever they bid. Offering fewer megawatts does not help, because the test uses the registered capacity.
The test is about local grid room, not the national volume. A project can pass it and still sit above the volume being bought, and then has to compete on price.
After that round, the auction is pay as bid. A project that remains within the volume being bought can keep its sealed offer and, if it wins on that offer, is paid that price.
Projects outside the volume must bid below the current price to move into contention, displacing another project. The marginal project, the last one within the volume being bought, sets the current price. None of this is open to a project already excluded by the grid test.
On a crowded Northeast bonus bus, rivals receive the same ranking adjustment, so it gives them no advantage over each other. A bidder may therefore need to submit an offer close to break-even in the sealed round, before knowing how much capacity will be bought.
How much margin it can keep depends on the volume and costs of rival projects competing for the same grid capacity.
Four questions decide whether a bid can win
A bidder's position comes down to four decisions:
- Does the registered project fit the room behind its shared grid limit, once the 2 December contracts are counted?
- What equipment price can it secure?
- Which financing route can it access: imported equipment with BNDES debt on the works, a Route 1 system or a local battery pack?
- What minimum payment earns its required return under those assumptions?
Qualification confirms eligibility; first-day awards help bidders reassess the grid room left for 4 December. From 17 November, bidders know whether EPE has qualified their own projects, but not how many rivals share their bus. Published totals are regional, and bids stay confidential until the result. The 2 December result will be the first public sign of how much grid room is taken.
We will cover each result on 3 and 5 December.
Model note: all figures are in May 2026 reais.





