A charging station is sold as a piece of equipment: pick the power rating, contract the installation, done. Anyone who has operated one knows the equipment is the cheap, settled part. What decides whether it pays for itself is in the electricity bill of the property it was bolted to — and that part is almost never modelled before the purchase. A 150 kW fast charger draws, in fifteen minutes, more power than an entire commercial building consumes in an afternoon. The grid does not ask whether you planned; it only charges.
Start with what usually stalls the conversation. There is no concession, permission or authorisation to operate charging in Brazil. ANEEL Normative Resolution 819/2018, later absorbed into and revoked by REN 1,000/2021, opted for minimum regulation: anyone interested may install and commercially exploit charging points, at a freely negotiated price. There is no approved charging tariff, there is no ceiling, and there is no list of licensed operators.
What the rule does require is something else, and it is the part people skip: the installation must be notified to the distributor in advance whenever there is a new connection, an increase or reduction of load, or a change in supply conditions. This is not decorative bureaucracy: it is the moment you find out whether the service line takes it, whether the transformer takes it, and what the works cost if they do not.
The figure that gives the market its size: in February 2026 the country had 21,061 public and semi-public points, 42% more than twelve months earlier — with direct-current fast charging growing 167% and already accounting for 31% of the network. Whoever installs now is not a pioneer; they are a latecomer with time to correct the project.
This is the number one reason for a charging station with good revenue and a bad result. Grupo A consumers (the high-voltage tariff group) pay two separate things: the energy they consume and the demand they contract — the capacity reserved for them, paid in full, used or not. A fast charger does not change monthly consumption much; it changes the peak brutally.
What happens in practice: the charger goes in, the peak shoots up, it exceeds contracted demand, and the excess demand charge arrives — punitive by design, because the distributor has to discourage exactly that. The owner then raises contracted demand to fit the peak. And from there on pays that amount every month of the year, including the months when nobody charged anything. The charger is used two hours a day; the demand is paid for twenty-four.
Arithmetic the owner can check alone: take today's contracted demand, add the power rating of the charger you intend to install, and multiply the difference by the demand tariff of the last twelve months. Compare that with the projected revenue from charging. If the second does not cover the first with room to spare, the project is not an energy project — it is a marketing one.
The second leak is the peak period — the three late-afternoon hours in which the tariff multiplies. A charger with no load management does exactly the worst thing: it serves whoever arrives, whenever they arrive. And whoever arrives, arrives at the end of the working day.
The fix is not technical, it is commercial: a charging price that varies by time of day. Since the price is free, nothing prevents charging more at peak and less overnight — which is what makes the customer choose the hour that suits both sides. A fleet is the simplest case and the most profitable: the vehicle sleeps in the yard, and charging overnight is a scheduling decision, not a technological one.
This is where storage comes in, and almost always for the wrong reason. Nobody installs a battery at a charging station to "stock cheap energy" — they install it so as not to have to contract more demand. The battery charges slowly through the day, within the demand that already exists, and delivers fast when a car arrives. The grid never sees the peak.
The effect is double, and it is what makes many projects work that otherwise would not: it avoids the service-line reinforcement works, which usually cost more than the charger, and it avoids the permanent increase in contracted demand. Where the arithmetic works, it works there — not on the price difference between overnight and peak, which on its own rarely pays for a battery.
What decides is the curve, not the catalogue: how many cars a day, at what hours, at what power each. Without that curve, sizing a battery is an expensive guess. With it, it is arithmetic.
And here is the part almost nobody tells whoever is about to invest. There is no settled definition of what is levied on charging: the states hold that it is a sale of energy, attracting ICMS (the state tax on goods and energy); the municipalities hold that it is the provision of a service, attracting ISS (the municipal tax on services). The doubt is not academic — it changes the rate, it changes the credit, and it changes who can assess you.
The ground has begun to move. ICMS Convênio 182/2025 — a convênio being an agreement struck among the states at CONFAZ, their ICMS council — concluded on 5 December 2025, opened the way for the states to deal with energy destined for charging stations. Santa Catarina moved first and regulated it through Decree 1,524 of 13 May 2026, creating an optional tax substitution regime in which the distributor takes on the collection of ICMS.
Note the word optional: whoever operates has to decide whether to join, and the decision has consequences for cash and for credit. Choosing on autopilot, or not choosing, is also a choice — only one made by omission.
The horizon changes everything again: Constitutional Amendment 132/2023 replaces ICMS and ISS with the IBS, with a transition between 2026 and 2032 and the extinction of the current taxes by 2033. In other words, the dispute ends — but after the investment has been made and the assessments have lapsed, or not. Whoever invests today has to cross the fog, not wait for it to lift.
Three owners, three sets of arithmetic. At the shopping centre or supermarket, charging is rarely the business: it is dwell time, and the charger pays for itself in the store's average ticket, not in the kilowatt-hour. In the condominium, the problem is legal before it is electrical — cost apportionment, individual metering and the bylaws; without that settled, the first bill becomes a residents' meeting. In the fleet, it is the only one of the three where the arithmetic is clean: diesel is replaced by electricity, with controlled timing and predictable use, and the saving is measurable the following month.
There is also the case of own generation feeding the charger. It works, and it improves the bill — but mind the fit: the sun produces at midday and the car arrives at six. Without storage, or without time-of-day management, the plant and the charger are two good projects that never meet.
First, the curve: the bills of the last twelve months, the contracted demand, the consumption profile by time of day. Then, the sizing: what power fits into the connection that already exists, and from what point a battery is worth more than construction works. Next, the tariff classification, which changes with the new load profile and is almost never reviewed after the installation. Then the tax decision, in writing, with the regime of the state where the operation runs. And, finally, the price of charging, which is where the project does or does not become a business.
What we do not promise: a return. Anyone who arrives with a payback period before seeing your bill is selling equipment. What we promise is that the arithmetic will be done with your numbers, and that the result will be said out loud even when it is no.
Official sources, at the exact point: the article of law, the service or the search you can use today. None replaces analysis of the specific case, which is our work.
Each note starts from a sourced figure, explains the mechanism behind it and ends with a test you can run yourself, today, without calling us. They are deliberately short pieces: what does not fit in them fits in a conversation.
A note does not replace the review — it shows the mechanism and gives the reader back the ability to check alone. That is deliberate: whoever understands the mechanism negotiates better, including with us.
If the reading raises a question the note does not answer, it is probably the right question. Worth sending: answering a new question is the cheapest way to find out what we have not explained properly yet.
We do not promise a percentage before looking at the data, we do not promise a homologation date — the procedure belongs to the State: and we do not chase speculative theses to fatten an opinion. Nor do we take commission for referring anyone: whoever executes is a dedicated partner of the house, contracted and paid by us, within our price.
A conversation, with no materials and no proposal, to understand what is blocked. If there is matter to work on, the next step is the letter of authorisation and secrecy, a document that limits our own access before anything else. If there is not, we say so, and the conversation ends there.
Every statement on this page comes from a public norm, and the links above lead to the full text: not to the agency's portal. That is deliberate: whoever wants to check can check alone, without asking anything of anyone, which is the only kind of trust that does not depend on whoever asks for it.
No thesis moves forward without its classification written beside it. That is what separates a survey from a promise.
Say in two lines what you need to resolve. An account manager replies personally, in business hours, and the conversation starts where it makes a difference.
This pagepoint the cameraIt reached the desk of whoever answers. The reply comes in writing, during business hours.
