Power grid resilience to extreme weather events: why reinforcing everything isn't the answer

At a substation in Assis, in the interior of São Paulo, Taesa's CEO asked a question that seems simple. Rinaldo Pecchio looked at a transmission tower and said: you can build one to withstand 120 km/h winds. Or you can build one for 400 km/h. The second one will be much heavier, will need a much larger foundation and the cost will rise considerably. "Society needs to decide how much it wants to pay to have oversized infrastructure," he said, on September 3, 2026 (Poder360, 09/03/2026).

That sentence is not about engineering. It's about money and about risk. And it sums up the dilemma the Brazilian power sector is facing right now, with no more room to postpone it.

There's no way to shield the entire grid against any wind. Nor is it possible to do nothing. The question that remains, the only one that truly matters, is: where is it worth reinforcing and where is it worth responding fast? And that answer only exists if someone knows, before the event, which stretch of the grid is exposed.

What 2026 has already shown, in numbers and in fallen towers

In the early hours of July 24, a storm in the interior of São Paulo brought gusts of up to 160 km/h, according to IAC (121 km/h in the official measurement). The result: 54 transmission towers knocked down and more than 250 CPFL Paulista poles damaged (Megawhat/UOL, 07/27/2026). Governor Tarcísio de Freitas called the event "unprecedented" and said current weather models still struggle to predict extreme events of that magnitude.

The ONS recorded automatic shutdowns between 4:20 and 4:46 that morning. Three high-voltage lines went down almost simultaneously: the 440 kV line between Água Vermelha and Araraquara (ISA Energia, 5 towers), the 500 kV line between Marimbondo II and Araraquara (Axia Minas, 5 towers) and the 500 kV line between Nova Ponte 3 and Araraquara 2, with 16 towers affected, 8 per circuit (Megawhat/UOL, 07/27/2026).

Two weeks later, it was Rio Grande do Sul's turn. A bomb cyclone knocked down transmission towers in the Lagoa Mirim region. Two were left submerged, leaving Santa Vitória do Palmar and Chuí without power. In total, 24 towers suffered damage (Canal Solar, 08/10/2026; G1 RS, 08/09/2026). A task force from CEEE Equatorial together with Axia mobilized more than 90 professionals and a helicopter. About 67,000 consumers were left without power on August 9. Generators were brought in to keep essential services running.

Two states, two different weather technologies (convective windstorm in one case, extratropical cyclone in the other), the same outcome: tower on the ground, power cut, crew scrambling to catch up with the damage.

What the ONS is saying, without mincing words

It isn't just Taesa that has this bill on the table. The very operator of the Brazilian power system is talking about it, publicly, without euphemisms.

Valter Cardeal, ONS operations director, told Valor that the country needs to reinforce grids to face extreme weather. He cited "an increasingly higher frequency of tornadoes and extratropical cyclones" and described the moment as "a scenario that had not occurred before" (Valor, 08/18/2026).

Solange Ribeiro, chair of ONS's board of directors, went in the same direction a week earlier: "one of the biggest challenges imposed on transmission and distribution companies is the capacity to anticipate new extreme events" (Valor, 08/11/2026).

Anticipate. Not react faster afterward. Anticipate beforehand.

This is where the conversation shifts. It leaves the structural engineering room and enters the risk management room.

Why "reinforcing everything" isn't the answer, and neither is "doing nothing"

Taesa operates 16,600 km of lines across 19 Brazilian states (Poder360, 09/03/2026). Reinforcing every kilometer to withstand the worst hypothetically possible wind is, in practice, unfeasible. Pecchio was direct: "there is practically no way to reinforce every line in Brazil for any extreme event." The foundation gets bigger, the tower gets heavier, the cost rises, and so does the execution timeline. Multiply that by the length of the Brazilian grid and the number turns into financial unfeasibility, not technical conservatism.

But the opposite also has a cost. Doing nothing means rebuilding the grid after the event, under pressure, with crews deployed in a rush, consumers without power, industry at a standstill, regulatory fines, a damaged image. The July event in São Paulo and the August one in Rio Grande do Sul show both sides of this equation: when the wind hits an unprepared stretch, the cost shows up anyway. It just shows up later, bigger, and without planning.

The strategy Taesa describes is neither one nor the other. It's a combination: preventive maintenance, weather monitoring, contingency plans and emergency towers for quick reconnection. The company has invested about R$ 15 million over the past two years in this maintenance and resilience front (Poder360, 09/03/2026). It isn't total shielding. It's targeted preparation.

The decision that remains: wind probability crossed with asset criticality

If reinforcing everything is unfeasible and doing nothing is expensive, the real decision is a different one: which stretch of the grid should be reinforced first, and which stretch should be handled with fast response instead of structural resistance?

This decision has two variables. The first is probability: what is the real chance of extreme wind, a cyclone or a tornado hitting that specific stretch in the coming years, not Brazil in general, not the whole state, but that particular span between two towers. The second is criticality: what that stretch feeds. A redundant line, with an alternative route, is not the same as the single line that sustains an industrial hub or a capital city.

Crossing the two variables, the investment priority emerges on its own. A stretch with high probability of an extreme event and high criticality: structural reinforcement, it comes first. A stretch with high probability but low criticality, with redundancy available: fast response, emergency tower, crew on standby. A stretch with low probability: monitoring, no heavy spending for now.

The problem is that this math only works if the first variable, local probability, is reliable. Public weather forecasting today has a resolution of about 25 km. That's enough to say strong winds will hit "the Araraquara region." It isn't enough to say which of the five towers between two substations lies in the gust's path. That's the scale gap Governor Tarcísio de Freitas unintentionally described when he said current models still struggle to predict extreme events of that magnitude (Megawhat/UOL, 07/27/2026). The problem isn't a lack of weather data. It's a lack of data at the asset scale.

What changes when the alert comes by asset, not by city

This is where method comes in, not product. Climate intelligence, in the sense that makes an operational difference, means turning weather forecasting (which speaks about a region) into risk by asset (which speaks about a tower, a line, a substation). i4sea works with a resolution of 1 to 3 km, against the roughly 25 km of public forecasting, crossing more than ten years of proprietary reanalysis with the monitoring of 18 hydrometeorological hazards, today present in more than 100 critical assets across Latin America and Europe.

The practical difference is this: instead of knowing that "strong winds are coming to the region," the operator knows that tower 12 on the line feeding the industrial hub has an elevated probability of a critical gust in the next 48 hours, while tower 40, on a redundant line, can wait. This doesn't prevent the wind. It prevents the wind from catching everyone by surprise at the same time, on the same line, with no plan.

Where most people see the unforeseen, it's possible to see a management variable. The difference between the two readings is the data that arrives before the event, not after it.

Extreme weather has become a design variable, not just an operational one

What happened in São Paulo in July and in Rio Grande do Sul in August 2026 was not an isolated coincidence. It was the same pattern repeating itself across different geography and different weather technology: grid caught by surprise, tower on the ground, emergency response. The ONS is already saying this publicly, through two different spokespeople, in different weeks. This isn't alarm, it's diagnosis.

Pecchio's question (120 km/h or 400 km/h) looks like an engineering choice. It isn't. It's a risk portfolio choice: how much to invest in structural resistance, how much to invest in response capacity, and how to decide that stretch by stretch instead of for the whole grid.

Whoever decides that allocation with local probability data gets ahead of the problem. Whoever decides without that data finds out about the problem when the tower has already fallen. The cost difference between the two approaches isn't a matter of philosophy. It's in the rebuilding bill, in the fine, in the consumer without power, in the industry at a standstill.

If you work with transmission, distribution, generation or grid operation planning, this is the moment to understand where your grid is exposed before the wind decides for you. Get to know i4cast and talk to the i4sea team about how to map climate risk by asset, not by city.

Sources

Poder360, 09/03/2026, "Taesa: Brazil will have to decide how much to spend against extreme events", https://www.poder360.com.br/poder-infra/taesa-brasil-tera-de-decidir-quanto-gastar-contra-eventos-extremos/

Megawhat/UOL, 07/27/2026, "Winds of up to 160 km/h knock down 54 transmission towers in São Paulo", https://megawhat.uol.com.br/distribuicao/ventos-de-ate-160-km-h-derrubam-54-torres-de-transmissao-em-sao-paulo/

Canal Solar, 08/10/2026, "Cyclone: transmission towers in Rio Grande do Sul", https://canalsolar.com.br/ciclone-torres-de-transmissao-rio-grande-do-sul/

G1 RS, 08/09/2026, "Cyclone knocks down power towers and leaves two cities without electricity", https://g1.globo.com/rs/rio-grande-do-sul/noticia/2026/08/09/torres-de-energia-eletrica-desabam-durante-passagem-de-cicl

Valor, 08/18/2026, "Country needs to reinforce grids to face extreme weather, says ONS", https://valor.globo.com/brasil/noticia/2026/08/18/pais-precisa-reforcar-redes-para-enfrentar-clima-extremo-diz-ons.ghtml

Valor, 08/11/2026, "The goal is to ensure the power system is prepared for the challenges", https://valor.globo.com/brasil/noticia/2026/08/11/objetivo-e-garantir-que-sistema-eletrico-esteja-preparado-para-desafio