José Mendieta, Energy Director, Elea Data Centers, says Brazil’s energy challenges highlight the need for better grid coordination – not blame on data centres.
Debates about data centers in Brazil often begin with a familiar assumption: that the sector is putting pressure on the power system and competing with the public for energy. But that framing is too narrow and it doesn’t reflect the broader reality of how the grid is evolving.
Brazil isn’t facing a simple shortage of generation. It has a largely renewable electricity mix and more installed capacity than many markets. In 2024, 88.2% of the country’s electricity came from renewable sources and that share has remained above 70% since 2004. By comparison, renewables account for about 41% of global electricity generation. Brazil’s advantage is clear. The issue is that strong generation does not automatically translate into efficient delivery or use.
The real challenge today isn’t demand alone. Instead, it’s the growing difficulty of balancing generation, transmission, storage and consumption in real time as the grid becomes more complex. The challenge is increasingly relevant in any market trying to align renewable growth with transmission capacity, grid reliability and large new sources of demand.
A system under strain for reasons beyond demand
Brazil’s energy debate often overlooks the role of transmission limits, intermittent generation and grid congestion. These pressures are becoming more visible as renewable capacity expands. Electricity may be available in one part of the country but that does not mean it can be moved efficiently to where it is needed or absorbed by the system at the right time.
This is where curtailment becomes an important signal. Curtailment, the forced reduction of generation to preserve system stability, has intensified as renewable energy growth outpaces the infrastructure needed to transmit and integrate it. It is driven by several factors, including temporary oversupply, operational reliability requirements and limitations in the transmission network.
The scale is significant. Between January and August 2025, Brazil’s renewable energy sector recorded estimated losses of R$1.7 billion due to curtailment, or US$340 million. During that same period, centralized solar generation experienced an average curtailment rate of 13.7%, peaking at 27.8% in June. Another study found that 17.2% of the country’s potential solar and wind generation had been discarded since the start of 2025.
These figures point to a structural problem. Brazil is generating large volumes of renewable power but system constraints are preventing part of that energy from being used productively.
Data centers are part of the conversation but not the villain
Against that backdrop, data centers are often treated as an emerging burden on the power system. But the numbers suggest a more measured view.
According to Brasscom, the association of Information and Communication Technology (ICT) and Digital Technologies companies in Brazil, data centers represented 1.7% of Brazil’s total electricity consumption in 2024, equivalent to 8.2 TWh out of 650.4 TWh. Even with projected growth, that share is expected to reach 3.6% by 2029. By comparison, the industrial sector accounted for roughly 36% of electricity consumption and the residential sector about 28%.
That doesn’t mean data centers are insignificant. They are large loads and they need to be planned carefully. But it does mean the public narrative often exaggerates their role while understating broader inefficiencies in the system.
One example is distribution loss. In 2023, total losses in Brazil’s electricity distribution network represented 14.1% of the energy injected into the grid, including 7.4% in technical losses and 6.7% in non-technical losses. That volume was roughly equivalent to the annual residential consumption of Brazil’s entire South region.
When those inefficiencies coexist with growing curtailment, the question is no longer simply who uses electricity. It becomes whether the system is being planned well enough to reduce waste and convert available energy into value.
Why siting and coordination matter
This is where data centers enter the discussion more constructively.
It would be overly simplistic to argue that data centers do not need to be near metropolitan areas or that they can broadly be located anywhere near generation assets.
Geographic location remains highly relevant and not all data center models have the same flexibility. Some workloads, particularly AI- or compute-intensive operations, may allow for greater distance from dense urban centers. But that is not a universal strategy nor should it be treated as one.
The better approach is coordinated planning among developers, utilities and regulatory bodies so that the expansion of generation and transmission can progress alongside the growth of digital infrastructure. Where possible, and where it makes operational and economic sense, data centers can be brought closer to power generation plants or underused transmission hubs. The objective isn’t to suggest that proximity alone solves system inefficiencies. It’s to improve the feasibility of new connections and reduce unnecessary infrastructure strain.
The most meaningful benefits of sitting near strong electrical infrastructure are practical. These sites can offer lower connection costs, better reliability and availability from the high-voltage network, a greater likelihood of securing capacity at that point of connection, lower operating costs tied to energy infrastructure and a more realistic path for serving large loads above 10 MW.
That is a more accurate rationale than suggesting line-loss reductions alone justify the decision.
Not all data center loads look the same
Another point that warrants greater precision is how data center demand is described.
Cloud data centers often have relatively stable and predictable load profiles, which can be valuable in a system increasingly shaped by intermittent sources.
But that profile does not apply equally across the sector. AI-focused facilities may have much larger fluctuations in demand and those variations may create additional pressure on the transmission system if they are not mitigated properly.
That does not weaken the case for data center growth. It simply reinforces the need for more sophisticated planning. Different workloads place different demands on the grid and those differences need to be reflected in infrastructure design, load management and long-term system coordination.
The real challenge is turning energy into value
Brazil’s debate over data centers shouldn’t be framed as a question of whether digital infrastructure is consuming too much electricity. The more important question is whether the country is building the systems needed to use its energy advantage efficiently.
Brazil has strong renewable generation. What it needs is stronger coordination among generation, transmission and large new loads so that curtailment, congestion and infrastructure bottlenecks do not continue to undermine that advantage.
That is why the ‘energy villain’ narrative falls short. It shifts focus away from the real challenge: modernizing the grid, unlocking transmission investment and aligning supply, demand and infrastructure planning in a way that supports both reliability and growth.
Brazil’s experience is a useful reminder that renewable advantage alone is not enough. The larger test, in any market expanding digital infrastructure and clean energy at the same time, is whether the power system can turn available energy into reliable, productive use.
Data centers, when well-planned and integrated into that process, aren’t the core problem – they’re the answer.
In a grid struggling with curtailment and congestion, well-planned infrastructure acts as a productive anchor. The challenge isn’t finding more power; it’s evolving the system fast enough to turn available energy into value.
