logo
Draining cities dry: the giant tech companies queueing up to build datacentres in drought-hit Latin America

Draining cities dry: the giant tech companies queueing up to build datacentres in drought-hit Latin America

The Guardian22-05-2025

It is a warehouse the size of 12 football pitches that promises to create much-needed jobs and development​ in Caucaia city, north-east Brazil​. But it won't have shelves stocked with products. This vast building will be a datacentre, believed to be earmarked for TikTok​, the Chinese-owned video-sharing app​, ​a​s part of a 55bn reais (£7.3bn) project​ to expand its global datacentre infrastructure​.
As the demand for supercomputer facilities rises, fuelled by the AI boom, Brazil is attracting more and more tech companies. The choice of Caucaia is no accident. Several undersea cables carry data from the nearby capital of Ceará state, Fortaleza, to other continents. The closer to the cables, the greater the traffic capacity and the lower the latency, or response time, between two points on the internet network.
Caucaia also has Pecém EPZ, a zone in which companies can produce goods and services for export with a range of tax benefits and simplified bureaucratic procedures.
Yet Caucaia also suffers from extreme weather events, including droughts and heavy rains, according to data from the Digital Atlas of Disasters in Brazil and the Integrated Disaster Information System.
In 16 of the 21 years between 2003 and 2024, a state of emergency due to drought was declared in the city at least once. In 2019, almost 10,000 people were affected by water shortages, the Digital Atlas of Disasters shows. As reservoirs were depleted, the water became unfit for consumption, leading to crop losses and difficulty accessing basic food for the population.
Datacentres use vast amounts of energy and water to cool their supercomputers. Nevertheless, public authorities are greenlighting their construction in cities that have persistently suffered from drought. Caucaia is not an isolated case.
According to the Digital Disaster Atlas, five of the 22 datacentres planned are located in cities that have suffered recurring droughts and water shortages since 2003.
So far, the Chinese social network has not been named in the permit applications in Caucaia. But in February, the state government's chief of staff, Chagas Vieira, confirmed in an interview with a local radio station that it was in talks with a Chinese company, and representatives from TikTok and its parent company, ByteDance, have met senior government officials, including the vice-president and the development, industry, trade and services minister Geraldo Alckmin.
ByteDance has been approached for comment.
The company officially responsible for the project is Casa dos Ventos, a Brazilian wind energy provider that has invested in the datacentre sector. The company's founder and president, Mario Araripe, announced last year that he aimed to attract a large global tech company, such as Apple, Amazon, Google, Meta or Microsoft, to fill the place with its computers.
Casa dos Ventos has already obtained one of the three necessary licences needed from the state of Ceará. According to the State Environmental Superintendence (Semace), the project received a licence for 'water consumption of 30m³/day in a closed circuit, supplied by an artesian well'. Access to further details has been refused due to commercial confidentiality.
Casa dos Ventos says it is 'committed to transforming Porto do Pecém into a technology innovation and energy transition complex' with 'the largest datacentre and green hydrogen project in the country, which will use renewable energy'.
Projects that require substantial power, such as datacentres, must apply for special permission from the Brazilian government to operate. In 2024, at least seven of the 21 authorisations published by the ministry of mines and energy were related to datacentres.
Casa dos Ventos is also responsible for another project for a datacentre under state review in Campo Redondo, Rio Grande do Norte, which has experienced drought for 14 of the last 21 years, analysis of data from the Digital Disaster Atlas shows. During a water shortage in 2022, the municipality requested federal support and water trucks were dispatched to meet demand.
A similar situation has occurred in Igaporã, Bahia, where a Brazilian renewables energy company, plans to build two datacentres. The city has declared a state of emergency due to drought and dry spells in 12 of the 22 years since 2003. In 2021, about 5,500 people were affected by water shortages in rural areas.
There is no transparency on water consumption by datacentres under construction in these cities. Companies do not publish this information voluntarily, and the government refuses to release technical documentation for licensing, citing industrial secrecy.
In early April, the National Electricity System Operator (ONS) rejected Casa dos Ventos's request for access to the grid, citing risks to grid stability. The ministry of mines and energy then asked for the calculations to be redone to explore possible grid adjustments.
Big tech companies admit that they are consuming water in sensitive areas due to the demands of AI. In its 2024 sustainability report, Microsoft said that 42% of its water came from areas of 'water stress', and Google said the same year that 15% of its water use was in areas of 'high water scarcity'.
The large amounts of water used by datacentres keep computers and machines cool, preventing them from overheating. However, some is lost to evaporation, which can exacerbate the climate crisis in the regions where they are set up. As AI models improve, they need more processing power, which requires more energy and cooling. This means water and energy consumption are set to increase.
The International Energy Agency conservatively estimates that datacentre energy consumption will double to 945,000 GWh by 2030 – the equivalent of Japan's annual energy consumption. Emerging countries such as Brazil will account for about 5% of the projected growth over this period.
Water consumption is set to rise significantly, with 4.2bn to 6.6bn cubic metres needed for global AI demand in 2027, according to researchers at the University of California, Riverside, and the University of Texas at Arlington. This is more than half of the UK's annual water consumption.
However, according to Shaolei Ren, a researcher at UC Riverside and co-author of the article, there is an essential difference between withdrawal (the water taken out of the system) and consumption (water withdrawal that evaporates) regarding datacentres.
Sign up to Global Dispatch
Get a different world view with a roundup of the best news, features and pictures, curated by our global development team
after newsletter promotion
'While residential users typically don't use much of the water they withdraw, datacentres often use 60% to 80% of it,' says Ren. In other words, the water is lost.
Datacentres can be cooled in two ways. One is air conditioning, a power-inefficient solution for extensive facilities. Water is the second option.
One technique is to use radiators with fans in a closed water circuit, where the water is recycled or reused, similar to the system in a car engine, but costs are high. Another option is cooling towers, which remove heat from hot water using evaporation, so the cold water can be pumped back into the system. The last method involves spraying water into the air to make it more humid and lower the temperature.
But there are still some inefficiencies. 'Both evaporation and spraying result in water loss,' says Emilio Francesquini, an associate professor at the Federal University of ABC.
A small datacentre with a 1MW capacity consumes 25.5m litres of water yearly, losing 1% (255,000 litres) via evaporation.
In Pecém, alternatives to water extraction would be seawater desalination or the purchase of recycled water from Fortaleza, says Francisco Assis de Souza Filho, a professor at the Federal University of Ceará.
He says it is up to the state government, as part of the environmental licensing process, to grant a datacentre a water concession.
Ronildo Mastroianni, a technical director at Esplar, an NGO that has worked in Ceará for 50 years, says a project that requires intensive water consumption in a semi-arid region does not make sense. 'A small push is all it takes to become arid,' he says.
Mastroianni cautions that these projects could result in permanent watershed changes. He warns of a weakening caatinga – a fragile ecosystem – and growing food insecurity due to water shortages in rural areas. No representatives from local non-governmental organisations or the dozens of quilombola and Indigenous communities were invited to discuss the project, he says.
Other Latin American countries have also experienced a datacentre industry boom. Chile has 22 datacentres, 16 located in the greater Santiago area. In December, the government announced a national plan for 30 new projects, as the country is expected to move from a medium to a high level of water stress by 2040, with the likelihood of a decrease in water availability.
In Chile, the government and companies are facing a growing resistance. In 2019, Google announced plans for a second datacentre in Santiago, where the activist organisation Socio-Environmental Community Movement for Land and Water (Mosacat) estimated that the project would extract 7bn litres of water annually.
After a series of demonstrations, a Santiago court reviewed the project. In early 2024, the court halted it pending Google's review of the environmental impact.
One of those who took part in the fight to stop the project was Tania Rodríguez, a member of Mosacat.
'It's turned into extractivism,' she said in an interview with Rest of World. 'We end up being everybody's back yard.'
This article was written in partnership with The Intercept Brasil and in collaboration with the Pulitzer Center's AI Accountability Network. Read it in Portuguese here

Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

EXCLUSIVE Starling launches AI chatbot Spending Intelligence in UK banking first
EXCLUSIVE Starling launches AI chatbot Spending Intelligence in UK banking first

Daily Mail​

timean hour ago

  • Daily Mail​

EXCLUSIVE Starling launches AI chatbot Spending Intelligence in UK banking first

Finding out how much money you spent on coffee last month will soon become a faster exercise than trawling through bank statements or paying for a budgeting app to work it out. That's because Starling Bank is ramping up its use of Artificial Intelligence for current account customers. It has become the first bank in the UK to integrate its very own Large Language Model (LLM) AI chatbot within its banking app. The chatbot, called Spending Intelligence, uses LLMs to answer questions customers pose to it about their spending. Starling's chatbot is powered by Gemini, Google's LLM and answer to OpenAI's ChatGPT, a move which made sense for Starling as the bank houses its tech with Google Cloud Platform. It's available to all customers, both personal and business accounts - sole and joint - from today. Spending Intelligence appears as a search bar and sits at the top of Starling's legacy 'spending' area within Starling's banking app. The spending area gives customers information about how much they spend, where they spend it and when. It can be split up into more than 50 categories which customers can customise. Starling's AI chatbot takes the spending area a step further by providing answers to customers about their spending within a specific period of time, for example within the last month or year, and showing them a breakdown. Customers can ask it questions like 'how much did I spend on groceries last week?' or 'how much did I donate to charity last year?' These questions can either be typed into the search bar or customers can ask the chatbot verbally. Spending Intelligence can only draw insights from Starling accounts and gives answers based on this data. The chatbot cannot tell customers what to do with their money - for example where to invest it or how to save more money - as this would fall into the remit of regulated advice. But the idea is to get customers to think about how they are spending. The hope is that it will help customers to see exactly where their money goes The Starling chatbot is free to use and chief information officer Harriet Rees says there are 'absolutely no plans to make it paid for use'. More AI coming down the line for Starling The Spending Intelligence chatbot is the first step towards a bigger ambition to implement AI within its banking app. The hope is it will nudge customers towards using more of its existing money management features as well as new tools Starling develops in the future. Rees says: 'Customers can use AI to feed their natural curiosity about their finances so that they can make informed decisions about their budgeting, and better utilise Starling's suite of money management tools.' While customers will learn about their spending habits by using the chatbot, at the same time Starling will be learning from its customers through the questions they ask to identify areas where it can roll out more AI in the future. 'This is the first step for us of putting AI in the hands of customers' Rees tells This is Money. 'But there's so much further we could go' she adds. These could include fraud detection and new customer onboarding according to Rees. The information Starling receives from customers through the questions they ask cannot be used by Google's Gemini to feed its Large Language Models, this is part of Starling's agreement with Google. Starling said: 'Customers have to opt in to use Spending Intelligence. It does not store sensitive information and customers can opt out at any time.' In the future Spending Intelligence could potentially evolve into telling customers how they could save money if they were to ask the chatbot to identify areas. But for the moment it is focused solely on spending. Rees says: 'We're having a think about how you could save. But we genuinely believe giving customers insights about how they are spending is the best way to get them thinking about where they could save.'

Companies considering 2 gigawatt of data centers near NY hydropower sites, power authority says
Companies considering 2 gigawatt of data centers near NY hydropower sites, power authority says

Reuters

time5 hours ago

  • Reuters

Companies considering 2 gigawatt of data centers near NY hydropower sites, power authority says

NEW YORK, June 9 (Reuters) - Companies are considering building 2 gigawatts of data centers near New York Power Authority hydropower facilities, New York Power Authority (NYPA) CEO Justin Driscoll said on Monday at a conference. Technology companies are spending tens of billions of dollars this year alone to build and power their energy-intensive AI data centers, which are expected to propel U.S. power use to record highs this year and in 2026. Prospective data centers near NYPA's hydropower plants would be on top of the 2.5 gigawatts of data centers, and other very large energy users, that New York's grid operator estimates will come online by 2035. 'Many may not materialize, but there is no question that we need more generation and transmission to attract and serve these companies,' Driscoll said at an event during New York Energy Week, without specifying any particular company. One gigawatt of electricity is enough to power about 800,000 homes.

Barclays to roll out Microsoft Co-Pilot to 100,000 staff
Barclays to roll out Microsoft Co-Pilot to 100,000 staff

Finextra

time5 hours ago

  • Finextra

Barclays to roll out Microsoft Co-Pilot to 100,000 staff

Barclays Bank is to extend its delivery of Microsoft Co-Pilot AI agents to 100,000 staff worldwide. 0 Microsoft 365 Copilot will be integrated into the bank's own staff productivity tool, creating a single agent for employee self-service such as booking business travel, checking policy compliance, or finding answers to HR-related questions. An agentic dashboard accessed by Microsoft Viva will be deployed to help staff with 'moments that matter', like booking desks or annual leave, as well as serving personalised news and announcements. Craig Bright, Group CIO and deputy Group co-chief operating officer, Barclays, says: 'At Barclays, we've been leveraging the power of AI, and now GenAI, to drive deeper insights, improve efficiency and create more intuitive experience across the organisation. 'Our roll-out of Copilot, integrated with our colleague productivity tool, is a significant step forward in simplifying the way we work, making it easier to get things done." The agreement follows a successful initial deployment of Microsoft 365 Copilot across 15,000 Barclays employees and builds on the bank's adoption of Microsoft Teams as its preferred collaboration platform, and the deployment of Microsoft Viva Engage as its employee community tool.

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into the world of global news and events? Download our app today from your preferred app store and start exploring.
app-storeplay-store