JupiteX Get the app
Environment & Climate20 Nov 2026 · about 7 min

Big Tech’s thirst for water is sparking a global resistance

The brief

Data centers need water to cool the equipment that powers digital services, including generative AI. As these facilities grow and add more servers, their water needs can rise. Communities may oppose projects when the facilities compete for clean water that residents also need for drinking. The conflict becomes sharper in places already facing shortages. The article describes how powerful GPUs used for generative AI run harder and produce more heat than traditional data-center equipment. Cooling them takes more energy or water. Facilities also need clean water to avoid damage to equipment and pipes. Using public drinking water can be cheaper than treating water themselves, so companies may seek the same potable supplies local communities depend on. In Chile, residents of Cerrillos challenged a proposed Google data center during a long drought. Its projected use was 169 liters a second, while more than half of Chile’s population lived in areas with severe scarcity by 2021. The article ends with activists building their case against the company; it does not state the project’s final outcome. Such disputes show why water needs can become a central concern when data centers are proposed.

01

Why are communities around the world resisting Big Tech data centers over water?

Data centers need water to cool the equipment that powers digital services, including generative AI. As these facilities grow and add more servers, their water needs can rise. Communities may oppose projects when the facilities compete for clean water that residents also need for drinking. The conflict becomes sharper in places already facing shortages.

The article describes how powerful GPUs used for generative AI run harder and produce more heat than traditional data-center equipment. Cooling them takes more energy or water. Facilities also need clean water to avoid damage to equipment and pipes. Using public drinking water can be cheaper than treating water themselves, so companies may seek the same potable supplies local communities depend on.

In Chile, residents of Cerrillos challenged a proposed Google data center during a long drought. Its projected use was 169 liters a second, while more than half of Chile’s population lived in areas with severe scarcity by 2021. The article ends with activists building their case against the company; it does not state the project’s final outcome. Such disputes show why water needs can become a central concern when data centers are proposed.

02

How much water can a hyperscale data center use compared with a town?

A hyperscale data center can use between 1 and 5 million gallons of water each day, according to the article. That range is comparable to the amount used by a city of 10,000 to 50,000 people. The comparison depends on the part of the world, but it shows that one facility can place a substantial demand on local water supplies.

Cooling is a key reason for this demand. Data centers contain servers and other equipment that generate heat. The article notes that generative AI GPUs run harder and throw off more heat than traditional chips and storage devices, so they need more energy or water to stay cool. Companies also often cluster data centers together, making it harder to supply them.

The article does not give one universal figure for every facility; use varies with size and equipment. It emphasizes that demand is growing as data centers expand and add servers. In water-stressed places, facilities drawing clean water can compete with communities for potable supplies. The scale comparison helps explain why residents may scrutinize a project even before it begins operating.

03

What is a data center, and what happens inside it?

A data center is a building or facility that houses computer systems and related equipment. This basic definition uses general knowledge; the article focuses on the water needs of these facilities rather than defining them directly. Data centers matter because they provide the physical infrastructure behind digital services, and their equipment requires power and cooling.

Inside, servers process and store information. The article specifically points to graphics processing units, or GPUs, which power generative AI. These chips are more resource-intensive than the chips and storage devices that traditionally filled data-center racks. They run harder and produce more heat. The facility must use energy or water to keep that equipment cool and prevent damage.

The article says data centers need clean enough water to avoid damaging their equipment and pipes. It also notes that operators can find it easier and cheaper to use public drinking water than to pay for treatment themselves. As facilities grow and companies cluster them, their demands can become significant for surrounding communities. The article gives a range of 1 to 5 million gallons a day for a hyperscale center.

04

What happened to Google’s proposed data center in Cerrillos after residents organized?

Residents in Cerrillos formed MOSACAT, a community organization whose name means Community Socio-Environmental Movement for Water and Land. Members examined available information about Google’s proposed data center and what it could mean for their community. Their efforts came amid concern about the project’s projected water use and the area’s limited water supply.

Google’s proposed facility was a $200 million project on the southwest side of Santiago. Company projections put its water use at 169 liters every second, or as much in a day as all 80,000 Cerrillos residents use for drinking in a year. The project planned to draw water from an aquifer already subject to restrictions, which became a complete prohibition on new extraction shortly before environmental approval.

The article says local activists ultimately built the case against one of the world’s largest companies, but the supplied text ends before explaining what followed. It does not state whether the project was approved, canceled, changed, or constructed. The known account is that residents organized, examined technical reports, and raised alarms about water use during Chile’s prolonged drought.

05

How can data center water use affect local access to drinking water, especially during drought?

Data centers can affect drinking-water access when they draw from public potable supplies that residents also rely on. The article says operators need clean water to protect equipment and pipes, and that tapping public drinking water can be easier and cheaper than treating water themselves. That creates direct competition between facilities and communities, particularly where water is already scarce.

Cerrillos shows how large the demand can be. Google’s proposed data center was projected to use 169 liters of water every second. In a single day, it would consume as much water as all 80,000 local residents use for drinking in a year. The project also sought water from an underground aquifer under restrictions, just as Chile was facing a prolonged drought.

By 2021, more than half of Chile’s 19 million residents lived in areas experiencing severe water scarcity. Hundreds of rural communities relied on tanker deliveries of drinking water. The article does not quantify how much the proposed facility would have reduced residents’ supply, but the scale and source of its demand raised alarms. During drought, additional large users can sharpen concerns about who has access to limited clean water.

06

What alternatives can reduce a data center’s use of freshwater for cooling?

The article explains the problem but does not identify alternative cooling methods. In general, well-established options include using treated wastewater instead of drinking water, reusing water in a closed-loop cooling system, and choosing cooling designs that rely less on water. These approaches can reduce demand for freshwater, though their suitability depends on local conditions and facility design.

The article’s example shows why alternatives matter. Google’s proposed Cerrillos data center was projected to use 169 liters of water per second and planned to draw from an underground aquifer. Residents were already concerned about the area’s limited supply. The article also notes that clean water protects equipment and pipes, while public drinking water can be cheaper and easier for operators to access than self-treated supplies.

Chile’s drought and local scarcity made the proposed demand especially contentious. By 2021, more than half of the country’s residents lived in areas experiencing severe water scarcity. The source does not say whether Google considered or adopted other cooling methods for the project. More broadly, reducing freshwater use could ease pressure on communities, but the article gives no estimates of savings or details on specific alternatives.

07

What is an aquifer, and how can limits on pumping help protect it?

An aquifer is an underground body of permeable rock, sediment, or other material that holds groundwater and allows it to move. This definition comes from general knowledge; the article mentions an aquifer but does not define one. Communities and regulators may set limits on pumping to reduce pressure on this underground water source and help preserve available supplies.

Google’s proposed Cerrillos data center wanted to draw water from an underground aquifer. The article says the aquifer had been under a decree restricting how much water could be removed since 2005. Just weeks before the project received environmental approval, that decree was upgraded to a complete prohibition on new extraction. Residents were alarmed by the center’s projected water use of 169 liters per second.

The article does not explain the technical basis for the decree or detail how its restrictions would be enforced. It does place the dispute in the context of Chile’s drought, which had continued since 2010, and severe water scarcity affecting more than half the population by 2021. Limits on pumping can matter when a large proposed user seeks groundwater in a region with little water to spare.

This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.

Read more in the JupiteX app

Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.

Or read more news on the web