Technology

Google Commits to Replenishing More Water Than It Uses by 2030

Martin HollowayPublished 4d ago4 min readBased on 8 sources
Reading level
Google Commits to Replenishing More Water Than It Uses by 2030

Google Commits to Replenishing More Water Than It Uses by 2030

Google has set a target to replenish 120% of the water it consumes by 2030—that is, to put back more water than it takes out. The company is aiming to generate more than 19 billion gallons of annual water replenishment, which is enough to supply Los Angeles with drinking water for over 40 days. This commitment comes as Google's data centers are using significantly more electricity, up 27% in 2024, largely because of artificial intelligence workloads that require more cooling.

Google's plan involves 165 water restoration projects across 97 watersheds (areas where water drains into rivers and streams) around the world. In 2024, the company replenished 4.5 billion gallons of water and increased its freshwater replenishment rate from 18% in 2023 to 64% in 2024.

Going Beyond Simple Efficiency

Google's 120% replenishment target is ambitious because it means the company will contribute more water to the environment than it removes. Most companies simply try to use less water—Google is trying to use less and also restore more.

As of late 2024, Google was running 112 active water stewardship projects. These range from teaching farmers how to irrigate more efficiently to restoring wetlands that naturally store water and help prevent flooding. In Chile, Google is working with an agricultural technology company to help local farmers use less water when irrigating their crops. In the Netherlands, it is funding structures that will create marsh areas suitable for both wildlife and certain types of farming, while also reducing flood risk. In the southwestern United States, Google is funding solar-powered irrigation systems for the Quechan Tribe to provide more reliable water access during droughts. In Belgium, the company is helping restore a nature reserve's wetlands, which improves both water storage and flood resilience.

How Big Is the Water Problem

Data centers cool their equipment using water. As artificial intelligence workloads have grown, data centers have needed more cooling, so they are using more water. Google notes that all U.S. data centers together use less than 1% of the water Americans use to water their lawns each year. However, this overall statistic hides an important issue: data centers often concentrate their water use in specific cities and regions that are already short on water.

For Google's data centers globally, 86% of the freshwater it pulls comes from areas where water is abundant. This suggests Google has been careful about where it builds. But as artificial intelligence becomes more powerful and demands increase, the company may need to build data centers in regions that have less water available, which could create real challenges.

While Google cut its data center carbon emissions by 12% in 2024 compared to 2023, it used 27% more electricity overall. This shows that the company is making its operations cleaner in some ways, but the growing demand for artificial intelligence is still pushing up overall energy and water consumption faster than efficiency improvements can bring it down.

The Power and Water Connection

Data centers need electricity to run computers and water to cool them. Google signed agreements in 2024 to buy over 8 gigawatts of clean electricity (enough to power roughly 8 million homes). The company made news by agreeing to buy nuclear power from small modular reactors—a newer type of nuclear plant that uses less water for cooling than traditional nuclear plants.

This shift reflects a pattern we have seen before. When the semiconductor industry expanded rapidly in the 1990s, companies realized they could not simply use less energy and water—they had to find new sources of power and new ways to cool their equipment. That industry ended up developing closed-loop cooling systems (where water is reused rather than dumped) and other technologies that became standard across the field.

Google's decision to spread its water restoration projects across nearly 100 different watersheds rather than focusing all of them near its data centers shows the company is thinking like it does with electricity: water availability is a system-wide problem, not just a local one. Just as companies can buy renewable energy from anywhere to offset their electricity use, they can invest in water restoration anywhere to help offset their water consumption.

The real question is whether Google's plan will succeed. The company will be working on water projects in very different parts of the world—some in agriculture, some in cities, some in natural wetlands. For a technology company used to controlling exactly what happens inside its data centers, managing water outcomes across rivers, farms, and ecosystems will be more complicated and less predictable.

The broader reason for this commitment is worth noting. As artificial intelligence workloads continue to grow, data centers will come under increasing pressure from regulators and communities in areas where water is scarce. By investing heavily in water restoration now, Google is positioning itself to operate in water-stressed regions and avoid the kind of public backlash that can delay or block construction. In that sense, water stewardship has become not just an environmental goal but an operational necessity for a company that needs to keep building and expanding.