Google’s water consumption is increasing rapidly as it dives deeper into the energy-intensive world of AI. Last year, the company used a staggering 5.6 billion gallons of water, which is 20 percent more than in 2021. This rise can largely be attributed to Google’s growing AI efforts, where training algorithms in massive data centers demands huge amounts of energy and water for cooling.
Surprisingly, most of the water they use is clean enough to be drinking water, but this doesn’t make the situation any less concerning. Water scarcity is becoming a serious global issue, and it’s particularly worrying for the West Coast, where Google is headquartered.
In response to water shortages in some areas, Google’s planned data center in Arizona had to switch to “air-cooled technology.” However, Google claims that 82 percent of its freshwater use last year came from regions with low water stress.
This water consumption issue isn’t unique to Google alone. Other tech giants like Meta and OpenAI are also using vast amounts of water to keep their data centers running, which is a hidden cost behind the ongoing AI race.
For instance, Meta’s latest AI model, Llama 2, doubled the company’s water intake compared to the previous model and while the situation is concerning, Google has expressed its commitment to improve and address this water consumption problem.
“[The increase in water consumption] was due to business growth, and aligns with other activity-based data,” a spokesperson told Gizmodo. “We’re working to address the impact of our water consumption through our climate-conscious data center cooling approach and water stewardship strategy, including our 120 percent replenishment target.”