AI Development Is Increasing Water Consumption and Contributing to the Global Water Crisis

• Recent reports suggest that the drastic AI growth is coming at the cost of inflating the water crisis.
• An unfathomable amount of water (gallons and gallons of it) is used to cool high-performing hardware that powers AI systems globally.
• The bigger concern is that a lot of this water is coming from areas with lower levels of freshwater.

Artificial Intelligence (AI) might have a lot of benefits, such as improved efficiency, better insights into decision-making, fraud detection, and better customer service, but it comes at a hefty cost.

While sustainability has always been a huge AI-related issue, given the technology's growing carbon footprint and massive energy consumption, a brand new concern has emerged: water scarcity.

Gallons and gallons of water are used in AI firms to cool the high-performing hardware. Approximately, 9 litres of water evaporate per kWh of energy used.

Although a lot of these AI firms use closed-loop systems to recycle water, a huge part of it is lost due to the need for humidity control, especially in dry regions. Plus, humid air can't be avoided either because it's necessary to prevent the air from becoming a conductor of static electricity - if that happens, it might damage sensitive equipment.

Virginia, which has the largest concentration of data centers in the world, noticed a drastic surge in water usage. From 2019 to 2023, water usage increased by nearly two-thirds, rising from 1.13 billion to 1.85 billion gallons.

What's even more concerning is that much of this water comes from areas with water scarcity. For example, Microsoft reported that 42% of the water it had used in 2023 came from areas with water stress."

Similarly, Google, which has among the largest data center footprints, also reported that 15% of the water it used came from areas that have a scarcity of freshwater.

At a time when a huge chunk of the world is suffering from a water crisis, such blatant wastage of water should be unacceptable.

According to reports, at least two-thirds of the world's population struggle with water shortage for at least one month out of the year. This means that approximately 4 billion people are having to deal with water issues while we wait with glistening eyes for Apple Intelligence to help us write birthday emails fluently.

Even worse, around 842,000 people die of water-related diseases (due to the unavailability of fresh water) every single year, out of which 361,000 are children.

What Can Be Done to Fix AI's Water Consumption Issue?

Servers that run AI algorithms generate more heat than classic computing systems, as a result of which more water is required to cool them down.

For example, a 5-50 prompt conversation with ChatGPT requires 500 ml of water for cooling. So, if the companies can make more energy-efficient systems, water requirements will automatically go down.

At the time of writing, a permanent fix to this problem hasn't appeared on the horizon. What's even more interesting is that this is a double-edged sword.

On one hand, there's a growing scarcity of water, but on the other, using water as a coolant is important to reduce greenhouse emissions.

It looks like the industry needs to come up with a way to efficiently cool down AI systems without water. Moreover, if they can develop a way to reduce heat generation, that can also help.

Amazon recently announced that it will be a good water steward" and its data centers will be water-positive by 2030. While it's undoubtedly good to see a market leader take such an initiative, a more industry-wide solution needs to be devised and implemented as soon as possible to solve the water crisis for good.

Related: AI will boost copper demand by 1 million tons per annum by 2030

The post AI Development Is Increasing Water Consumption and Contributing to the Global Water Crisis appeared first on The Tech Report.