What the People Have Been Saying:
I first noticed that AI had a negative impact on the environment from a TikTok post saying how one AI question takes the same energy as running 14 light bulbs for 1 hour. I wasn’t sure if this conversion was right, so I looked into it. With, of course, the power of AI because now Google generally has AI generated answers when you search anything. I learned that AI uses 2.9 Wh, Watt per hour, on average (www.euronews.com). Which is not the same amount as 14 light bulbs, definitely not LED and not Fluorescent either. Asking AI one question is the same as running an LED light bulb for 20 minutes, and a Fluorescent bulb for about an hour. Though, AI queries use about 10 times more energy than a normal Google search. But this is explanatory as AI usually gives you much more information. In a way, if you were to Google a question and then have to continue to search for a more in depth answer – you’d be spending the same amount of energy on a computer.
AI is a pretty helpful tool, and I don’t understand why the fault of how much energy it uses is being pushed on the consumer. This is generally the case with a lot of our environmental concerns. When in reality, the producers should be looking into more sustainable ways of running things.
Why is AI Hurting the Planet?
Well, AI data centers take a lot of freshwater to cool down their servers as they generate a significant amount of heat. The amount of freshwater we use has become an issue over the years depending on location and population. According to asce.org, most of the water that is used to cool the servers at data centers is used from municipal or regional water utilities(/www.asce.org) Which raises a concern for the local communities within the radius of the data centers, the larger they are creates more of a concern. For instance, in Austin Texas where the city is already under a stage 2 water restrictions due to drought issues, one hyperscale AI Data center is said to use 5 million gallons a day (https://www.youtube.com/watch?v=OaLIqqBfelU) . This amount can vary based on the cooling methods used within the data center. This has been a concern for Austin’s government recently as they try to think of ways to reduce these issues. Though, the water that is used to cool down these servers can be recycled within a closed loop system but that doesn’t change the fact that so much water is used compared to one home’s consumption of 154 gallons. 32,500 homes would use the same amount of water a day to consume 5 million gallons a day. Which is a similar number within the amount of electricity a hyperscale data center uses a day.
AI data centers use a lot of energy of course. 4% of the nation’s electricity is used to generate the AI data centers nationwide, and this is expected to rise to 9% in 2030. (www.epri.com) It is estimated that average data centers have a daily energy consumption of 1,000 kWh to 10,000 kWh. 30 to 40 average homes would need 1,000 kWh of electricity on the daily. Though, on the extreme side hyperscale AI data centers use over 1,000,000 kWh per day which according to my math …
Would be the same amount as providing electricity to 33,333 homes a day. And the United States has 54% of the world’s hyperscale data centers, so we are using quite a bit of electricity globally around here. (www.datacenterfrontier.com)
But think about it, these are super computers. Of course they are going to take a crazy excess of power. Though, we should find new ways of producing electricity that are more sustainable than the processes we use today. And by we, I mean the super smart people that were able to create these data centers in the first place. If they are able to create these, they are able to think of more environmentally friendly ways to run them.
What are the Data Centers Doing to be More Environmentally Friendly?
Thankfully the big AI data centers are looking into more environmentally conscious ways to fulfill the centers electric needs. Meta had released a report last year that they are teaming up with Sage Geosystems to work with geothermal based electricity. The Earth’s center will never run out of heat!
Geothermal energy uses the Earth’s natural heat that produces steam under the Earth’s service to power a turbine which is connected to and powers a generator to produce electricity. Geothermal energy is one of our less explored ways of electricity, but it is renewable, pollution free, and consistent compared to our more popular ways of energy like fossil fuel.
Mainly California, Nevada, and Utah use geothermal energy at some of their AI data centers at this time, but Meta is looking into expanding these practices with their other centers across the united states. Sage’s technology has made significant advancements within the clean energy sector – showcasing the ability to harness geothermal energy virtually anywhere and promising a new era of reliable, sustainable baseload power and enhanced grid stability (https://about.fb.com) . Meta has contracted more than 12,000 megawatts in renewable energy projects, making them one of the largest corporate buyers of renewable energy globally (https://about.fb.com).
So, Should We Continue Using AI?
Honestly, yes. Just be mindful of what type of questions you may ask AI. At the end of the day, we don’t need to rely on AI to answer all of our questions anyway. I have been at fault for this by asking AI some silly questions about recipes or a quick math question that my phone calculator would have no issue completing. But I’ve also asked AI some really helpful questions that have saved me a lot of time that I may have used to research instead.
At the end of the day, the environmental concerns with AI shouldn’t fall on the consumer – but on the producers and engineers at data centers. This is a common scapegoat for a lot of our environmental concerns that the consumers need to ‘do better’, or ‘use less’ which yes I do agree but not to the extent that it is presented. It is always good to be aware of how much energy is used when doing anything as we are living in a world where most things we do or use seems to come out of thin air. Everything we do has an impact on the world around us, but the effect doesn’t fall solely on the consumer, the producers need to be influenced to spend more money and time on environmentally friendly ways to move forward.
Citations:
Engineers often need a lot of water to keep data centers cool by Rasheed Ahmad (April, 2024) https://www.asce.org/publications-and-news/civil-engineering-source/civil-engineering-magazine/issues/magazine-issue/article/2024/03/engineers-often-need-a-lot-of-water-to-keep-data-centers-cool
U.S. Contains Fully Half of 1,000 Hyperscale Data Centers Now Counted Globally, as Cloud Giants Race Toward AI by Matt Vincent (April, 2024)https://www.datacenterfrontier.com/hyperscale/article/55020441/us-contains-fully-half-of-1000-hyperscale-data-centers-now-counted-globally-as-cloud-giants-race-toward-ai
EPRI Study: Data Centers Could Consume up to 9% of U.S. Electricity Generation by 2030 (May, 2024) https://www.epri.com/about/media-resources/press-release/q5vu86fr8tkxatfx8ihf1u48vw4r1dzf
ChatGPT, Deepseek & Co: How much energy do AI-powered chatbots consume? By Inês Trindade Pereira (March, 2025)https://www.euronews.com/my-europe/2025/03/17/chatgpt-deepseek-co-how-much-energy-do-ai-powered-chatbots-consume
New Geothermal Energy Project to Support Our Data Centers (August, 2024) https://about.fb.com/news/2024/08/new-geothermal-energy-project-to-support-our-data-centers/#:~:text=Geothermal%20energy%20is%20a%20viable,buyers%20of%20renewable%20energy%20globally.
Go Green Girl 
Leave a comment