I’m really excited for today’s EcoMinutes podcast since the topic was suggested to me by a listener. He asked that I discuss Artificial Intelligence, or AI. You can’t listen to the news or pick up a newspaper without there being an article about AI, and if you’re here to learn about AI going rogue and taking over the world, you’re at the wrong podcast. I’ll be discussing the environmental effects of AI and, in particular, the big data centers that are needed to support it.
So, first, what is AI?
According to the United Nations Environmental Program (UNEP), which did a recent study on the subject, “AI is a catch-all term for a group of technologies that can process information and, at least superficially, mimic human thinking. Rudimentary forms of AI have been around since the 1950s. But the technology has evolved at a breakneck pace in recent years, in part because of advances in computing power and the explosion of data, which is crucial for training AI models.”
RELATED: Escambia and Santa Rosa Counties ban data centers
To be fair, AI has some valuable environmental benefits. According to the Yale School of the Environment, AI can enhance energy efficiency and reduce energy usage, assist in environmental monitoring and management, assess life cycle impacts of products, and support process and supply chain optimization to minimize environmental impacts.
Sounds great so far, but unfortunately there’s more to be considered, in particular, the direct and indirect environmental costs of AI. When it comes to the environment, there is a negative side to the explosion of AI and its associated infrastructure, according to a growing body of research. They are large consumers of water, which is becoming scarce in many places. They use massive amounts of electricity, spurring the emission of planet-warming greenhouse gases. They rely on critical minerals and rare elements, which are often mined unsustainably. They use a lot of land, and they’re noisy. An enormous amount of resources are used in building the data centers and in the production, use, and disposal of electronic equipment. These are some direct effects, and I’ll go through them individually.
Data centers use water during construction and, once operational, to cool electronic components since water evaporation is one of the most efficient ways to cool down the data center facilities. Globally, AI-related infrastructure may soon consume six times more water than Denmark, a country of 6 million, according to one estimate. That is a problem when a quarter of the world’s population already lacks access to clean water and sanitation, and many data centers are located in water-stressed regions.
In the US alone, data centers account for 2-3% of public water use, with projections rising to 8% by 2030, with approximately 80-90% coming from potable water sources.
Even a mid-sized data center consumes as much water as a small town, while larger ones require up to 5 million gallons of water every day—as much as a city of 50,000 people.
Powering and cooling these centers take an enormous amount of electricity. A conventional data center—like cloud storage for your work documents or streaming videos—draws as much electricity as 10,000 to 25,000 households, according to the International Energy Agency. But a newer, AI-focused “hyperscale” data center can use as much power as 100,000 homes or more. Meta’s Hyperion data center in Louisiana, for example, is expected to draw more than twice the power of the entire city of New Orleans once completed. Another Meta data center planned in Wyoming will use more electricity than every home in the state combined.
The number of US data centers more than doubled between 2018 and 2021, and that number doubled again by 2025. Early in the AI boom, in 2023, US data centers consumed 176 terawatt-hours of electricity, roughly as much as the entire nation of Ireland (whose electric grid is itself nearly maxed out, prompting data centers there to use polluting off-grid generators), and that’s expected to double or even triple by 2028.
The problem is these data centers are contributing to the rise in demand for power and growing faster than both grid capacities and growth in renewable energy. This delays retirement of polluting fossil fuel plants and contributes to the continued rise in global emissions and climate change.
And data centers require land. Lots of it. Some of the largest data centers being built today will cover hundreds of acres with impermeable steel, concrete, and paved surfaces—land that will no longer be available for farmland, nature, or housing—and require new transmission lines and other associated infrastructure as well.
Another problem with data centers is the 24/7 noise. This stems from all the equipment that can produce noise levels of up to 100 dB (A)s. Since most of this equipment is located outside on the data center’s roof, this allows for low-frequency sound waves to travel far and wide to impact nearby communities.
According to the World Health Organization, this low-frequency humming noise affects people’s health. Although you may not be cognitively aware of the sound, your brain is, and it’s constantly processing the noise.
Over time, long-term exposure to low-frequency noise can lead to:
- Cognitive impairment
- Sleep disturbances
- Stress
- Anxiety
- Increased risk of cardiovascular disease.
Last, resources and electronic waste. Most large-scale AI data centers rely on a lot of raw materials. The data centers produce electronic waste, which often contains hazardous substances, like mercury and lead. And the waste is being produced five times faster than our ability to recycle it, becoming the fastest-growing waste stream on the planet. I intend to do another podcast on electronic waste in the future.
A lot, although not all, environmental impacts of data centers can be mitigated – unfortunately, governments don’t insist that the big tech companies take steps to decrease their environmental damage.
Construction of these facilities can be more sustainable using fewer polluting choices, fewer impervious surfaces, and renewable options. Their locations can be chosen to minimize the negative impact on ecosystems, protect existing natural habitats, and promote diversity of species within and around them by conducting early-stage ecological surveys to identify protected species, habitats, and ecological corridors, as well as engaging landscape architects.
Data centers can further minimize their environmental footprint by implementing circularity practices in which all materials are reused, reusable, or recyclable, with zero output to landfill or incineration, and waste heat recovery and reuse.
Data centers’ water usage can be addressed with overall water neutrality, achievable with different cooling technologies, appropriate water reduction and reuse strategies, such as closed-loop systems or air cooling and use of additional water resources such as rainwater, and site location in cooler climates. Data centers can also be asked to invest in local systems instead of competing with them, by sponsoring water efficiency, building water desalination plants, and restoring wetlands.
These centers can minimize their energy consumption from the grid by building their own renewable on-site power plants, and they can be asked to finance improvements to the grid.
Noise reduction can be achieved with better site location and acoustic screens and barriers. Generators should be a last resort and used for short times with the least polluting fuels.
What you can do:
There’s a growing bipartisan movement to ban data centers; you can join a group who is opposed or sign a petition. Both Escambia and Santa Rosa counties have bans in place, and the state has a new law, SB 484, that establishes regulations for hyperscale and large-scale data centers. Let your state legislators know your opinion. If you find there’s a data center being planned, which cannot be stopped, ask your local government to insist that they use the best mitigation for the worst of the environmental impacts.
But the best way to slow down construction of data centers and their resource use is to use less AI.
A single generative query consumes four to five times the energy of a standard web search, and in case you hadn’t noticed, it’s almost impossible to do a web search without an AI response being your first answer. If you don’t want AI, just add “-AI” to your query.
Do you really need AI writing your emails? Generating just two emails could use as much energy as a full charge on the latest iPhone. And according to a Pew Research Center study, that 100-word email could use up a whole bottle of water for the cooling that’s needed at these data centers.
References:
