Data centers can't come at the cost of community health and climate solutions

Image
A graphic of a data center with a blurred image of a forest in the background

Key Takeaways

  • Data centers affect where electricity loads grow and how climate solutions in the electricity sector are adopted.
  • The most critical environmental impacts of data centers are localized.
  • Privately powered data centers threaten both climate and communities. 

You’ve likely heard that artificial intelligence is changing the landscape of work. 

But you may not have seen how the facilities that provide the infrastructure for AI are literally changing the landscape unless you’re in a corridor of data center development. Data centers provide storage and computing power for the technologies we use in our daily lives, including those deemed “AI.” Other than electric utilities themselves, these tech facilities have become the most important single entity shaping the development of the North American power grid. They comprised less than 2% of our electricity use in the United States just a decade ago; now they’re almost 5%, and projections consistently point to data centers as the fastest-growing slice of that pie.

Here, we’ll look at data centers broadly, encompassing a variety of computing resources. We’ll also focus on data center growth in the United States, where I’m based, and where the International Energy Agency shows that data centers use more electricity than in any other country, though growth in other countries follows similar patterns.

Data centers have two big environmental impacts. First, they use a lot of power, with unique patterns of power usage based on the type of computing taking place. Second, the computer chips they hold require extensive cooling, and the most cost-effective way to cool them is to use a lot of water. And that’s in addition to the water the power generation facilities use to produce the electricity data centers use.

Data centers affect where electricity loads grow and how climate solutions in the electricity sector are adopted.

The electricity sector is preparing for more rapid load growth than anything we have seen since the 20th century. And they’re doing so with aging infrastructure, delays in connecting new resources, and the need to navigate a changing climate, which brings challenges such as weather extremes and reduced water supply. Keeping costs low and reliability high for consumers is a top priority, even where efforts focus on increasing the use of clean energy resources.

Companies site data centers based on their financial interests and in locations where it’s politically possible to stand them up quickly, not on where the most environmentally friendly power is generated. In fact, data centers tend to be located in places that have more climate pollution per kWh of electricity generated than the national average, and in places already experiencing water stress. 

Image
A graph showing the growth and projected growth in electricity consumption from data centers by region

On top of that, the water impacts are borne by local communities, who may see their water supplies dwindling or water pollution increasing.

Data center growth shouldn’t take priority over the growth of climate solutions and the rapid displacement of fossil-fuel pollution from electricity generation. But that means data center development must change. Data centers should be transparent about and accountable for their water use, energy use, and greenhouse gas emissions tied to that energy use. Moreover, regional policies should guide data center rules, because the impacts go beyond individual cities and are more complex than any single project.

Privately powered data centers threaten both climate and communities.

While most data centers built to date have been connected to local utilities, as is typical of most infrastructure, bypassing the grid to build faster is a growing trend. Twenty-five percent of new data center capacity in the United States is planned to generate its own power rather than relying on utility generation.

...we should first and foremost be asking: what kind of future do we really need?

Tech companies powering data centers on their own is a much bigger concern than data centers that cooperate with local utilities. Newly created power generation facilities for data centers, particularly those that emerge outside a central planning process or don’t require integration with broader electric utilities, like Microsoft’s Chevron-fuel Texas data center, are mostly powered by natural gas. Most of the new electricity generation capacity we’re adding in the United States is renewable, but our country’s on-site power facilities for data centers run on natural gas, reversing climate progress in the electricity sector.

Because the utilities, system operators, and regulators responsible for developing the power grid are already stressed about meeting their goals for providing reliable electricity while demand grows, and because the system doesn’t move at Big Tech speeds, many tech companies want a way around the typical queues. But when companies are in a hurry, with large amounts of money at their disposal to quickly stand up data centers without accountability to those who bear the environmental price of their infrastructure, the health of local communities suffers.  

Importantly, the health of the grid can suffer, too: For electricity sector climate solutions to grow, we need coordinated planning that incorporates load growth projections, energy storage, demand flexibility, and incentives to cut emissions. Planning for power demand from buildings, data centers, and other industries together gives us more options to integrate climate solutions and enabling technologies such as batteries that we already have in hand. Transforming the electric grid into the clean energy machine we need and deserve won’t happen on private grid islands owned by tech companies. It will emerge from coordinated efforts by electricity grid operators, regulators, utilities, and policymakers. Data center developers will say they’re building the future, but we should first and foremost be asking: what kind of future do we really need?


About the Author
Amanda D. Smith, Ph.D., is a senior scientist at Project Drawdown with expertise in building science and distributed energy systems.

About Project Drawdown 
Project Drawdown is the world’s leading guide to science-based climate solutions. Our mission is to drive meaningful climate action around the world. A 501(c)(3) nonprofit organization, Project Drawdown is funded by individual and institutional donations.

This work is published under a Creative Commons CC BY-NC-ND 4.0 license. You are welcome to republish it following the license terms.