Data centers powering artificial intelligence are generating massive amounts of waste heat that's warming neighborhoods, straining electrical grids, and sparking pushback from over 140 local groups who have blocked or delayed more than $60 billion in data center investments in roughly a year. As cities including Denver, Minneapolis, and Seattle consider restrictions on new facilities, environmental experts are weighing the real costs and trade-offs of AI's explosive growth.

How Much Heat Do Data Centers Actually Produce?

Nearly all electricity consumed by a data center converts directly into heat, requiring continuous cooling systems that release that energy into surrounding neighborhoods. Research conducted in Phoenix, Arizona, found that a single large data center can emit waste heat comparable to that produced by tens of thousands of households.

The temperature effects are significant and measurable. Studies show that data center waste heat can increase surrounding land surface temperatures by as much as 16 degrees Fahrenheit and raise air temperatures in nearby neighborhoods by up to 4 degrees Fahrenheit. These warming effects have been observed as far away as half a kilometer, or about five city blocks from the facility.

"Because nearly all of the electricity consumed by a data center ends up as heat, servers and other equipment require continuous cooling. These cooling systems remove the heat generated by computing and release it into the surrounding environment as waste heat," explained Karen Seto, the Frederick C. Hixon Professor of Geography and Urbanization Science and Director of the Hixon Center for Urban Sustainability at Yale School of the Environment.

Karen Seto, Frederick C. Hixon Professor of Geography and Urbanization Science, Yale School of the Environment

The concern is particularly acute in neighborhoods already experiencing the urban heat island effect, where cities are typically 1 to 7 degrees Fahrenheit warmer than surrounding areas. Additional warming from data center waste heat can further increase air conditioning energy demand and human health risks in these communities.

What's Driving the Explosive Growth of Data Centers?

The primary driver of data center expansion is the race to train better artificial intelligence models, particularly large language models (LLMs), which are AI systems trained on vast amounts of text data to generate human-like responses. Beyond training, data centers also run generative AI queries and store cloud data for billions of users worldwide.

Some large data centers under construction are slated to use 3 gigawatts or more of peak electricity, which is more than 75 times Yale University's peak demand. This unprecedented growth is creating substantial anxiety about generative AI in public polling and political pushback from some lawmakers.

"The main driver of data center expansion today is the race to train better AI models. In addition to training large language models, data centers are also used to run generative AI queries. They are also used simply to store all of the data that we all have in the cloud. But the primary driver of the growth really is generative AI," stated Kenneth Gillingham, the Grinstein Class of 1954 Professor of Environmental and Energy Economics at Yale School of the Environment.

Kenneth Gillingham, Grinstein Class of 1954 Professor of Environmental and Energy Economics, Yale School of the Environment

Electricity prices are already rising and are likely to continue increasing in the near future. While higher electricity prices can help accelerate the adoption of renewable energy, the new electricity demand has largely been met with natural gas generation so far, which means increased greenhouse gas emissions.

What Are the Real Environmental Impacts Beyond Heat?

The environmental footprint of data centers extends far beyond the waste heat they generate. The impacts are highly context-dependent and include several interconnected factors:

• Electricity Use and Emissions: The most visible impact is electricity consumption, which translates into greenhouse gas emissions depending on when and where the power is consumed. A data center located on a coal- or gas-heavy electrical grid can have a vastly different carbon footprint from one using low-carbon electricity sources.
• Water Consumption: Water use is a major concern, especially for cooling systems and for electricity generation upstream. The severity of this impact depends greatly on whether a facility is located in a water-stressed region where water scarcity is already a problem.
• Manufacturing and Supply Chain Impacts: Data centers require servers, chips, cooling equipment, buildings, backup power systems, and grid infrastructure, all of which carry embodied environmental impacts from manufacturing, construction, and eventual replacement.

"The environmental impacts of data centers are real, but they are also highly context-dependent. The most visible impact is electricity use, which can translate into greenhouse gas emissions depending on when and where the power is consumed," noted Yuan Yao, the Manufacturer's Association Professor of Industrial Ecology and Sustainable Systems at Yale School of the Environment.

Yuan Yao, Manufacturer's Association Professor of Industrial Ecology and Sustainable Systems, Yale School of the Environment

How to Minimize Data Center Environmental Impacts

• Facility Design and Neighborhood Cooling: Cities and industry leaders should collaborate on better facility design and neighborhood cooling strategies, including the installation of parks and reflective surfaces that can help dissipate heat more effectively.
• Energy Efficiency and Improved Cooling: While efficiency gains remain important, they can be offset if total electricity demand grows rapidly. Continuous innovation in cooling technology and server efficiency is essential to slow the growth of energy consumption.
• Clean Electricity Procurement: Data centers should prioritize sourcing electricity from renewable and low-carbon sources, evaluated by location and time of use rather than just annual averages, to ensure the power actually comes from clean sources when the facility operates.
• Transparent Reporting and Life Cycle Management: Solutions need to consider siting and infrastructure decisions, operational strategies, and the full life cycle management of hardware and equipment. Transparent reporting and metrics that capture carbon, water, reliability, and life cycle environmental impacts together are critical for accountability.

Experts emphasize that the solutions that work best are not single fixes and may differ across locations, technologies, and operating models. A comprehensive approach considering all environmental dimensions is necessary to ensure that AI's growth benefits society without overwhelming local communities or global environmental systems.

What Does the Broader AI and Sustainability Movement Look Like?

Beyond the United States, the conversation about AI and environmental sustainability is gaining momentum globally. In Hong Kong, the TVB Green Summit 2026 brought together government, business, technology, and research leaders to explore how AI accelerates the green transition and drives sustainable economic development.

Industry leaders at the summit discussed how enterprises can leverage green data centers and smart energy systems to balance AI computing demands with carbon reduction goals. Companies like Tencent Cloud and Tianqi Lithium are working on strategies to manage the tension between growing AI infrastructure needs and environmental responsibility.

The broader message from experts is clear: society needs a more clear-eyed approach to both the potential benefits and the real costs of AI expansion. Only with ongoing policy-guided investments and concerted effort across government, industry, and research institutions can we be confident that AI will move society toward a sustainable future rather than accelerating environmental degradation.