Protect Florida from Data Centers

Impacts of Hyperscale Data Centers

Background

Hyperscale data centers (HDCs) are rapidly proliferating in the United States. As of March 2026, there were 68 HDCs in the U.S., with an additional 267 planned (Greenfield 2026). They play an important role in modern digital infrastructure, including the storage and distribution of data, support services for cloud computing, artificial intelligence (AI), and e-commerce (Berger 2025, Guidi et al. 2026); however, the explosion of HDCs are setting off alarm bells regarding the environmental impacts. HDCs are those facilities with an average energy capacity of 50 MW or greater. Between 2024 - 2025, these facilities accounted for more than 2-3% of total US energy consumption (Cruzes 2025, Guidi et al. 2026) and consumption is expected to rise to 6.7-12% by 2028 (US Dept. of Energy 2024). Consumer Reports (Greenfield 2026) likens this to “adding a country with the energy needs of Spain in just three years”. To further illustrate the significance, the energy demand of a 100 MW HDC is roughly equivalent to the energy use of 100,000 U.S. households, according to the International Energy Agency (Berger 2025, IEA 2025b). Of the energy consumed by HDCs, about 60% of that is coming from the burning of fossil fuels (specifically natural gas and coal). Additionally, carbon emissions from HDCs in 2024-2025 are estimated to have been 52 million tons of CO2 (Guidi et al. 2026).

Modern large-scale data centers, including HDCs, are different from the data centers of the 2000s and 2010s, in that they now support highly complex workloads that result in a pulsed power demand, causing sudden spikes in power consumption and cooling needs on the timescale of milliseconds. These sudden spikes in energy usage and chilling needs on such a large scale can result in local electrical grid instability and high energy costs. It is estimated that 40% of the energy used by HDCs are from cooling alone (Cruzes 2025).

Policymakers and business leaders view HDCs as a strategic asset. They have the potential to bring in large amounts of revenue through investments and taxes and impact local job and housing markets. Local governments are often interested in HDCs and their operations as “digital sovereignty” - ensuring that data within their jurisdiction is secure (Cruzes 2025; Posey et al. 2026). Communities, however, tend to view them in a more negative light, focusing on the environmental impacts and limited revenues due to tax breaks.

Types of Cooling Systems

As mentioned previously, the energy demand of the cooling systems alone equates to approximately 40% of energy used by the HDC. There are several types of cooling systems used. These are a few:

Traditional Room-Based Cooling (CRAC/CRAH)

Computer room air conditioners (CRACs) and computer room air handlers (CRAHs) are the standard cooling systems of data centers. Both of these circulate conditioned air through cold-aisle containment and then return warm exhaust from hot aisles back for reconditioning. While these processes are different in their set-up, both rely on heat exchangers (coils or compressors) and water or cooling fluids (Cruzes 2025). 

Rear-Door Heat Exchangers (RDHx)

This is an air-to-liquid cooling system that is mounted to the rear of server racks and uses liquid-cooled coils to capture heat directly. Absorbed heat is then transferred to a coolant loop and eventually the heat is released through external chillers or dry coolers. Once cooled, air is returned to the rack for cycling. It is estimated that this method of cooling can cut energy demands 30-66% compared to CRAC/CRAH (Cruzes 2025).

Direct-to-Chip (D2C) Liquid Cooling

D2C does not rely on heat sinks or forced airflow, like the previous methods, but instead applies coolant directly to cold plates that are mounted on processors and accelerators. The coolant is then circulated through a closed loop that is in direct thermal contact with high-power chips and a liquid-to-liquid heat exchanger. It is estimated this method can cut cooling costs by 90% (Cruzes 2025).

Liquid Immersion Cooling

This technique fully submerges electrical components in a thermally conductive dielectric fluid, resulting in direct heat transfer. The coolant is moved to a bath and heat exchanger, in which heat is released, before being recirculated. This method is estimated to reduce energy consumption by 95% and reduce water consumption by 90% (Cruzes 2025). 

Evaporative Cooling

Direct evaporative cooling (DEC) brings in warm, dry ambient air and moves it through a wetted, evaporative pad, where latent heat is absorbed and air is cooled. The air then passes through the data hall to absorb heat before the now hot air is expelled, cooled, and recirculated. This method is still considered high water demand due to the wetted, evaporative pad (Cruzes 2025).

Environmental, Health, and Cost Concerns

Energy Use & Carbon Footprint

A conventional grid power is the most widely used source of energy by HDCs (Cruzes 2025). Because this relies on fossil fuels, primarily natural gas, it results in a large carbon footprint and greenhouse gas (GHG) emissions (Selvakumar 2026). In addition to the standard energy consumption, because these facilities need to be operational 24/7, they require backup generators - often running on diesel or natural gas generators and turbines. This not only contributes to the carbon footprint, but to noise pollution (discussed below). Not only are GHGs released by HDC facilities, but so is fine particulate matter (FPM). FPM is recognized as a severe health hazard, contributing to respiratory and cardiovascular disease. Additionally, the equipment, itself, is put through significant strain, necessitating regular replacement and repair. This requires extraction of rare earth minerals and other raw materials and a global supply chain. Mining for such materials further degrades ecosystems (Selvakumar 2026).

According to Consumer Reports (Greenfield 2026), energy costs rose in 2025 anywhere from 7.1-20% across states. This increase is more than twice that of inflation and has been attributed to both an aging power grid and the HDC boom. Virginia provides a stark example of how citizens can be impacted by the energy demand of HDCs. The state is home to more than 600 data centers, which pull electricity from the state or regional grid. A Bloomberg analysis found that in 2025, Virginia saw energy costs soar by 267% as compared to five years ago. Because of the demand on the grid and being a wholesale commodity, the increased cost has been passed on to households, with residents, on average, paying 80% more on their monthly electric bill than they did three years ago. 

Water Consumption & Pollution

HDCs generate a significant amount of heat, due to the densely packed servers and processors, which require continuous cooling (Selvakumar 2026). According to the Environmental Protection Agency (EPA), a large data center can use upwards of 5 million gallons of water per day for direct cooling - the equivalent of 16,000 U.S. households (Greenfield 2026, Sierra Club n.d.). In 2023, data centers in the United States used approximately 17.4 billion gallons of water with hyperscale facilities using 84% of that (Shehabi et al. 2024). According to the IEA, an average 100 MW HDC, may consume around 530,000 gallons of water per day (IEA 2025a). Similar to what was discussed previously with energy costs being passed on to households, the increased water demand and associated costs are also passed on to residents. 

Although there are cooling methodologies that reduce water consumption, CRAC/CRAH and evaporative cooling are the most common cooling techniques and both have a high water demand (Cruzes 2025). In dry and/or hot regions, this presents a greater sustainability challenge and often results in a greater carbon footprint (Selvakumar 2026). Furthermore, since 2022, about ⅔ of HDCs have been built in areas experiencing water stress due to drought (Greenfield 2026), much like is currently being experienced in Florida. 

While, in theory, closed loop and more modern cooling systems save on water usage and protect from water pollution, they are not 100% effective. As discussed, these systems rely on coolants that contain chemical additives, such as nitrite inhibitors, glycol antifreeze, PFAS, and heavy metals, like copper and zinc. As these chemicals move through the cooling system and evaporation takes place, the chemicals become more concentrated. There is risk of these chemicals being released both as a result of malfunction as well as normal systems maintenance, known as “bleeding the lines” (removal of air from the system). During the bleeding process, chemicals and contaminants can be released, and even in small amounts, the pollutants are many times higher than the limits set for surface waters (Hedge 2025).

Heat Island Effect

Studies have found that data centers act as heat islands, due the significant energy usage and generation, raising the surrounding ambient temperatures. This effect is not unique - the same phenomena is found in urban centers and industrial areas where vegetation is removed and replaced with cement and large, heat-absorbing surfaces along with heat-generating activities. But Marinoni et al. (2026) found that HDCs can actually increase temperatures of the surrounding area anywhere from 0.3-9.1oC (0.54-16.38oF) and the effect can extend up to 10 km (6.2 mi) from the facility. Another study, focusing on HDCs in Arizona, including a 36 MW data center and a 169 MW HDC, found something similar, though less dramatic. In this study, ambient temperatures increased by an average of 2.2oC (3.96oF) and the heat island effect was still detected at 500 m (0.3 mi) from the facility.

The heat island effect is of substantial concern. According to the EPA, urban centers already experience increased daytime temperatures of 1-6oF as compared to rural areas and up to 22oF higher nighttime temperatures. The construction of HDCs will add to this temperature increase. The heat island effect presents both a human health and environmental risk. It increases heat exposure, which increases respiratory and cardiovascular risks, like heat stroke and asthma, in vulnerable groups, including the elderly, children, and pregnant women. The increased temperatures also promote the formation of air pollutants - particularly ozone - further exacerbating respiratory and cardiovascular disease  (CalEPA 2026). The heat island effect is additive to the already increased global temperatures due to global climate change.

Noise Pollution

Noise pollution coming from HDCs and their use of back up generators and turbines are also a significant concern. A data center can emit between 50-98 adjusted dB, with 105 dB having been recorded. That noise will be most noticeable within 400 ft of the facility, but may be heard up to 3000 ft away. For reference, the Center for Hearing and Communication states that continued exposure to 70 dB will contribute to hearing loss over time (Sierra Club n.d., Yañez-Barnuevo 2026); however, noise over 85 dB is considered harmful to a person’s ears (Yañez-Barnuevo 2026). Noise pollution is also associated with mental health issues, such as anxiety, and learning challenges. Children are at increased risk of these disorders (Sierra Club n.d.). Infrasound, associated with data centers, is being linked with a variety of disorders and ailments, including head aches, sleep disorders, nausea, and vertigo (Collins 2026). Infrasound is below 20 Hz and can’t be heard by humans, but rather is felt and the vibrations from infrasound can travel hundreds of miles if unimpeded.

Habitat & Biodiversity Impacts

Not only do HDCs present costs in terms of energy and water, but they also come at a cost to biodiversity. HDC facilities have a huge land area footprint with many of them being 100s-1000s of acres in size. Natural ecosystems, rural lands, and wildlife corridors are paved over in the process, leading to habitat loss, fragmentation, and degradation. The infrastructure built to support such facilities further cause habitat fragmentation and the increased traffic on roadways increases the risk of car strikes for wildlife. Species are also negatively impacted by water and air pollution and the heat island effect, the same as our own communities. When we are in the midst of a sixth mass extinction event, we need to protect our species that support our ecosystems and their services, rather than further contributing to the global decline of biodiversity. 

Can they be Sustainable?

Addressing the environmental impacts focuses on two key aspects: 1) clean energy, and 2) smart cooling. Providers, including Google, Microsoft, Meta, and Amazon have all pledged to operate on 100% renewable or carbon-free energy by the end of the century. While wind and solar power offer a low-carbon pathway for HDCs, they are weather-dependent and require energy storage. Another low-carbon option are small modular reactors (SMRs), which are compact nuclear units. These can provide clean energy, however, obvious obstacles are safety concerns and nuclear waste management (Cruzes 2025). Geothermal energy is also an option in areas that have access to geothermal resources for energy production (Selvakumar 2026). If HDCs are able to improve their cooling systems, for instance through high-efficiency heat exchangers and compressor-free cooling, it can significantly lower energy demand and efficient cooling, like D2C, reduces water usage. Additionally, switching to “green refrigerants”, such as hydrofuoroolefins (HFOs), which are found naturally in the environment and have minimal environmental impact, would reduce the risk of water pollution and habitat degradation (Airsys 2026).

To be truly sustainable or “green”, sustainability needs to be considered at all steps, including all aspects of operation, governance, supply chains, and stakeholder engagement. This requires HDCs and their operators to establish measurable targets, outcomes, and sustainability goals. It also requires transparency and true corporate responsibility (Selvakumar 2026). In the end, businesses and policymakers need to view green HDCs and the shift to sustainable technology, not as a constraint, but as a driver of innovation. If large-scale facilities, such as HDCs, are able to effectively implement the use of renewable energy and efficient cooling, then it will reduce energy and water demands, decrease the cost passed off to consumers, and provide a pathway for other industries to follow suit and decrease human impacts on the environment.

Stay tuned for more info and a letter to the SFWMD and Polk County Board of Commissioners!

Literature Cited

Airsys. 2026, March, 6. How to make AI data centers more sustainable. https://airsysnorthamerica.com/how-to-make-ai-data-centers-more-sustainable/ 

Berger, A., 2025. Thesis: Artificial Intelligence Data Centers and United States Based Hyperscalers: Impacts and Solutions. 

California Environmental Protection Agency. 2026. Understanding the urban heat island index. CalEPA. https://calepa.ca.gov/climate/urban-heat-island-index-for-california/understanding-the-urban-heat-island-index/ 

Collins, B. 2026, May 11. ‘Dizziness, nausea, vertigo, and sleep disruption’: The undetectable hum of AI data centers is making local residents sick. Techradar. https://www.techradar.com/pro/dizziness-nausea-vertigo-and-sleep-disruption-the-undetectable-hum-of-ai-data-centers-is-making-local-residents-sick 

Cruzes, S., 2025. Data centers in the age of AI: A tutorial survey on infrastructure, sustainability, and emerging challenges. Authorea Preprints. 

Guidi, G., Dominici, F., Squartini, T., Sprinkle, C., Gilmour, J., Butler, K., Bell, E., Delaney, S. and Bargagli-Stoffi, F.J.. 2026. Assessing the Carbon Emissions of United States Hyperscale Data Centers. 

Greenfield, Nicole. 2026. AI Data Centers: Big Tech’s Impact on Electric Bills, Water, and More. Consumer Reports. https://www.consumerreports.org/data-centers/ai-data-centers-impact-on-electric-bills-water-and-more-a1040338678/ 

Hedge, G. 2025, June 11. Closed- loop cooling: Water saver or chemical time bomb? KETOS. https://ketos.co/closed-loop-cooling-water-saver-or-chemical-time-bomb 

IEA. 2025a, April 10. Energy and AI: A world energy outlook special report. https://iea.blob.core.windows.net/assets/34eac603-ecf1-464f-b813-2ecceb8f81c2/EnergyandAI.pdf 

IEA. 2025b, April 10. AI is set to drive surging electricity demand from data centres while offering the potential to transform how the energy sector works - News - IEA. IEA. https://www.iea.org/news/ai-is-set-to-drive-surging-electricity-demand-from-data-centreswhile-offering-the-potential-to-transform-how-the-energy-sector-works  

Marinoni, A., Cambria, E., Lin, W., Mura, M.D., Chanussot, J., Ragusa, E., Tso, C.Y., Zhu, Y. and Horton, B. 2026. The data heat island effect: quantifying the impact of AI data centers in a warming world. arXiv preprint arXiv:2603.20897. 

Posey, M., Fellows, W., Lenoir, T., and Vaden, H. 2026, May 12. Compute sovereignty: The strategic importance of digital infrastructure. S&P Global. https://www.spglobal.com/en/research-insights/special-reports/compute-sovereignty-strategic-importance-of-digital-infrastructure 

Sailor, D.J., Abolhassani, S.S. and Martin, E.P. 2026. Data center waste heat as an emerging urban thermal hazard: First field measurements of neighborhood-scale air temperature impacts. ASME Journal of Engineering for Sustainable Buildings and Cities, 7(2), p.024501. 

Saul, J., Nicoletti, L., Pogkas, D., Bass, D., and Malik, N. 2025, September 29. AI data centers are sending power bills soaring. Bloomberg Tech. https://www.bloomberg.com/graphics/2025-ai-data-centers-electricity-prices/ 

Selvakumar, P. 2026. Carbon Footprint Monitoring and Reduction Strategies Green Data Center Governance: A Review. Journal of Carbon Sequestration and Climate Engineering, 1(1). 

Shehabi, A., Newkirk, A., Smith, S.J., Hubbard, A., Lei, N., Siddik, M.A.B., Holecek, B., Koomey, J., Masanet, E. and Sartor, D. 2024. 2024 United States Data Center Energy Usage Report. 

Sierra Club. n.d. Water and local impacts of hyperscale data centers. https://www.sierraclub.org/sites/default/files/2025-12/water-and-local-impacts-of-hyperscale-data-centers.pdf 

U.S. Dept. of Energy. 2024, December 20. DOE releases new report evaluating increase in electricity demand from data centers. https://www.energy.gov/articles/doe-releases-new-report-evaluating-increase-electricity-demand-data-centers 

Yañez-Barnuevo, M. 2026, March 23. Communities are raising noise pollution concerns about data centers. Environmental and Energy Study Institute. https://www.eesi.org/articles/view/communities-are-raising-noise-pollution-concernsabout-data-centers