Statement on Proposed Artificial Intelligence (AI) Data Centers in the Rogue River Watershed
The purpose of this statement is to provide general education and recommendations to our members, the local community, and municipal decision-makers pertaining to the development of AI Data Centers and the subsequent impacts to the land and cold water resources of the Rogue River Watershed. The Rogue River is a defining natural feature in the greater Grand Rapids region, spanning across five counties (Kent, Montcalm, Newaygo, Ottawa, and Muskegon) and serves as a significant cool-cold water tributary of the Grand River. As such, it supports critical wildlife habitat for fish like trout and other species, offers opportunities for public recreation, and maintains downstream water quality in both the Rogue and Grand Rivers. The Rogue, along with several of its tributaries (Barkley, Rum, Shaw, Stegman, Cedar, Duke, Spring Creeks), have been designated with permanent protections under the Michigan Natural Rivers Act of 1994.
With a watershed spanning nearly 140,000 acres of largely forested and agricultural land to the north of urban Grand Rapids, the Rogue River not only offers aesthetic beauty and recreational opportunities, but also provides natural buffers for flooding control and water quality improvement in communities downstream. Due to its large expanses of rural land area, the Rogue River Watershed has been and will likely continue to be of interest for the development of new AI Data Centers in the state of Michigan.
As the Rogue River Watershed Partners (RRWPs), our mission is to protect, preserve, and promote the Rogue River, its tributaries, its watershed, and its communities. We acknowledge that data centers can bring economic benefit, and that rapidly-evolving technologies are continuously in development that could ensure a higher level of sustainability through their implementation at AI Data Centers. However, there are several risk areas that these data centers present to local communities and their natural resources, particularly in light of the fact that to date there is little government oversight or regulation of these facilities. We outline our concerns regarding these risk areas below.
With a watershed spanning nearly 140,000 acres of largely forested and agricultural land to the north of urban Grand Rapids, the Rogue River not only offers aesthetic beauty and recreational opportunities, but also provides natural buffers for flooding control and water quality improvement in communities downstream. Due to its large expanses of rural land area, the Rogue River Watershed has been and will likely continue to be of interest for the development of new AI Data Centers in the state of Michigan.
As the Rogue River Watershed Partners (RRWPs), our mission is to protect, preserve, and promote the Rogue River, its tributaries, its watershed, and its communities. We acknowledge that data centers can bring economic benefit, and that rapidly-evolving technologies are continuously in development that could ensure a higher level of sustainability through their implementation at AI Data Centers. However, there are several risk areas that these data centers present to local communities and their natural resources, particularly in light of the fact that to date there is little government oversight or regulation of these facilities. We outline our concerns regarding these risk areas below.
Key Concerns of the RRWPs regarding the development of AI Data Centers within the Rogue River Watershed:
|
1. Site placement
a) AI data center acreage varies significantly, but the current trend is toward massive campuses ranging from several dozen acres to over 1,000 acres. b) Land use decisions and zoning in Michigan are handled at the municipal level (township, county) through zoning boards and municipal government with -at present- little regulatory direction at the state and federal government levels regarding construction of new AI Data Centers. Therefore, individual communities play a significant role in regulating the development of AI Data Centers in their regions. c) The increase in impervious surfaces (such as rooftops, parking lots, roads, and sidewalks) as a result of new development can have negative impacts on communities and the environment by increasing stormwater runoff volume and speed, pollutant loading during and after construction, increased turbidity of nearby surface water sources, and thermal impacts to coldwater streams. These impacts often lead to issues like flooding, habitat degradation, alterations to natural water cycles, and decreased water quality. |
Image Credit: Google
|
2. Water usage, consumption, & impacts to watersheds
a) AI Data Centers contain large numbers of servers that generate heat which must then be removed from the data center, a process that often involves water. There are a few main ways that data centers may use water for cooling - 1) air cooling using water evaporation, which is a more water-intensive open-loop system; 2) liquid cooling where the liquid coolant (typically water) is applied directly to the Graphic Processing Units (GPUs) and Central Processing Units (CPUs), a process that can also be described as a ‘direct-to-chip’ method, requiring minimal water consumption; and 3) hybrid cooling systems that dynamically switch between air-based and water-based cooling, depending on ambient conditions. The latter are intelligent systems, using water only when necessary while leveraging outside air cooling whenever possible.
b) Data centers have been known to consume large amounts of water that may not be returned to the source watershed, potentially leading to the depletion of regional surface and groundwater sources. A mid-size data center consumes approximately 300,000 gallons of water per day, as much water as 1,000 U.S. households, while a large data center can use 1- 5 million gallons of water per day, rivaling the water usage of a town with a population of 10,000 to 50,000 people.
c) Water consumption also occurs in the production of electricity used to power data center systems as well as throughout the AI ‘training’ process before the units become operational. Thermal power plants generate this electricity using boiling water and steam-powered turbines.
d) As much as 80% of the water withdrawn evaporates throughout the cooling process and is therefore not necessarily returned to the watershed from which it was withdrawn, leading to potential drops in downstream water levels or aquifer exhaustion, particularly in water-stressed regions.
e) The additional wastewater released from AI Data Centers may overwhelm local infrastructure, disrupting or slowing the wastewater treatment process.
a) AI Data Centers contain large numbers of servers that generate heat which must then be removed from the data center, a process that often involves water. There are a few main ways that data centers may use water for cooling - 1) air cooling using water evaporation, which is a more water-intensive open-loop system; 2) liquid cooling where the liquid coolant (typically water) is applied directly to the Graphic Processing Units (GPUs) and Central Processing Units (CPUs), a process that can also be described as a ‘direct-to-chip’ method, requiring minimal water consumption; and 3) hybrid cooling systems that dynamically switch between air-based and water-based cooling, depending on ambient conditions. The latter are intelligent systems, using water only when necessary while leveraging outside air cooling whenever possible.
b) Data centers have been known to consume large amounts of water that may not be returned to the source watershed, potentially leading to the depletion of regional surface and groundwater sources. A mid-size data center consumes approximately 300,000 gallons of water per day, as much water as 1,000 U.S. households, while a large data center can use 1- 5 million gallons of water per day, rivaling the water usage of a town with a population of 10,000 to 50,000 people.
c) Water consumption also occurs in the production of electricity used to power data center systems as well as throughout the AI ‘training’ process before the units become operational. Thermal power plants generate this electricity using boiling water and steam-powered turbines.
d) As much as 80% of the water withdrawn evaporates throughout the cooling process and is therefore not necessarily returned to the watershed from which it was withdrawn, leading to potential drops in downstream water levels or aquifer exhaustion, particularly in water-stressed regions.
e) The additional wastewater released from AI Data Centers may overwhelm local infrastructure, disrupting or slowing the wastewater treatment process.
3. Energy Use
a) The massive uptick in energy consumption is outpacing the growth of sustainable energy sources, resulting in overall increases to energy consumption. Many of the costs of this increased energy usage are transferred to local communities, in some cases driving residential energy bills up.
b) The use of fossil fuel sources to generate electricity to power AI Data Centers and backup generators contributes to carbon dioxide emissions, further contributing to climate change.
a) The massive uptick in energy consumption is outpacing the growth of sustainable energy sources, resulting in overall increases to energy consumption. Many of the costs of this increased energy usage are transferred to local communities, in some cases driving residential energy bills up.
b) The use of fossil fuel sources to generate electricity to power AI Data Centers and backup generators contributes to carbon dioxide emissions, further contributing to climate change.
In light of these concerns, the RRWPs has the following recommendations for communities:
1. Attend local municipal meetings to gather information, ask questions and address any potential impacts to our cold water resources.
2. Contact state and federal government representatives about the importance of government regulatory measures that include protections for land and water resources regarding construction and ongoing maintenance of new AI Data Centers.
3. Promote use of appropriate sites like brownfields (former industrial sites) instead of using forest, farmland and other natural areas.
4. Utilize robust riparian buffers and best management practices whenever possible to control stormwater and mitigate impacts to our cold water resources.
5. Promote development and use of more sustainable cooling methods and energy use.
6. Request that state and municipal authorities require increased public transparency of actual water consumption by any proposed AI Data Center, as well as written protection plans for other potential impacts to land and water resources, such as stormwater management, wastewater removal, erosion & sediment controls, stream temperature monitoring, and noise pollution.
7. Request that state and municipal authorities require long-term control measures such as third-party monitoring, maintenance bonds for stormwater systems, and legally enforceable remedies if impacts occur.
8. Promote renewable energy sources both on campus, e.g. rooftop solar, and in support of the grid.
9. Promote increased transparency of energy use and regional cost distribution.
2. Contact state and federal government representatives about the importance of government regulatory measures that include protections for land and water resources regarding construction and ongoing maintenance of new AI Data Centers.
3. Promote use of appropriate sites like brownfields (former industrial sites) instead of using forest, farmland and other natural areas.
4. Utilize robust riparian buffers and best management practices whenever possible to control stormwater and mitigate impacts to our cold water resources.
5. Promote development and use of more sustainable cooling methods and energy use.
6. Request that state and municipal authorities require increased public transparency of actual water consumption by any proposed AI Data Center, as well as written protection plans for other potential impacts to land and water resources, such as stormwater management, wastewater removal, erosion & sediment controls, stream temperature monitoring, and noise pollution.
7. Request that state and municipal authorities require long-term control measures such as third-party monitoring, maintenance bonds for stormwater systems, and legally enforceable remedies if impacts occur.
8. Promote renewable energy sources both on campus, e.g. rooftop solar, and in support of the grid.
9. Promote increased transparency of energy use and regional cost distribution.
RESOURCES:
AI data centers & the fight for Michigan’s water. (2023) Economic Development Responsibility Alliance of Michigan. https://edraofmi.org/ai-data-centers
Copley, M. (2022) Data centers, backbone of the digital economy, face water scarcity and climate risk. NPR, https://www.npr.org/2022/08/30/1119938708/data-centers-backbone-of-the-digital-economy-face-water-scarcity-and-climate-ris
Garcia, Mya, AI Uses How Much Water? Navigating Regulation Of AI Data Centers' Water Footprint Post-Watershed Loper Bright Decision (December 13, 2024). Available at SSRN: https://ssrn.com/abstract=5064473 or http://dx.doi.org/10.2139/ssrn.5064473
Gupta, J., Bosch, H., and van Vliet, L. (2024) AI’s excessive water consumption threatens to drown out its environmental contributions. Science-Policy Brief for the Multistakeholder Forum on Science, Technology and Innovation for the SDG. https://sdgs.un.org/sites/default/files/2024-05/Gupta,%20et%20al._AIs%20excessive%20water%20consumption.pdf
Handbook for Municipal Officials - Section 3. Operations. Chapter 16: Planning and Zoning. (2024) Michigan Municipal League. https://mml.org/wp-content/uploads/2024/12/HMO-CH-16-Planning-and-Zoning.pdf
Ignaczak, N.M. (2025) Data Centers in Michigan: What you need to know. Michigan Public. https://www.michiganpublic.org/environment-climate-change/2025-11-21/data-centers-in-michigan-what-you-need-to-know#data-centers-water-use
Kwong, E., Barber, R.G., Chinn, H. Why the true water footprint of AI is so elusive, (2025). Podcast, NPR Shortwave. https://www.npr.org/2025/05/07/1249592906/energy-water-ai-climate-tech
Minnesota Trout Unlimited. (2025). Data Centers and Coldwater Fisheries. https://mntu.org/data-centers-and-coldwater-fisheries/
Modern Data Center Cooling. (2023) Microsoft Corporation. https://datacenters.microsoft.com/wp-content/uploads/2023/05/Azure_Modern-Datacenter-Cooling_Infographic.pdf
Pengfei, Li; Yang, Jianyi; Islam, Mohammad A.; Ren, Shaolei (2025). Making AI Less “Thirsty”: Uncovering and Addressing the Secret Water Footprint of AI Models. https://arxiv.org/pdf/2304.03271
Policy Statement on Artificial Intelligence Data Centers. (2025) PA Council of Trout Unlimited. https://patrout.org/policy-statement-on-artificial-intelligence-data-centers/
PUE vs. WUE: Balancing Efficiency and Sustainability in Data Centers. (2026) AIRSYS Cooling Technologies Inc. https://airsysnorthamerica.com/puw-vs-wue-balancing-efficiency-sustainability-in-data-centers/
Yanez-Barnuevo, M. (2025) Data Centers and Water Consumption. Environmental and Energy Study Institute. https://www.eesi.org/articles/view/data-centers-and-water-consumption
AI data centers & the fight for Michigan’s water. (2023) Economic Development Responsibility Alliance of Michigan. https://edraofmi.org/ai-data-centers
Copley, M. (2022) Data centers, backbone of the digital economy, face water scarcity and climate risk. NPR, https://www.npr.org/2022/08/30/1119938708/data-centers-backbone-of-the-digital-economy-face-water-scarcity-and-climate-ris
Garcia, Mya, AI Uses How Much Water? Navigating Regulation Of AI Data Centers' Water Footprint Post-Watershed Loper Bright Decision (December 13, 2024). Available at SSRN: https://ssrn.com/abstract=5064473 or http://dx.doi.org/10.2139/ssrn.5064473
Gupta, J., Bosch, H., and van Vliet, L. (2024) AI’s excessive water consumption threatens to drown out its environmental contributions. Science-Policy Brief for the Multistakeholder Forum on Science, Technology and Innovation for the SDG. https://sdgs.un.org/sites/default/files/2024-05/Gupta,%20et%20al._AIs%20excessive%20water%20consumption.pdf
Handbook for Municipal Officials - Section 3. Operations. Chapter 16: Planning and Zoning. (2024) Michigan Municipal League. https://mml.org/wp-content/uploads/2024/12/HMO-CH-16-Planning-and-Zoning.pdf
Ignaczak, N.M. (2025) Data Centers in Michigan: What you need to know. Michigan Public. https://www.michiganpublic.org/environment-climate-change/2025-11-21/data-centers-in-michigan-what-you-need-to-know#data-centers-water-use
Kwong, E., Barber, R.G., Chinn, H. Why the true water footprint of AI is so elusive, (2025). Podcast, NPR Shortwave. https://www.npr.org/2025/05/07/1249592906/energy-water-ai-climate-tech
Minnesota Trout Unlimited. (2025). Data Centers and Coldwater Fisheries. https://mntu.org/data-centers-and-coldwater-fisheries/
Modern Data Center Cooling. (2023) Microsoft Corporation. https://datacenters.microsoft.com/wp-content/uploads/2023/05/Azure_Modern-Datacenter-Cooling_Infographic.pdf
Pengfei, Li; Yang, Jianyi; Islam, Mohammad A.; Ren, Shaolei (2025). Making AI Less “Thirsty”: Uncovering and Addressing the Secret Water Footprint of AI Models. https://arxiv.org/pdf/2304.03271
Policy Statement on Artificial Intelligence Data Centers. (2025) PA Council of Trout Unlimited. https://patrout.org/policy-statement-on-artificial-intelligence-data-centers/
PUE vs. WUE: Balancing Efficiency and Sustainability in Data Centers. (2026) AIRSYS Cooling Technologies Inc. https://airsysnorthamerica.com/puw-vs-wue-balancing-efficiency-sustainability-in-data-centers/
Yanez-Barnuevo, M. (2025) Data Centers and Water Consumption. Environmental and Energy Study Institute. https://www.eesi.org/articles/view/data-centers-and-water-consumption
*Updated January, 2026