It has become apparent that humans are working for artificial intelligence (AI) more than AI is working for humans.
AI data centers are voracious consumers of human resources and a bane to our environment. They exacerbate global warming and greedily swallow up an enormous proportion of available capital, labor, land, energy, and water. We are investing heavily in the future of artificial intelligence, while blindly neglecting our own current and future needs.
Statistics Tell the Tale
Data centers house the servers, storage systems, networking equipment (switchers, routers), cooling systems, power infrastructure (backup generators), and security required to support AI. They run 24/7 in enormous facilities scattered all over the world—sited where there is cheap land, power, water, and tax incentives.
There are now more than 12,000 data centers in 174 countries (5427 in the US, Germany with 529, UK with 523, and China with 449).1Data center investment is exploding and dominates all infrastructure spending ($489.5 billion in 2025, a 46.8% increase from 2024’s $333.4 billion).2In 2023, data centers already accounted for about 4.4% of total US electricity consumption (it is growing 12% per year and will soon be about 10% of total use).3Servers generate enormous heat and already require 560 billion liters of water per year for cooling. Oddly, 25% of existing data centers and 30% of planned ones are in regions projected to have water scarcity by 2050.4Existing data centers are very large, and new ones will be truly enormous (one in planning will occupy 2000 acres, about 2000 football fields).Data centers result in increased consumer costs for power and water; strain power grids, and provide few jobs once construction is completed.Data centers are vitalizing the fossil fuel industry and destroying any possibility of meeting climate control targets (60% of fossil fuel guzzling power plants that had been scheduled for retirement will remain operative).5Lack Of Regulation
Current presidential executive orders are intended to reduce or block AI regulation at both the federal and state levels.6 Federal policy has portrayed data centers as essential for economic growth and national security. Agencies are empowered to “cut red tape” by minimizing or even eliminating reviews of environmental harms, community impacts, workers safety, and public health concerns. As a result, data center projects will advance with minimal scrutiny despite the substantial harms they will cause via extravagant energy, water, and land use. Collectively, these policies foster an environment in which speed and scale take precedence over scrutiny, allowing data centers to operate with limited oversight despite their significant destructive impacts.7
Pushback
Community attempts to halt data center expansion are constrained by the lack of Big AI transparency, rushed by expedited public review periods, and hindered by nondisclosure agreements between developers and local governments. Community groups usually lack the funding, organizational capacity, and top legal talent needed to sustain legal challenges against giant AI companies. Most data center projects have advanced with minimal opposition despite their substantial footprints. But there are some exceptions that can serve as a model for community opposition:
Ashburn, Virginia mobilized against proposed high voltage transmission lines intended to provide energy for a nearby data center. The citizens argued the lines would cut through neighborhoods and schools, pose health risks, and lower home values. Thousands of written comments and hundreds of registered speakers weighed in at local hearings.8Kansas City successfully halted a $12 billion data center after its zoning had been initially approved. They argued that the company lacked transparency with the community, not sharing sufficient details about the project despite receiving hundreds of public information requests.9Midwest Environmental Advocates sued Racine, Wisconsin after city officials failed to respond to requests to release records on the projected water use of a Microsoft data center. Only after court intervention were documents released estimating annual water consumption at 8.4 million gallons.10Minnesota Center for Environmental Advocacy sued developers for keeping the public uninformed by refusing to disclose its projected water and energy use.11Imperial, California filed suit to overturn approvals for an AI linked data center, alleging that county officials violated environmental law by rushing approvals without adequate public input and failing to conduct a thorough environmental review.12The Future
There are 4 possible futures for data center metastasis—all of them grim for energy use, water depletion, and global warming.
Extrapolating from the present: Global electricity consumption from data centers doubles by 2030, reaching about 945 terawatt-hours (comparable to Japan’s total annual electricity use).Liftoff: AI adoption accelerates rapidly and policy and infrastructure bottlenecks are minimal. Electricity demand grows even more dramatically to about 1,700 terawatt-hours by 2035.High efficiency model: Improvements in data center technology reduce their energy demands, but electricity use still rises significantly, reaching 970 terawatt-hours by 2035.Headwinds: Slower AI uptake results from economic, technologic, supply chain, or regulatory constraints. Yet even here, data center electricity demands do not decline, but instead plateau at around 700 terawatt-hours by 2035.Concerns for Psychiatrists
The massive financial, energy, water, and global warming costs associated with data centers have a powerful, if indirect, impact on the world’s mental health. Psychiatric symptoms do not ever occur in a vacuum; diagnosing the context in which symptoms arise and persist is as important as applying a DSM label to them. Socioeconomic risk factors (eg, financial distress, joblessness, chronic stress, homelessness) very much influence who suffers from mental illness, severity of illness, whether treatment is available, and the treatment outcome.13,14 Vast investments in data centers prevent investment in mental health services and other human infrastructure. Patients who might benefit greatly from treatment have no access to it. The higher energy and water bills routinely caused by nearby data centers can tip people at the boundary into financial crisis and homelessness, which in a vicious cycle worsens mental disorders which further worsen the socioeconomic triggers. Mental health services were shamefully
underfunded before data centers—there exponential growth ensures the continuing neglect of existing patients and the creation of new ones.
Concluding Thoughts
Global warming was a clear and present danger to human survival even before data centers unleashed a shocking and unexpected increase in the use of fossil fuels. Previous international commitments to reduce carbon emissions have been superseded by the mad rush to create massive data centers with completely unsustainable energy and water needs. We are perilously close to a disastrous global warming tipping point, if we have not already passed it. Data centers are greasing the slippery slope toward a possible human extinction.
Is there any hope that our species will wise up before it is too late? Probably not much. Evolution has shaped us to be smart enough to create remarkably powerful technologies, but not wise enough to use them safely. Our only hope for a safer future may arise from the increasing destruction caused by climate change in the present. Policy makers will eventually realize the madness of our energy policies as we begin to experience ever more severe climate disasters. Hopefully this wakeup call happens before additional suffering and before it is too late to reverse climate damage. But the explosion of data centers in the face of global warming does not inspire confidence in the wisdom of our species or bode well for our survival.
The opinions expressed are those of the author and do not necessarily reflect the opinions of Psychiatric Times.
Dr Frances is professor and chair emeritus in the department of psychiatry at Duke University.
Ms Beaver is a psychology student at UCLA interested in how artificial intelligence will impact mental health.
References
1. Leichter R. Number of data centers by country (November 2025). CARGOSON. January 4, 2026. Accessed January 28, 2026. https://www.cargoson.com/en/blog/number-of-data-centers-by-country
2. Roach A. Data center deals hit record $61 billion in 2025 amid construction frenzy. CNBC. December 19, 2025. Accessed January 28, 2026. https://www.cnbc.com/2025/12/19/data-center-deals-hit-record-amid-ai-funding-concerns-grip-investors.html
3. DOE releases new report evaluating increase in electricity demand from data centers. US Department of Energy. December 20, 2024. Accessed January 28, 2026. https://www.energy.gov/articles/doe-releases-new-report-evaluating-increase-electricity-demand-data-centers
4. Nicoletti L, Ma M, Bass D. AI is draining water from areas that need it most. Bloomberg. May 8, 2025. Accessed January 28, 2025. https://www.bloomberg.com/graphics/2025-ai-impacts-data-centers-water-data/
5. Kearney L. AI data center are forcing dirty ‘peaker’ power plants back into service. Reuters. December 23, 2025. Accessed January 28, 2025. https://www.reuters.com/business/energy/ai-data-centers-are-forcing-obsolete-peaker-power-plants-back-into-service-2025-12-23/
6. Keener B, Sathyamurthy L. Executive order issues to restrict state regulation of artificial intelligence. Phillips Lytle LLP. December 11, 2025. Accessed January 28, 2025. https://phillipslytle.com/executive-order-issued-to-restrict-state-regulation-of-artificial-intelligence/
7. Accelerating federal permitting of data center infrastructure. The White House. July 23, 2025. Accessed January 28, 2025. https://www.whitehouse.gov/presidential-actions/2025/07/accelerating-federal-permitting-of-data-center-infrastructure/
8. Heckt S. Loudoun residents take fight against high-voltage power lines for data center alley to SCC. Virginia Mercury. December 16, 2025. Accessed January 28, 2025. https://virginiamercury.com/2025/12/16/loudoun-residents-take-fight-against-high-voltage-power-lines-for-data-center-alley-to-scc/
9. Socha E. Lawsuit delays $12B data center in Kansas City as community, environmental group voice concerns. Kansas Reflector. November 7, 2025. Accessed January 28, 2025. https://kansasreflector.com/2025/11/07/lawsuit-delays-12b-data-center-in-kansas-city-as-community-environmental-group-voice-concerns/
10. Demanding data center transparency – city of Racine. Midwest Environmental Advocates. 2025. Accessed January 28, 2025. https://midwestadvocates.org/our-work/legal-action/demanding-transparency-about-the-environmental-impacts-of-data-centers/
11. Marohn K. Lawsuit alleges Hermantown kept public in dark over data center proposal. MPR News. November 5, 2025. Accessed January 28, 2025. https://www.mprnews.org/story/2025/11/05/lawsuit-says-hermantown-kept-public-in-dark-on-data-center-plan
12. Suzuki K. Imperial Valley city sues to force environmental review of massive data center project. KPBS. December 12, 2025. Accessed January 28, 2025. https://www.kpbs.org/news/environment/2025/12/12/imperial-valley-city-sues-to-force-environmental-review-of-massive-data-center-project
13. Batstra L, Frances A. Diagnosing the context is as important as diagnosing the individual. Front Psychiatry. 2025;16.
14. Salem M, Robenson J. The impact of socioeconomic factors on mental health: a conceptual framework. Cureus. 2025;17(7):e88244.