SEMINOLE, Okla. — The Seminole Nation of Oklahoma turned a local fight over a proposed technology project into a national flashpoint by voting 24-0 to impose a moratorium on generative AI technology and hyperscale data center development within its jurisdiction. The council acted after a company approached tribal leaders about a data center proposal, as advocates warned that AI data centers could consume scarce water, strain power grids and industrialize tribal and rural lands, March 7.
The vote, reported by Native News Online, made the Seminole Nation one of the first Native American tribes to enact a moratorium on data center development on tribal lands. Mekusukey Band Representative Glen Chebon Kernell, who introduced the resolution, framed the decision as a water-protection issue, saying, “This is the threshold. There’s no turning back.”
AI data centers become a sovereignty fight
For Indigenous organizers, the dispute is not simply about whether artificial intelligence should exist. It is about who controls the land, water, energy and data systems that make AI possible. The Seminole case added urgency because tribal members said the proposal arrived with requests for a nondisclosure agreement and a letter of intent, raising concerns that communities could be locked out of decisions before they fully understand the environmental and political costs.
Honor the Earth, an Indigenous-led climate justice group that supported the Seminole effort, has launched a data center tracker for Indigenous lands to help communities identify proposed and expanding hyperscale projects. The group describes the fight as a new form of extraction, comparing the demand for water, power and land to earlier battles over mines, pipelines and industrial development.
The backlash is spreading because data centers are no longer small warehouse projects. The U.S. Department of Energy said in a 2024 report on data center energy demand that data centers consumed about 4.4% of total U.S. electricity in 2023 and could consume 6.7% to 12% by 2028. Those numbers have become central to community concerns that local ratepayers, rural utilities and water systems may be asked to absorb the costs of a private AI boom.
Water risks are driving the Indigenous backlash
Water is the most politically explosive issue. Many data centers use water to cool servers, and even air-cooled systems can raise electricity demand, which can indirectly increase water use at power plants. In the Mountain West, tribal leaders and water advocates say the problem is not just total gallons used nationwide, but where the facilities are built and whose watersheds bear the risk.
That concern is already visible in Nevada, where the Pyramid Lake Paiute Tribe has raised alarms about the massive buildout around the Tahoe-Reno Industrial Center. The center sits upstream from Pyramid Lake, and The Guardian reported that U.S. data center water use has tripled over the past decade to more than 17 billion gallons a year, with further growth expected by 2028. For tribes whose cultural, ecological and legal fights are tied to rivers and lakes, even “reclaimed” water can become controversial if it changes flows that would otherwise return to a watershed.
Those conflicts also intersect with tribal digital sovereignty. In comments to federal officials, the National Congress of American Indians said Tribal Nations must retain authority over AI-related data collection, management and ownership, including data tied to citizens, cultural knowledge, environmental conditions and public health. That position broadens the debate beyond zoning and utilities: Indigenous leaders are asking whether AI infrastructure can be built without undermining tribal law, consent and self-determination.
A backlash years in the making
The Seminole Nation’s moratorium did not appear in isolation. Earlier reporting showed the same pattern emerging across the United States and abroad: data center developers promise investment, then communities discover the projects may require large volumes of water, new power infrastructure and long-term industrial commitments.
In 2023, Axios reported that Microsoft’s five West Des Moines data centers used as much as 11.5 million gallons of water a month for cooling during peak summer periods. In 2024, Reuters reported that a Chilean environmental court partially reversed a Google data center permit and ordered further climate-related review of impacts on Santiago’s stressed aquifer. Later that year, The Washington Post reported that data center growth in Arizona was intensifying power-grid debates while many Navajo Nation households still lacked electricity.
Together, those examples explain why Indigenous advocates view the AI buildout as more than a technology story. It is a land-use story, a water story and a governance story. The economic promises are real for some local governments, but so are the questions over who gets jobs, who pays for grid upgrades, who loses water security and who controls the decision-making process.
Tribes are not rejecting technology outright
Some tribal leaders and Indigenous technologists support AI tools when they are governed by tribal law and used for language preservation, climate planning, health care or emergency response. The sharper dispute is over outside companies building hyperscale facilities that may lock communities into decades of water and electricity demand while offering limited permanent employment.
That distinction matters. A tribally governed digital infrastructure project designed around local consent, cultural protection and resource limits is different from a corporate data center campus built primarily to serve national or global AI platforms. Indigenous opponents say developers often blur that line by presenting data centers as innovation while downplaying the physical footprint: land conversion, transmission lines, backup generators, water withdrawals, wastewater and noise.
What comes next for AI data centers
The Seminole vote is likely to encourage other tribal governments, rural counties and city councils to consider moratoriums, disclosure rules or stricter permitting standards. The most immediate demands from opponents are clear: early public notice, no nondisclosure agreements that block community oversight, full water-use estimates, grid-impact studies, enforceable community benefits and respect for tribal consent.
For tech companies, the conflict is a warning that the race to build AI infrastructure will face resistance wherever communities believe the costs are being hidden. For Indigenous nations, the Seminole moratorium signals a wider strategy: slow the process, force transparency and make sovereignty the starting point rather than an afterthought.
The fight over AI data centers is now moving from abstract questions about algorithms to concrete fights over rivers, aquifers, substations and land. In Indian Country, that makes the AI boom look less like a virtual revolution and more like the latest test of who has the power to decide what happens to the places that sustain life.
