By Koohan Paik-Mander

For the last five centuries, Indigenous Peoples and regions have been ravaged by outsiders pillaging resources: land, water, bodies, minerals—even DNA. In the 21st century, the latest extraction rush is driven by the hubristic vision of an economy driven by Artificial Intelligence (AI). The resource extraction required for this new fever dream delves into previously unmolested territory, at a scale and pace that is staggering. The sooner we can understand the nature and nuances of current technocolonialism, the sooner we can identify the most effective points of intervention for demanding adherence to the principle of Free, Prior and Informed Consent.

The centerpiece for the current extraction frenzy is the plan for a global, industrial AI industry. This would include establishing its supply chain and constructing its infrastructure. In the first press conference that President Trump held after his 2025 inauguration, he announced the “Stargate Project,” an ambitious, half-trillion-dollar boondoggle intended to cover the continent with data centers. Those are the gargantuan concrete warehouses that are supposed to provide the U.S. with AI as an “essential utility,” like water or electricity.

At the press conference, three tech oligarchs who are leading the project—Sam Altman, Larry Ellison, and Masayoshi Son—lined up to make dubious promises of cancer cures and climate solutions, all thanks to AI. In the following months, a cavalcade of executive moves deregulated environmental, nuclear, and privacy policies while lavishing funding on new AI projects, particularly in the military.

An AI economy would be global and comprised of two main pillars. One would be data center construction. The other would be a surveillance/extraction infrastructure that would provide a continual flow of data into the number-crunching data centers. An AI economy would fail without these two pillars. This makes them key points of intervention for resistance.

It’s important to understand that Artificial Intelligence like ChatGPT is not “intelligence.” It’s actually a very sophisticated classification process that makes predictions off of vast amounts of data that are collected as the first step in building an AI machine. You can’t build an AI without the data. The data is used to program the AI to recognize sentences, images, or DNA sequences. Simply put, AI is a fancy pattern recognizer. 

The term “data” refers to any samples of text, images, sounds, or behavioral changes—even DNA. It’s all digitized into a mathematical language so that it can be processed by a machine.

Once the AI is adequately programmed by enough data samples, that model is equipped to recognize certain patterns and make predictions based on those patterns. You could give it part of a sentence and it would be able to finish it with the right word. You could give it a series of prompt words and request that it compile an image, and because it was programmed to recognize the patterns of millions of images, it would be able to predict what you were requesting. 

Millions of samples of biological samples have been pirated by biotech corporations from Indigenous and marginalized communities to build AI models. This is why the corporations are so non-transparent in revealing the origins of their genetic data. Using AI and all that genetic data, tech companies are able to generate genetic sequences with certain attributes, such as a virus that doesn’t show symptoms but is highly contagious, which might be used in biowarfare.

The legal methods of extracting data are almost as nefarious as the illegal ones. Already, the Trump administration is funneling the data from every agency database in the U.S. government to Palantir Corporation—the same company that uses AI to generate the “kill lists” of Palestinians for the Israeli Defense Forces. And the same one which works with the U.S. Immigration and Customs Enforcement (ICE) to track migrants. 

In addition to government data, Palantir would have access to all the data continually generated through what is loosely called the “Internet of Things.” That’s the network of smart appliances and gadgets and voice and facial recognition and biometric sensors; the doorbell, the stove, the washer, the car, everything that’s been rigged out as “smart” technology. And, of course the Big Brother device in your purse or pocket masquerading as a phone. All these extraction devices are hooked up to talk to each other and the manufacturers to continually send data to companies and/or governments. Every move we make feeds into the valuable data stream.

Greed for wealth and power is driving the greed for more and more data. Data is being called “the new oil.” And if data is the new oil, then data centers are the oil refineries. Only there, refining takes place in the form of AI computing. Data-center computing is the second pillar required for bringing an AI economy to reality.

AI data center. Photo courtesy of Rawpixel.

Data centers, such as those envisioned by the Stargate Project, are comprised of clusters of warehouses built upon a slab of grim, Earth-smothering concrete. Each warehouse is a half-million square feet, filled from floor to ceiling with stacks of metal boxes—computers that process all the AI calculations, to the tune of billions per second—so much number-crunching that the machines heat up. To cool them, giant pipes filled with water snake through the basements of these buildings with capillaries of cooling liquid that branch off to run upwards alongside each of the machines. The water consumption for cooling data centers is enormous. Crops in drought-stricken Queretaro, Mexico are withering while water is diverted to the development of newly constructed data centers there owned by Microsoft, Amazon, Google, and other tech companies.

Data centers are what is euphemistically called “the cloud,” where our Google docs and emails and other files are stored on a virtual drive. Calling it “the cloud” conjures something abstract and light and fluffy, when in fact, data centers built specifically for AI are wreaking environmental havoc on Indigenous and other marginalized communities around the world. As awareness is raised globally on just how ecocidal these data gulags are, communities are rising up in protest. This is essential, since obstructing data centers is the best way to knock down a principal pillar in an AI economy before it can get started.

A $2 billion data center has been proposed on the traditional territory of the Sturgeon Lake Cree Nation. The so-called “green” development is advertised as “one of Canada’s leading net-zero, eco-industrial developments.” But the First Nations of the area are having none of it. On January 13, 2025, in an open letter to Alberta Premier Danielle Smith, Chief Sheldon Sunshine called for the provincial government to cease and desist. He pointed out that First Nations members have traplines in the area and rely on water from the Smoky River.
In southwest Memphis, Tennessee, people are choking on air that has been made unbreathable by Elon Musk’s data center, named Colossus. There’s nothing light and fluffy about “the cloud.” 

Energy consumption by data centers is so astronomical that their existence would foreclose the possibility of ever reaching carbon emission goals. According to the International Energy Agency, by 2030—only five years from now—the data centers will consume more electricity globally than manufacturing of steel, cement, chemicals, and all other energy-intensive goods combined worldwide.
Jeff Bezos, Bill Gates, Altman, and others have responded by plowing hundreds of billions of dollars into a “nuclear energy renaissance,” as if that would be a viable alternative to over-consumption of carbon-based energy.

Just when Indigenous Peoples thought the worst uranium mining and radioactive waste storage practices might be in the past, the tycoons of Silicon Valley have begun research and development of Small Modular Reactors, or SMRs. The ambitious goal is to develop the first mass-produced SMRs so that every data center can have its own SMR. But this vision is still years away. In the meantime, the mandate remains, “drill, baby, drill” until functional SMRs come online.

The state of New Mexico has been uniquely burdened with every stage of the nuclear industry life cycle. Ever since J. Robert Oppenheimer developed the atom bomb, the result has been that uranium mining, enrichment, testing, and waste disposal have irreversibly poisoned native lands and water. The plans to power AI with nuclear energy will impact Indigenous communities hardest in New Mexico, but also in other western states reaching as far north as Montana that are targeted for uranium mining.

It is no coincidence that AI is being researched at Los Alamos, birthplace of the atom bomb. Artificial intelligence and algorithms emerged as the support infrastructure to Oppenheimer’s invention. They were first developed at the start of the Cold War as the operating logic for systems to deploy missiles armed with nuclear warheads. And now, eight decades later, AI is driving the development of nuclear energy so that it can be used to develop more AI. It’s an elaborate symbiosis of nihilism that the Indigenous Peoples in New Mexico have tragically had to witness in real time for the better part of a century.

Petuuche Gilbert photographed by Koohan Paik-Mander.

Despite the entwinement of nuclear energy and nuclear power, the industry has always conceptually cleaved them into two separate and distinct parts. However, Indigenous Peoples who have been impacted know better than anyone that the two are inextricably linked. Petuuche Gilbert, an Acoma Elder who has been a nuclear activist and protector of his lands, water, and people for close to three decades, says that the division between weapons and energy is a false dichotomy: “Land, earth, air, water, people. Indigenous philosophy understands that we live in coexistence. Any part of the nuclear industry, whether for energy or for weapons, is harmful to that balance. This is not an unrealistic position.”

Japanese anti-nuke activists have always called nuclear energy and nuclear weapons “two heads of the same snake.” In fact, the life cycles of both energy and weapons share most processes, such as uranium mining, enrichment, and disposal of radioactive waste. All of these are irreversibly damaging to Indigenous lands, water, and health.

Los Alamos National Laboratory is currently seeking to increase the number of nuclear pits manufactured each year. The pit is the part of the bomb that explodes into apocalyptic destruction. Each one of these bombs will have about 20 times the destructive force of the bomb that leveled Hiroshima. 

Nuclear pit manufacturing had been dormant since 1989, but was revived in 2025 at the rate of 30 pits per year. The new plan is to manufacture 80 to 100 pits per year by 2030. It is safe to assume that with the increase in pit production, there will also be an increase in underground nuclear testing. Understandably, the suggestion is traumatizing to the Indigenous downwinders, who are still suffering from serious health problems from the testing that took place between 1951 and 1992.

Abandoned uranium mining site in Arizona. Photo by doctress neutopia.

The combination of the initiatives to increase nuclear pit production along with developing SMRs have sent uranium stocks skyrocketing, with absolutely no regard for the horrific impacts on Native communities, let alone their right to Free, Prior and Informed Consent. Since 1940, Indigenous Peoples in the American Southwest have been opposing uranium mining and the nuclear industry. Every stage of nuclear production has been a scourge on generations of innocents, with cancer, birth defects, learning disabilities, kidney disease, and other unjust health crises, as mining and milling release radioactive poisons into their air, water, and land. Today, there are over 500 abandoned uranium mines on and near the Navajo Nation alone. These mines, left unregulated, continue to pose grave environmental and health risks to local Indigenous communities.

New Mexico is also home to the Waste Isolation Pilot Plant (WIPP), the only U.S. facility where radioactive waste from weapons manufacture is stored. It was a pilot program, opened in 1999 with a 25-year contract that expired last year. During that period, the Department of Energy promised to build new facilities in other states and close WIPP in April 2024. But the DOE broke that promise. According to the WIPP’s own Environmental Impact Study, the government plans to expand the landfill for 69 more years—even though it’s surrounded by fracking wells. New tunnels are being dug to house the waste from future weapons production nationally.

One can only imagine the potential for trucking accidents involving radioactive waste being transported across Indigenous lands to WIPP; several have already taken place. Once again, the burden of this cataclysmically toxic industry falls on Native communities concentrated in New Mexico, as well as those in other parts of North America.

As to the question of where the radioactive waste will go from the hundreds of SMRs that are being planned to power AI in the future, Gilbert observes, “That’s the whole problem with high-level nuclear waste, like the spent fuel rods coming from the SMRs or other nuclear reactors. Where do you put it? They know it’s a big problem, but they don’t have any place in mind to put it.” 

Such slipshod planning is not reassuring as corporate AI interests barrel forward in their endeavor to mass produce SMRs. If they succeed, Indigenous communities will have to contend with the enrichment process as well. Enriching uranium releases highly dangerous chemicals and radioactive materials into the environment. SMR uranium is enriched almost to the level of weapons-grade uranium, which makes it far more radioactive than uranium enriched for larger, traditional nuclear reactors.

Abandoned uranium mining site in Arizona. Photo by doctress neutopia.
 

Native Peoples, mostly in the American Southwest, are extremely vulnerable to the guaranteed dangers of the nuclear buildout to support AI Manifest Destiny. The existential threat posed to their lands, water, health, culture, and sovereignty is the proverbial elephant in the room that is almost never included in any discussion about AI.

The good news is, an AI economy is not a done deal. The infrastructure simply doesn’t exist yet, and there can be no industry with no infrastructure. That’s why the Trump administration, puppeteered by Silicon Valley, is now scrambling so hard to throw up data centers and their accompanying nuclear energy sources. This, along with their efforts to establish and continually expand an infrastructure for widespread surveillance and data extraction, are the structural pillars to be knocked down for most effective resistance. 

We can put the brakes on the AI juggernaut simply by blocking data center construction at every turn, and by being offline as much as possible. Stop the data centers. Stop the biopiracy. Stop the nukes. Stop the AI. Stop the surveillance. They’re all interconnected. Keep the uranium in the ground. Respect Indigenous Peoples’ rights.

—Koohan Paik-Mander is a pan-Pacific peace activist with intimate ties to Korea, the Mariana Islands, and California. She co-authored the book “The Superferry Chronicles: Hawaii’s Uprising Against Commercialism, Militarism and the Desecration of the Earth” with her late husband Jerry Mander, and has written for many publications about environment, Indigenous struggles, technology, and Asia-Pacific geopolitics.

Top image via www.vpnsrus.com