The Cloud Is Not Clean:

The myth of the cloud

We have been sold a picture of artificial intelligence as immaterial: something that lives in a shimmering, weightless "cloud." But as Professor Brevini made plain, those clouds are data centers. They are noisy, dusty, thirsty, and energy-hungry. 

The radioactive violence of the nuclear age and the extractive violence of the AI age share the same logic: communities on the frontlines bear the burden, while those profiting keep their hands clean.

The numbers are stark. Global data center electricity consumption is on track to more than double by 2030. A single request through Google's AI-powered search consumes roughly ten times more electricity than a standard search query. And behind every data center lies a supply chain of so-called "critical minerals" including lithium, cobalt, rare earth elements. Metals and minerals that are extracted from Indigenous and First Nations lands with devastating consequences for communities, ecosystems, and water.

Then there is the waste. E-waste has grown by over 80% in the last decade. The digital dumps of the Global North's AI economy are, predictably, located in the Global South, like countries Kenya, India, and the Philippines. This continues the same colonial logic that has always determined who is exposed to toxic harm and who is not.

Nuclear power is not AI's solution

In the last two years, the major tech companies, Google, Amazon, Meta, Microsoft, have made a series of high-profile announcements pledging investment in nuclear power to meet AI's energy demands. 

Professor Ramana walked us through why these announcements deserve serious scrutiny.

"...many of these announcements do not even mention how much money they are investing... they say we are part of a consortium of a group of companies or entities that are investing and together we are investing $650 million, $500 million, whatever.”

“So in the United States right now, there is not a single nuclear reactor under construction… The last time there was a nuclear reactor under construction was in the state of Georgia. This is the Vogtle project. And those two reactors that were built cost somewhere between 36 and 37 billion.” 

“…that's the kind of scale you should be thinking about if you're thinking about even a reasonable amount of nuclear energy being generated in order to meet some fraction of the demand growth that these AI and data centers are going to cause."

According to Professor Ramana, the timing does not add up either. New nuclear reactors take many years to build, and historically, most have been significantly delayed. Small modular reactors, the technology most frequently cited by tech companies, have never been successfully built anywhere in the world. There is no realistic nuclear pathway that meets today's AI energy demand.

"I would suspect that the few hundred million dollars actually comes out of their public relations budget rather than their energy budget. It's much more about keeping the public discourse under control."

What is really happening is a narrative strategy. The AI industry is watching and has responded by positioning nuclear power as its alibi, a future-tense promise that allows data center construction to continue unimpeded today.

The danger is not only environmental

The radioactive violence of the nuclear age has always been inseparable from militarism, and AI is accelerating that connection. 

From the use of AI-assisted targeting systems in Gaza, like "Lavender" that identified people for bombing, often in their own homes, and "Where's Daddy," designed to track when targets were with their families, to the integration of AI into nuclear command-and-control systems, this technology is being built into the machinery of mass killing at speed.

The people developing these tools talk explicitly about speed as the goal. Where it once took six months and a team of fifty people to generate a target list, AI can do the same work in two weeks, one person. This is not progress. This is the industrialisation of killing, and it demands the same abolition response we bring to nuclear weapons themselves.

What we are called to do

Professor Brevini named what this moment requires: building a movement that connects Indigenous and environmental groups, and those fighting for a world without nuclear weapons. 

"We have to actually keep calling these mythologies for what they are... we need to make sure that this is actually generating a movement that brings together indigenous groups, environmental groups, tech anti-big tech groups, you know, the no kings groups because this is the type of capitalism that is damaging humanity at multiple levels."

The logic of AI capitalism and nuclear capitalism is the same logic, accumulation at the expense of frontline communities, ecological limits treated as obstacles, and the public bearing costs while private actors pocket the gains.

We are seeing resistance. Communities across the United States are organising against data center construction. Young people are recognising the cognitive and ecological costs of generative AI. Civil society is reclaiming the right to ask: who decides what technologies are built, where, and at whose expense? As Professor Brevini states:

"If we really embrace this understanding of non-market forms of value, the logic of repair, the values of care that indigenous knowledge especially has been bringing to us as a kind of stewardship and guidance of how actually to manage and how to thrive as humans with environment,  I think that actually we can find answers to how to manage and how to dispel this idea that AI is something that can ever be sustainable." 

The Nuclear Truth Project's commitment to our Protocols: Rights, Respect and Reciprocity is not a niche concern within nuclear policy. It is the framework that every technology conversation requires. 

Affected communities are not subjects of study or problems to be managed. They hold lived expertise that must be centered, whether we are talking about radioactive contamination in the Marshall Islands, lithium mining in Western Australia, or e-waste dumps outside Nairobi, Kenya.