Contents

1. The Tip of the Iceberg
2. A Building Full of Computers
3. A Nuclear Renaissance
4. The Ring of Fire
5. For a Self-Organized and Diffuse Project of Struggle

Preamble

Capitalism is always restructuring itself. We all know this, and it is why analyzing and creating proposals for struggle is an ongoing process, never something that can be fully concluded. So in a sense, change is the only constant: a comforting story for a world that is often disorienting. And yet, capitalism, as simultaneously an economic and technological system, goes through phases that might last more or less long and that create a sense of stability within them: the more things change, the more they stay the same (again, as we are told).

In this moment, Summer 2026, it seems we are on the cusp of a new technological and economic phase. The economic and political systems have thrown their weight into developing a specific technology even more dramatically than for the expansion of the internet around the turn of the millenium. And unlike the early internet, this technology is highly centralized in the hands of billionaires, with no liberatory potential to speak of (although time has not been on the side of those who saw the internet as a force for freedom thirty years ago). And like all high technologies, it comes with a whole web of industrial, extractive, and financial projects, because, if you will allow us one more tired proverb, misery loves company.

That technology is of course artificial intelligence, or AI, and we surely won’t be the first to tell you that it is poised to change everything. Ever since the first AI-powered chat bots appeared a few years ago, the models have been steadily improving and increasing the number of tasks they can perform, from making memes to piloting killer drones. But the actual uses of the technology are in some ways secondary to what’s going on behind the scenes, which is a frantic race to build as many data centres as possible, to produce the chips and electricity those centres need, to control and mine the minerals used in the chips, and, of course, to mobilize the vast sums of capital needed to do all of the above.

Some people say there is a financial bubble in the AI sector right now. We hear that this is a problem and will lead to more infrastructure being produced than will be useful long term. Or we hear this is no problem because bubbles are normal and don’t typically stop the underlying technology — after all, the dot-com bubble didn’t sink the internet, and even the Panama Canal, the great grandmother of all bubbles, still ended up being built. Probably both of those things are true. But this is not a text about bubbles, because the risk is real that, regardless, AI will actually deliver on its promise, transform society, and prove all the billionaires right.

There are aspects of the civilizational push for AI that are playing out the same across Canada, the US, and Europe: the same kinds of jobs are being eliminated, the same kinds of data centres are being built with the same arguments, which triggers the same kinds of resistance, and the electricity always has to come from somewhere. There is a lot anarchists and others who oppose AI can learn from each other across our contexts, and this proposal will likely have a lot in it that applies elsewhere. But, as always, we find ourselves in a specific place dealing with specific companies, data centre projects, financial interests, politicians, energy projects, infrastructure, and extractive projects. For our analysis to be grounded and our proposals to be actionable, we need to talk about the specifics of that place. And for us, that place is Southern Ontario, Canada, in the area around Toronto.

The Greater Toronto Area is a huge continuous urban area with a population of about 8 million, making it the third largest metropolitan area in the US and Canada and home to fully a quarter of Canada’s population. It is simultaneously Canada’s industrial heartland and financial centre. We provide this bit of context to support our premise that there is no way to effectively challenge power in Canada without acting in the Toronto area, despite the near total absence of any history of combative struggle in this area.

In Ontario, the AI era is shaping up to be defined by data centres and nuclear reactors in the South, and by nuclear waste and critical mineral extraction in the North. We see all of this as part of a single process: the AI applications churning out slop and taking jobs, the old nuclear plants being refurbished and the dawning new era of nuclear construction, the 20-year hunt for Canada’s nuclear waste dump finally picking a site, the data centres popping up like weeds, and of course mining critical minerals in the Ring of Fire region. All of these are in their early phases right now, but they will happen quite quickly, especially considering the legislative changes at the provincial and federal level to accelerate major projects.

The purpose of this text is to make plain the links between these various elements — AI applications, data centres, nuclear power, mining — as they are playing out on the ground in Ontario. Because analysis is meaningless without action, this will be accompanied by a proposal for a determined struggle against the AI hell world that is being prepared for us. Our goal is to identify areas of social tension and opportunities for intervention rather than to paint an exhaustive picture of the issues.

The future is not written, and a world run by AI is far from inevitable, so we should do everything we can to make sure it stays in the realm of science fiction and the sick dreams of billionaires.

Part 1 – The Tip of the Iceberg

AI applications are the most visible part of the overall push for AI, and they are supported by all the less visible infrastructure and extraction that makes them possible. For that reason, and because there is little to say that is specific to any particular location, we aren’t going to get into too much detail about these applications and will just make a few general observations.

Work. AI is indeed poised to transform or eliminate a lot of jobs, but as people who are opposed to work, to the capture of life by the economy, is this really something we should be upset about? And we could reply that we aren’t defensive over our work or the tasks that make it up — in fact, we crave nothing more than to finally put an end to the useless, destructive activity we are engaged in every day. What we observe, though, is that many people who had previously been included in the supposedly productive work of society — and who were, in many cases, well rewarded for this — are going to join the ranks of the excluded, those with no valuable skills, whose labour has become flexible (precarious) and competitive (low wage). If AI fulfils its promise, we are poised to see a restructuring of work greater than that brought about by computers or the internet, as whole sections of the workforce are eliminated or boiled down to a small number of workers overseeing AI output.

This has an effect of de-skilling the population. Those of us who used to write, translate, draw, or design may find ourselves instead editing machine output. Similarly, those who write code, do tech support, or conduct research may end up essentially machine operators, guiding the AI through tasks that we would no longer have the skills to do ourselves. AI may well come for truck drivers and pilots, architects and engineers, data analysts and mathematicians, musicians and artists, doctors and therapists. It doesn’t mean that these skills will be lost entirely — there were still people who knew how to weave on equipment they could make themselves after the mechanical loom had taken over the industry — but they will be more rare and much less valuable. This will have many impacts, but one will be similar to what happened to the manufacturing sector as supply chains became global — less control over the work, less possibility for work as a site of autonomy, fewer skills gained through work that are useful outside of it, less ability to produce anything of value. All this makes social revolution less possible, as it becomes harder to break free of structures of domination and rebuild autonomy.

We don’t want to live in a world like this. In fact, our activity is already so alienated that we feel we should have already been pushing to take back more of the activities necessary to our daily life. AI is just going to make this problem vastly worse, in addition to pushing more people out of the productive economy and into precarity. Whether this exclusion is addressed by the powerful through something like a universal basic income is beside the point.

Social control. At this point, we have all heard about the ways that the sadly ubiquitous security cameras are being networked and fed into AI to make for more effective surveillance. No more will police have to go around to homeowners asking for footage after an event has occurred — now, they will already have access to as much footage as they want and can search through it almost instantly with AI, identifying people and tracking them across multiple cameras or creating profiles of their movements and associations over months. In some contexts, this process will be overseen by judges and in others it won’t, but this frankly makes no difference. If the state puts rules on its use of surveillance technologies, it is only to maintain its own legitimacy based on what the public is prepared to accept, not to tie its own hands and put certain tools out of reach.

Military applications. What else is there to say about the emergence of autonomous weapons or the use of AI in selecting targets for aircraft piloted by humans? Some AI companies have drawn a theatrical line in the sand about whether their incredibly destructive technologies can be used in these ways, but it makes no difference whether this or that billionaire wakes up and decides to have ethics one day. AI is already being used by the military in the US attack on Iran and in Israel’s destruction of Gaza and Lebanon, and this will only grow as the AI sector itself continues to grow. The development of AI applications that spit out slop memes for social media also feeds the development of AI that is used to kill people around the world — it is not that the two are moving in lockstep, but that they are simply the same technology.

Alienation. So much of human society has already been captured by the internet. What does it mean that we can no longer be sure we are interacting with a real human in that space? Some say that if you can’t tell you are talking to an AI then it makes no difference whether or not you’re talking to an AI — but this seems more like a condemnation of the alienation inherent to the internet than a defence of AI. When I communicate, I want it to be with a person who understands. When I read or listen, I want it to be a sincere expression of someone’s thoughts. When someone is helping me, I want it to make sense for me to help them in return or to express gratitude. The rise of AI makes much of online communication into something of no value, into slop. Every time a chat bot wishes us a nice day at the end of an interaction, something inside us dies, because the bot not only does not wish anything, it doesn’t care about our well-being in even the most banal and generic way that we care about the well-being of the customers when working at a cash register. People are already lonely and cut off from anything that might feel like community, and AI is already making that far, far worse.

There is, of course, a lot more that could be said here about why the AI applications we are supposed to be so excited about are bullshit. We’ve met enough university instructors giving up on grading assignments themselves and getting AI to do it because all the assignments are written by AI anyway — at best, it is slop all the way down. At worst, AI means being tracked by AI-enabled cameras before being assassinated by an AI-guided drone. However, focusing on the AI applications can make the situation seem hopeless, despite the evident social tension around them. There is plenty of opportunity for intervening against AI and its world, but it will just involve looking beyond the screens.

Part 2 – A Building Full of Computers

There are already about a hundred data centres across Ontario, most of them quite small and often hosted within office buildings, but this is going to change. There is a wave of large data centre projects coming to Southern Ontario, notably 15 proposed as of 2025 that will, combined, use as much electricity as 2.2 million homes. And these are just the ones about which there is information available — there are few transparency requirements around data centres in Ontario, despite the fact that they are required to apply for permits to connect to the electrical grid.

The Government of Ontario estimates that data centres will eat up about 15% of all new power generation over the next ten years (which will almost all be nuclear, as we will discuss in Part 3). This is likely an underestimate, as it is based on existing investment and development patterns, and these dynamics will probably continue increasing for some time (unless the people predicting an AI financial collapse are right). Do we just accept that the use of electricity will continue to go up? That, in the face of climate change, society will continue to exploit the Earth ever faster? And not for any grand purpose, but for AI billionaires to get a bit richer?

Data centres don’t just run AI though — the whole internet requires data centres. The other main reason for the rapid construction of data centres is for cloud services. AI and cloud services are closely related, since AI requires vast amounts of data and the cloud allows lots of data to be organized and stored in a readily retrievable format. To return to AI-powered surveillance, this requires cameras to be networked to a cloud — devices in different locations pooling their data, which, although it may be stored in different physical locations, can then be accessed all in one place. This is as opposed to local storage on a device you physically control. Cloud computing also involves using the processing power of many computers to carry out demanding tasks, like breaking encryption on devices seized by police, to name just one totally random example.

With all that said, let’s take a look at those 15 big data centre projects in Ontario’s immediate future:

• Napanee
  • Napanee Data Park, 250 MW, owned by Napanee Environmental Complex
• Milton
  • Milton Data Centre 720 MW, owned by Logistics Land Investment, completed by 2032
  • Datacentres Milton-Substation, 342 MW, Bird Construction, completed by 2028
• Hamilton
  • Steelport Data Centre, 180 MW, owned by Slate Asset Management
  • It seems like there is a second, larger data centre also being planned for the same site, but nothing has been made public yet
• Toronto
  • StackTORO1, proposed expansion to 56 MW, owned by Stack Infrastructure, completed 2027
  • Yondr TOR 1X0, 27 MW, owned by Yondr, completed by 2026
  • Microsoft Etobicoke, around 55 MW, completed by 2026
• Markham
  • Microsoft YTO14 Campus, 96 MW, completed by 2028
• Richmond Hill
  • Urbacon DC5/6, 70 MW, 2026
• Vaughan
  • Microsoft YTO12, around 48 MW (early design phase)
  • Microsoft YTO11, 48 MW, completed by 2026
• Brampton
  • 55H Data Centre, expansion to 20 MW, owned by Grain Management and Stratcap, operated by Core Data Centres completed by 2026
• Cambridge:
  • Ascent TOR1, expansion to 60 MW, owned by Ascent, Related Digital, and TowerBrook Capital Partners, operated by CoreWeave, completed by 2026
  • Intermarket CAM data centre, 30 MW, Intermarket Properties, completed by 2026
• Mississaugas of the Credit First Nation
  • Mikinak Data Centre, 200 MW, Mississaugas of the Credit and Red Jar Energy, completed by 2027

This list teaches us a number of things. One is that there are, so far, only a few “hyperscale” data centre projects (over 100 MW) proposed for the province, alongside a multiplication of mid-sized ones. This could change in the future as investors gain more confidence in the regulatory environment and the stability of energy prices. At the moment, Ontario has favourable energy prices, as the largest energy consumers in the province pay a wholesale rate for energy that is substantially less than the cost paid by households (about 2/3). Despite household energy prices in Ontario being relatively high, the largest consumers pay less in the Toronto area than they do in all but a few Canadian and American major cities. However, Ontario’s largest power generation company, OPG, has asked regulators to allow them to nearly double the rate they charge for electricity over five years.

Another thing this list shows us is that there seems to be only one hyperscale data centre being proposed in a dense urban area: the Steelport project in Hamilton, which is proposed for the site of a former steel company in the city’s industrial portlands. Slate Asset Management seems to be mostly talking about a data centre specialized in AI-driven academic research while it tries to get the zoning changed for its site, but there are some hints that they are also quietly planning a much larger data centre there once they have the permits in place. The site, on the shore of Lake Ontario, has abundant water for cooling, is already set up to draw large volumes of power, is at a distance from residential neighbourhoods, and is an improvement over what was there before in terms of air and noise pollution — it seems like a perfect fit. And yet, there is potential for a social tension around this project, as it has already drawn hundreds of people to protest and speak at City Hall. Hamilton city council has responded with a moratorium on data centres, but this likely won’t last, whether because council changes course or because it gets overruled by the Province.

Considering how few requirements there are for transparency around data centres, it seems much easier to hide data centre projects in industrial parks outside the city than to risk opposition by being as visible as Steelport. There is no provincial consultation process for data centre construction, and the energy regulator (Independent Electricity System Operator, or IESO) does not publish details about requests to draw large volumes of power. In terms of regulation, the Province recently instituted an accelerated process for major projects over 10 MW, which includes helping them connect to the grid, so there is clearly a desire to open the door to data centres provincially. As well, municipalities across the province are preemptively changing their zoning laws to facilitate data centre growth and avoid contestation like in Hamilton — one example is Napanee, which has moved to allow them by default in all areas zoned as industrial, which removes a layer of per-project transparency. Basically, people may not care about most of these projects because of where they are located and because of the lack of information.

Local resistance to data centres typically focuses on noise, water pollution, and increased demand for electricity driving up prices for individual consumers. All of these things seem true, but when compared to a steel mill, for instance, a data centre is a good neighbour. These site-level nuisances don’t explain the scale of resistance to a project like Hamilton’s Steelport. We see this opposition as a manifestation of a deeper social tension against the AI era itself, one motivated by both environmental and social concerns, and it is a tension that won’t go away based on what happens with any specific project. Rather, we anticipate the tension spreading, and data centre construction will remain a flashpoint. A question for us, then, is how do we expand that tension to other aspects of the broader push for AI to create even more opportunities to intervene? How do we encourage an analysis and terms of struggle that go beyond not-in-my-backyard concerns? How do we model and spread self-organization and offensive action?

One advantage of thinking about the push for AI more broadly than just AI applications and data centres is that it opens possibilities for diffuse action and for solidarity between people who are differently impacted. Data centres need chips, which are made of critical minerals, and they also need abundant electricity, which in Ontario will mostly come from nuclear. People protesting in Hamilton are struggling against the same beast as those in Northern Ontario who are preparing to resist open-pit mining, or those east of Toronto who will be having a massive nuclear plant built next door. We don’t need to approach any of these struggles as occuring elsewhere, as having a front line and a most-affected community that the rest of us must be content to support. There are countless opportunities to resist the push for AI — we just need to make the links visible between its different manifestations through words and actions.

Part 3 – A Nuclear Renaissance

Nuclear power in Ontario has kind of faded into the background over the decades. There are nuclear plants spread out across the southern part of the province, quietly providing half of the electricity consumed — much more than the more famous Niagara Falls. It is like society has collectively forgotten all the reasons to oppose this technology — or at least, we were content to continue using the ones built in the 70s until the end of their lifespan. But now, the old nuclear plants are being renovated, small modular reactors are being developed, and a massive new nuclear plant is in the early stages of planning. From producing over half of Ontario’s power today, nuclear is projected to produce 75% by 2050, meaning Ontario will overtake even France and become the jurisdiction most reliant on nuclear power in the world.

This nuclear renaissance is mostly happening quietly, so far, but the massive new nuclear plant (being described as potentially the world’s largest) will likely bring all the downsides of nuclear back into focus. Far from being as safe and clean as it is reputed to be, nuclear accidents, when they occur, render areas uninhabitable permanently, and these nuclear plants are all close to Canada’s largest urban area. What level of risk is actually acceptable? (Especially when the risk is being taken to fuel nonsense like AI data centres.) And it is not that nuclear power doesn’t produce waste. The concrete used to make them produces large amounts of carbon dioxide, and then there is the radioactive waste. The spent nuclear fuel is so dangerous that there is no way of making it safe, and the only so-called solution is to bury it deep underground. Forever. So the obvious question, then, is where this poisonous waste will go. And wouldn’t you know, Canada has been working on that problem for the last 20 years and finally has a proposal.

The Nuclear Waste Management Organization was tasked with creating a plan for dealing with Canada’s nuclear waste in 2005, and they have been shopping around for a dump site since that time. Obviously, every community considered had no shortage of good reasons why their community, specifically, was not the right place. But now, the NWMO has proposed a site near the township of Ignace and Wabigoon Lake Ojibway Nation in Northern Ontario, 250 km outside of Thunder Bay. (For context, Ontario is really big, and Thunder Bay is about as far from Toronto as Tampa Bay, Florida.) The local First Nation and the township of Ignace have agreed to the plan, and a federal impact assessment is currently in progress, so this project is still in its early stages. But it is already attracting opposition, though mostly of a bureaucratic kind: other First Nations are saying their consent is required as well, and people have been flooding the review panel with negative comments.

The proposed dump would be active for 160 years (though of course the buried waste will stay radioactive far longer), and it will receive two or three truckloads of nuclear waste per day for at least the next 50 years from across Eastern Canada. This means trucks full of nuclear waste driving all the way from the Gentilly plant in Quebec or up from Southern Ontario across the entire length of the province, every single day. On this kind of scale, an accident starts to feel like a virtual certainty.

Currently, the nuclear waste produced in Canada is being stored on site where it is produced, but that isn’t safe or permanent. So it needs to go somewhere, right? This question is an example of how the nuclear industry holds society hostage once it gets started. Of course, something needs to be done with nuclear waste, but its dangerousness can’t be used to force us to go along with the continued expansion of the nuclear industry. The nuclear dump should be opposed no matter where it is built and no matter how safe it is, except in the context of ending the use of nuclear power once and for all.

Another aspect of nuclear power is, of course, nuclear weapons. Canada has no nuclear weapon program, but there is a history of sending nuclear fuel from plants in Canada to weapons producers in the United States, and the underlying technology is intimately tied to the arms industry. Similar to how the development of AI is always the development of AI weapons, the civilian nuclear industry is hand in glove with the military. We see this with Canada’s famous Candu reactors, which are sold to the Indian government who uses them to make nuclear weapons. Canada is working hard to develop and build small modular reactors, which have the potential to further spread nuclear arms (including the privately owned nukes being dreamed of by a subset of tech billionaires). Nuclear war is probably the greatest risk we all collectively face, so let us ask the question again: With the stakes this high, what level of risk is acceptable with nuclear power?

Here is a list of nuclear energy projects currently underway in Ontario and a few details about them that we found interesting:

New production

• Canada’s first small modular reactor is currently under construction at OPG’s Darlington site (to be completed by 2029), and it is meant to be one of four, with each producing 300 MW of power (enough for 300,000 homes). They are able to use the existing electrical grid infrastructure from the Darlington nuclear plant. They are being pitched as partnerships with nearby First Nations.
• OPG’s Wesleyville site in Port Hope may become the home of potentially the world’s largest nuclear plant, on land long owned by OPG and Hydro One on the shore of Lake Ontario. The site isn’t fully confirmed yet, but local governments have approved it (including First Nations), and construction is expected to start in 2030 (OPG has built a “nuclear discovery centre” in Port Hope in the meantime). It is close to the Durham Region Energy Corridor, which already connects to nuclear sites in Pickering and Darlington.

Refurbishment of existing plants

• The Darlington Refurbishment Project just finished after nine years of work and extended the lifespan of the reactors there by 30 years.
• Bruce Power is Ontario’s largest nuclear plant. It is made up of eight units producing 6,400 MW, or 30% of Ontario’s electricity. Their lifespan is being extended by 30 years each through phased construction, with work on Unit 3 just completed and Unit 4 underway.

It seems extremely difficult to intervene in the refurbishment projects, and there is no social tension around them of any kind. However, the Wesleyville site is already facing some local opposition from locals who remember being poisoned by the uranium mine that used to exist near Port Hope. This may escalate into a genuine tension, but that will likely depend on how much people are able to dramatize the issue and make it controversial. Considering the history of opposition to nuclear in Ontario back when all these plants were being built, it does seem like that is possible, although the threat posed by nuclear weapons doesn’t motivate people as much as it used to.

That said, the new nuclear plant project is so large and so important that finding ways to oppose it feels like a priority, and one way of doing this is to draw attention to the link between AI and nuclear energy in Ontario. One way to slow AI development in this province is to drive up the price of electricity for big consumers, and the surest way to do that is to prevent new generation capacity from coming online. Nuclear power is very expensive to produce, and the plants take decades to pay for themselves, so controling the price of electricity is already a major problem in Ontario. The Ontario government currently spends $7 billion a year subsidizing energy prices to mask the high cost of nuclear and delay a ratepayers revolt for as long as possible.

We aren’t politicians, and so aren’t in the business of coming up with policy solutions to massive problems like how to produce enough electricity so that everyone can commute to work in electric cars or so that “green steel” becomes realistic. The best solutions to the problem of how to reduce the amount of energy society consumes will come from cycles of crisis and revolt that give us all the chance to experiment with new ways of relating to each other and of surviving that are less disastrous. Every nuclear plant that gets built makes a liberatory future harder to imagine, because of the eternal liability they represent. Which is especially hard to justify when all those small modular reactors being built in Darlington won’t even cover the power needs of the data centre projects we know about.

So to be clear, we aren’t arguing that opposing nuclear is THE key to opposing AI or that opposing AI is the only thing that matters. Our goal is always for people to struggle against domination in all its forms, and we are proposing a focus on AI as one impactful way of coordinating our efforts and contributing to social tension. People who want to engage with this proposal should struggle against whatever aspect of the push for AI they are close to and feel motivated by — using decentralized and self-organized forms of struggle and without compromising in their critique, of course. We do want people to struggle against nuclear, though, and this isn’t just because of AI. The expansion of energy consumption should be resisted, and as we said above, nuclear specifically holds society hostage and makes social revolution less possible, because it requires constant attention by specialists to avoid unimaginable disaster.

Part 4 – The Ring of Fire

The Ring of Fire is the name given to a large area in the Hudson Bay lowlands, which is the world’s largest undisturbed peatlands, which is in turn part of the world’s largest boreal forest. It is about 500 km north of Thunder Bay, far beyond where the roads end. Almost all of the Ring of Fire is in the basin of the Attawapiskat River, which flows to James’s Bay where the Attawapiskat First Nation is found, a community already dealing with serious water contamination. The area is called the Ring of Fire because of the roughly ring-shaped concentration of critical mineral deposits below the surface. The Ring of Fire is potentially very large, at 5,000 km2, though so far most concretely identified deposits are in a 20 km-wide strip of land. So far, there have been discoveries of chromite, nickle, copper, platinum, gold, zinc, and palladium.

The Ring of Fire is in Treaty 9 territory, and there are nine First Nations in or around it. Most of these Nations are part of the Matawa First Nations and are represented in this matter by the Matawa First Nations Management Tribal Council. The provincial government had initially been planning to designate the area as a special economic zone, allowing them to bypass the usual Indigenous consultation requirements and environmental assessments, but following protest from First Nations, the government revised itself and now says that it won’t use the measure, supposedly because local First Nations are supportive.

There aren’t a ton of details about what negotiations between First Nations and the Province have looked like so far, but one major piece is that development in the Ring of Fire will come with all-season roads, better electrical infrastructure, and high-speed internet access for these remote communities. Notably, three major road projects are either underway or about to begin: the Marten Falls Access Road, the Northern Road Link (going from Martin Falls to the main Ring of Fire sites) and the Webequie Supply Road.

To us, it is understandable that a lot of people would value reliable road access to the rest of the country (which means lower prices for all goods) and a level of internet access closer to what is standard elsewhere. And if this is being made contingent on their area being “open for business,” then it is also understandable that people would accept the project. But why do these services have to be contingent on strip mining the area? The process of displacement of Indigenous peoples in the North and forcible settlement onto reserves was always meant to be “out of sight, out of mind,” but now their land is valuable and further displacement isn’t politically possible, so the state offers a few small measures to some of the most excluded people in the Canadian territory.

It is not a problem that some residents of First Nations will support these projects, whether it is for jobs, for the promise of roads and internet, or for any other reason. That is because there will also always be people who refuse to accept projects like these and who are determined to fight back, even against the roads that will, we are told, bring so much good to the communities. In June 2025, members of Neskatanaga and Atawapiskat First Nations set up a settlement in the path of the projected road development, where there will be a bridge on the Attawapiskat River. Although it doesn’t seem like there is always a presence there, the intention was for the settlement to be permanent.

As always, our task as anarchists is to identify points of tension, spread ideas and practices, and show solidarity to other rebels through action. There will certainly be Indigenous-led blockades against the Ring of Fire, and these blockades deserve support. However, it is also important to avoid repeating past patterns of how anarchists have related to Indigenous front lines that have been too limiting. By all means, go there, meet people, build relationships. But let’s not treat those blockades as the only sites at which struggles occur, and let’s not go looking to their participants to direct how the rest of us struggle away from those sites. We are able to act against the Ring of Fire development — and the world that needs it — starting immediately and in a diffuse manner, attacking the interests and infrastructure that are tied to it and also to the push for AI as a whole. We don’t need to wait to be told it is okay. And we don’t need to travel to Northern Ontario to do it.

We want to see a struggle where actions against nuclear and data centres in the south are widely understood to be in solidarity with actions against the Ring of Fire in the north. For this link to exist, it will require both words and actions as well as actual relationships between people in the north and south. Mining companies and governments have known about the Ring of Fire deposits for a long time, but they are only ramping up the pressure to develop now in response to specific pressures related to international trade and the growing importance of these minerals for computing and electrification. It is very difficult to act in the far north if you don’t live there (and even if you do, it seems challenging), and only about 8,000 people live north of Lake Nipigon. We can’t leave it entirely on them to fight this project, and our efforts will definitely fail if we narrow our imagination to acting only near the Ring of Fire.

We need a way of understanding the Ring of Fire that is broader than just the mines and roads themselves, and we are proposing the push for AI as one lens for approaching this. Others have made proposals that electrification and the green energy transition are the broader tension, but we believe AI is a more fruitful framing for a couple of reasons. One is that, currently, the green energy transition in the US and Canada is state-driven, rather than economically driven, and so it will be put on hold any time it becomes inconvenient, and we are currently in one of those moments. Another reason is that AI promises to upend a lot of aspects of our lives, whereas the green transition is all about keeping our lives the same while being a little less ecologically devastating. We believe that a genuine social tension exists around AI, and that isn’t true around the green transition (although local projects will always produce local opposition).

Because the Ring of Fire is so far away from where we are (and almost certainly where you, the reader, are), we feel that a specific struggle would not be feasible. However, practices of diffuse attack could succeed at dramatizing and escalating the tension around that specific project and linking it to other major struggles elsewhere in the province.

Right now, the provincial government is the main actor in the area, but there are a large number of mining companies involved too. We want to mention just two of them here: Wyloo and Juno Corp. Wyloo, as it is currently called, owns the claim to the Eagle’s Nest site and its deposits of nickle. This will likely be one of the first concrete projects to move ahead once the infrastructure is in place. Juno Corp. is an exploration company, and it has laid claim to almost the whole Ring of Fire, 4,600 km2. Their role in this project is critical, even though they will likely hand off individual deposits to actual mining companies as the project moves ahead.

The Ring of Fire, if it is built at the scale being envisioned, will become one of the most significant extractive projects in Canada. It will involve open pit mining that clears away huge swaths of the boreal forest, the lungs of the Earth, and draining the peatlands, which are vast carbon sinks. For those of us who care deeply about stopping the ravages of industrialism and moving towards a more livable world, we can’t let this project’s remoteness stop us from attacking it and showing our solidarity with those resisting on the ground.

Part 5 – For a Self-Organized and Diffuse Project of Struggle

We’ve spent a lot of words now describing how some major projects in Canada’s most populous province are interconnected and in the service of a major technological transformation of society driven by elite interests, which we’ve called the push for AI. We’ve also offered some starting points for how to struggle against this process, and in this final section, we will get into those starting points in a bit more detail. Briefly, here is what we propose:

• A diffuse struggle against AI-related projects and infrastructure across the province (or more widely if this proposal speaks to you where you are)
• Solidarity between the northern and southern parts of the province through actions against projects in our respective areas
• Constantly finding creative ways to make the link between AI applications, data centres, nuclear, and mining
• Spreading practices of attack
• Acting at sites of social tension while remaining autonomous
• Informality and self-organization

First, though, what is a project of struggle (also called a projectuality)? Basically, it is the determination to organize informally and act towards a specific goal over the medium term, so about three to five years. To be a truly autonomous project of struggle, we need to prioritize direct action, which we would argue should be destructive or disruptive to the process being targeted. This is easier said than done, because it is easy to end up in a position where you are basically “lobbying with bricks” by doing escalated action with a symbolic goal. If our goal becomes to make a powerful person change their mind, then our project of struggle is no longer autonomous, and we need to take some time to reassess and figure out how to make it so our actions directly disrupt the process in question (which in our proposal is the push for AI).

Communication is also important to a project of struggle, both with people with whom we already have affinity and with people broadly in order to reach those who might enter into struggle on similar terms. This can take the form of leafleting at rallies, speaking at meetings, publishing on anarchist news sites or in anarchist journals, starting a street newspaper, putting up posters and stickers… There are lots of possibilities. But the important thing is to be clear and uncompromising — there is no point entering into a struggle like this if you are going to pretend to be a reformist or like your opposition to AI is limited to this or that specific project. It is important to defend action, make links between different projects, and advocate for self-organization (people should organize themselves, and not be organized by those who have, or aspire to have, power). It is also important to rely on communications channels we can control so that the medium doesn’t shape our message based on their written and unwritten rules: posters and street papers, not op-eds in mainstream news or interviews with journalists; counter-info websites, not corporate social media. This might make it feel harder to get the word out, but we promise it is more effective long term, especially if the struggle escalates.

Now, what do we mean by diffuse struggle? Basically, a project of struggle can either be diffuse or specific. A specific struggle is against a particular project in a particular geographic area, and it needs to define its focus quite, well, specifically in order to be effective. There have been attempts at specific struggles in Southern Ontario, notably against the Line 9 pipeline during the 2010s, but we don’t believe this is the right approach to take here. Both the Ring of Fire and the new nuclear facilities (especially the dump) would be good targets for a specific struggle, but the ones in the north are extremely remote, and even the nuclear plants happening in Eastern Ontario are quite far away from where most comrades live, and they are less likely to have local tension of a scale and character that could easily serve as a starting point anyway. It will probably be important, for a diffuse struggle to succeed, for us to learn to act farther away from where we live, but Northern and Southern Ontario are just too far apart, and focusing on one project would mean overlooking many equally good opportunities closer to home that might be easier to act against consistently. As well, there is no single project that is particularly critical — no one data centre is more important than another. It is possible a specific struggle will emerge around one or more of the projects mentioned, but right now, the tension and criticality aren’t there.

Diffuse struggle, on the other hand, defines its target more broadly and spans multiple different (but related) sites of domination in different places. We believe it is appropriate in our case because there are related projects happening all over the province, and it allows us not to simply show solidarity with people in the North, but to be part of the same struggle in different locations. Data centres in particular are interesting in this way, and they have many opportunities to target infrastructure that is both all over the province and often quite critical: fibre optic cables and electrical infrastructure. This will require skill-building, and this process of researching and sharing techniques can begin immediately. Diffuse attack allows for more trial and error, since there are virtually unlimited potential targets that allow us to resist the push for AI.

(There is a lot to say about the need to share skills. In our opinion, the most urgent ones relate to operational security and reducing the risk of DNA contamination during actions. The ability to act is not useful without the ability to not get caught. The website notrace.how is your best friend for this. We should also note that people use the terms “specific” and “diffuse” struggle in different ways, which is why we wanted to spell out what we mean. Our intention is not to impose our definitions on others.)

We want to spread practices of attack, of literally destroying the things we wish did not exist. This is not to say that other kinds of action don’t matter or are less important — far from it. Rather, we see attack as a missing ingredient in what has been happening in Ontario in the last five or so years, and a determined project of struggle is the perfect way to build up a continuity of action, help people get their sea legs, and bring back some offensive capacity in the anarchist milieu here. In this, we propose moving away from targeting the facades of buildings (windows, locks…) and towards targeting infrastructure (fibre optic and electrical cables), though there are a lot of valid ways this escalation could go. We should try to be specific in what we aim to accomplish and precise in execution, while recognizing that sometimes actions will necessarily impact more than just the target, and this reality should be approached deliberately.

We leave it to other texts — and to people’s actions — to further clarify what attack can look like, but it feels important to talk about how practices spread. Even in a context like Southern Ontario with little history of resistance, there is still a common imaginary of what resistance can look like. In our context, this mostly looks like holding protests, but also includes land defence blockades and, more recently, sabotage targeting vehicles (the wave of Tesla vandalism even made it to the Toronto area). New practices enter the imaginary when we see them play out repeatedly in contexts similar to our own and when they speak to a tension we also feel. Bringing practices of self-organized, small group attack into the imaginary would be a huge step forward, and other proposals in this direction have been made by anarchists elsewhere in Canada. It is not that practices need to spread to be effective — our actions should have value in themselves — but developing practices that enter the imaginary and then are reproduced when people decide to enter into struggle is a major goal.

We imagine a sequence of struggles characterized by regular disruptions to centralized systems that would lead to more people having backup plans for living without those systems and, due to moments of mass participation in upheaval, experimenting with different forms of life. That participation piece is important, because there is nothing about discomfort in itself that makes people look up at the night sky, see the stars, and realize that we don’t need any of this shit. Without a broader social tension and opportunities to engage, disruptions are more likely to seem like impositions that require a law and order response. We aren’t saying that actions need to be understood by everyone to have value, but we also aren’t saying we should just ignore whether or not they are understood.

While we were working on this text, another proposal was published in the journal Petrel about preparing to attack AI-driven robots as they get deployed in our communities in coming years, because these actions have a high potential to spread. We feel that proposal is very much in the spirit of our own, and there is space to move in both directions. One difference we see is between targets that are everywhere and relatively accessible but not very critical to the systems they are part of (for instance, a self-driving car) and targets that are high criticality within the AI system (for instance, an important internet cable). Anarchists can contribute to bringing actions against both these types of targets into the imaginary. We surely won’t be alone when it comes to destroying delivery drones, but there needs to also be a place for actions that, in the present, only anarchists are likely to take.

Throughout this text, we have tried to identify areas where we think social tensions are likely to emerge, and that line of inquiry should be an ongoing process. Ideally, we would like to be a bit ahead of the curve, identifying areas of tension as early as possible so that we are ready to act when they start to bubble up. But this inquiry is only meaningful if we insist on staying autonomous and not lending our skills to organizing that doesn’t share our goals (helping the good citizens lobby city hall, for instance). When we do this, we risk ending up the pointy end of the spear for reformist demands (engaging in symbolic attacks on a contractor who people are trying to pressure) or watering down our analysis under the pretext of making it more accessible (arguing over the location of a data centre or for a different kind of “green” energy). We need to establish our own goals, priorities, and timelines and not let them be dictated to us.

As well, we need to distinguish between opposition and tension. Opposition is a built-in part of the democratic system and is not actually a threat to power. This includes when citizens organize to pressure government. Social tension can exist within citizen’s movements, though, if they contain ideas that go beyond the limited framework democracy provides (such as hostility to AI as a whole, rather than just particular projects) or when disruptive practices start to spread. We want to identify social tension precisely because it has the potential to overflow the limited space democracy provides to opposition and become something autonomous and threatening to power. We want to give people an option that isn’t either the good citizens or staying home, and in our experience, anarchist ideas and practices of attack can be readily taken up in the right circumstances.

As a rule, don’t let yourself become part of the left. Say the things that only anarchists will say, and then match words with actions. Your freedom of action is precious, and the left will only parasitize you and funnel your energy towards reformist or electoral goals. Even more depressingly, they might try to get you to join their org.

A lot has been written about informality in recent years, and we won’t be trying to replicate it here, but basically, our organizing structures should be flexible, adapted to the nature of the actions we want to take, and based on affinity (which is actively built and maintained). Groups form around projects and don’t need to outlive the desire to do those projects. When maintaining the structure is seen as a goal in itself, rather than in service to action, something critical has been lost, and it is worth slowing down and reflecting.

One critique of informal organizing is that it lacks a unified strategy. It is totally possible for groups to communicate with each other — often, this is done totally openly through journals or counter-info sites, or, if relationships allow it, through meetings that might bring together several affinity groups. This communication is enough for some coordination; however, it is worth calling into question the idea of strategy in the context we are discussing — why is it important to all agree on a single course of action? It seems more interesting to us to build a common understanding of the situation we find ourselves in, circulate proposals, elaborate a project of struggle, and then encourage groups to take initiative within that. A set strategy would just get in the way.

We have a chance now to initiate a project of struggle before most of the big developments in the push for AI get started in Ontario. This will make it more possible for anarchists to resist the impulse to prioritize so-called front lines rather than targeting the infrastructure of domination that’s right in front of us and thus avoid the same limitations past struggles have had. We have a chance now to build up dialogue and a common analysis between Northern and Southern Ontario so that, in coming years, we can answer action with action and throw sticks into the wheels of technological development in this region. This will make us all less isolated and more powerful. We have a chance now to get out in front of the NIMBYs and leftists to offer a fundamental critique of AI and its world. It is much easier to get clear about our ideas and perspectives early in a struggle, rather than once the space is swamped by leftists and not-for-profits who wish for a gentler, greener system of domination.

The push for AI is playing out all over the world, and everywhere there is resistance. We can look to contexts similar to our own for inspiration, and we can aspire to be that inspiration for others. There is nothing inevitable about the future that the state and its billionaire allies are preparing — no matter how bleak things get, the situation isn’t hopeless. We have the ability to organize ourselves, communicate, and go on the offensive, and in doing so, we carry the possibility for radically different ways of life beyond anything the AI hell world has to offer.

From North-Shore