In January, the Rural Municipality of Ritchot in Manitoba publicly asked the developers behind a proposed hyperscale AI data centre to engage with the community.
The project, a partnership between Las Vegas-based Jet.AI and Vancouver-based Consensus Core, would have been powered by natural gas turbines on 142 hectares (351 acres) of farmland just north of Île-des-Chênes.
The municipality said in a statement that it had received only an outline of what the developers wanted to do, with no formal application and no rezoning request. It had also asked developers for community engagement so residents could ask questions and get accurate information.
By June, a petition against the project had reached more than 13,500 signatures, and Premier Wab Kinew killed it.
Consensus Core CEO Wayne Lloyd said in a statement that the company is reviewing Kinew’s decision and wants to begin a “robust engagement process” with the province. He said this after the project was shut down.
The Manitoba rejection is one project among many proposals across the country, and the pressure on developers to build more is intensifying every month. AI is driving a massive surge in demand for computing power, which means more data centres and more electricity to run them.
The federal government’s national AI strategy, released earlier this month, forecasts Canada will need enough AI computing capacity by 2030 to run the equivalent of roughly four million homes, with a portion of that coming from publicly backed projects.
Demand is not slowing, communities have real questions, and the cost of skipping consultation is paid by the developers who skipped it.
That conversation is one oil and gas has been having for 60 years, through pipelines, refineries, and projects that hit the same walls AI data centres are hitting now.
Last month at Upper Bound, a large-scale AI event hosted in Edmonton each year, four veterans of the energy sector sat down on a panel to tell the AI industry to look at their scars first.
Mark Little, the former president and CEO of Suncor who now leads Jotson, was joined on stage by energy economist Peter Tertzakian, Parminder Sandhu, who leads AI deployment at Innerva, and moderator Marla Orenstein, director of partnerships and impact at the Energy Futures Lab.
The evidence for what they were saying was already showing up across the country. In the past 90 days:
The complaint in each place involves consultation, as communities said they were not consulted in a way that mattered.
The Global Energy Show is taking place in Calgary this week, and data centre content has been part of the agenda every day.
The energy sector spent decades developing the playbook AI now needs, and its leaders will be back in conversation about what comes next.
What the energy sector learned
Orenstein has spent her career working on Canada’s energy future. At the Energy Futures Lab, she sees the same arc developing in AI that the energy sector has already lived through.
“If you were to look around at all the industry sectors that exist today, and you were to try to find the one that has the most parallels to AI and how it’s developing, I would argue that the energy sector is where you should be looking,” she said.
Orenstein laid out the parallels.
First, the speed of rollout once the underlying technology was figured out. She noted that it delivered obvious and measurable value to everyday life that made adoption inevitable. AI is moving through the same arc, with hyperscale infrastructure being committed before regulators, communities, or utilities have caught up.
And then, the money.
“It triggered an immediate gold rush, a flood of entrepreneurs, wildcats, and capital that just came rushing in to chase a stake in something where nobody fully understood the implications,” she said. “It generated extraordinary wealth, but concentrated it, at least initially, in the hands of a few early movers and infrastructure owners.”
The AI buildout is following the same pathway. As Little said on the panel, the most lucrative parts of AI are sitting “in organizations where the public can’t win” and the value is locked up in private markets retail investors cannot access.
Orenstein also pointed to the regulatory and societal costs that came later. New regulatory and legal frameworks were required that governments were completely unprepared for.
“It created externalities that weren’t priced in for decades,” she said. “Social, environmental, and geopolitical costs that the original innovators never had to answer for.”
She described energy as a technology that became so foundational to modern life that it stopped being an industry and became infrastructure.
AI is heading toward the same status as the technology moves into the foundations of how businesses operate, how governments make decisions, and how people work.
Tertzakian, energy economist and founder of Studio.Energy, said the parallels between energy and AI are direct because AI itself runs on energy.
“AI is 120% an energy story,” he said, describing data centres as “massive energy heaters.”
Every unit of electricity that goes in comes out as heat, and another fifth on top is needed just for cooling. That makes AI an energy industry in everything but name.
As a result, the same questions about siting, emissions, water, grid capacity, and community impact that have shaped pipeline and refinery debates for decades are the questions data centre developers face now.
Tertzakian said the AI version of the story is “far bigger than the oil and gas story.”
Energy’s harms hit specific groups, like Indigenous communities near drilling sites or farmers along pipeline routes.
In his words, AI is “much more broad-based.” Workers worry about losing their jobs. Households worry about electricity prices rising. Whole societies are absorbing a technology that moves faster than the institutions around it.
“This is a way bigger story,” Tertzakian said.
And the backlash has already started.
“I think it’s past the point of no return,” he said. He pointed to polls in the United States showing growing public opposition, and to former Google CEO Eric Schmidt being booed during a University of Arizona commencement speech for his comments about AI.
Opposition is being driven by the same combination of concerns that hit oil and gas, including the energy footprint of data centres, rising electricity prices, and social disruption. Oil and gas took two decades to face the convergence and AI is facing it in weeks.
This is the cycle Little has watched run through the energy sector.
Critics would raise early alarms about safety, regulation, social and economic harm, or environmental damage. The industry, which is logic-based and grounded in STEM, would dismiss them when their evidence and reasoning sounded like “the sky is falling.”
Years later, the criticism would turn out to be right.
Tertzakian named climate change as the example, where oil and gas dismissed it for years before the science caught up and the public organized around it.
“Nobody’s going to believe that,” Little said, describing how warnings about climate, safety, and other issues were treated when they first arrived. “But 10 years on, after the industry’s ignored it for a long period of time, they realize, wow, a lot of people now believe this.”
Sandhu pointed to a Canadian example of what happens when developers fail to bring communities along, citing two large natural gas plants in Ontario that sat in good locations near major demand centres.
Both were cancelled before the 2011 provincial election after local opposition organized against emissions, health risks, and proximity to homes. The cancellations cost the province up to $1.1 billion.
The lesson, Sandhu said, applies directly to data centres. Infrastructure projects get killed when communities are not part of the conversation early enough.
“If [data centre developers] don’t get their head wrapped around the social impacts, the political risks, and the stakeholder engagement that’s going to be needed and necessary, then we might be looking for a little bit of a slower trajectory for adoption.”
Tertzakian offered a blunt warning.
“The person who doesn’t get a cheque has the ability to disrupt the whole project,” he said. “It’s a big learning for AI.”
The model the energy sector already built
Little spent his career in oil and gas during the decades when the industry was forced to reckon with who carried the impacts of energy infrastructure and who got the benefits. The lessons it learned are directly relevant to the AI buildout now.
“Indigenous communities are probably the prime example of this across our nation,” Little said. “They tend to be a rural population, so they’re often out where oil and gas is operating. They’re often having big impacts to their traditional hunting and such. When the industry started, it meant that their trap routes were taken over.”
The model the energy sector eventually built for handling this evolved over decades.
“This has become far more sophisticated, literally over 60 years, to the point now where there’s the First Nations Major Projects Coalition, there’s equity ownership,” Little said. “They’re getting some significant and generational benefits associated with projects.”
That framework is one the data centre industry needs to consider. Little said it’s still in the early stages even in the energy sector, where roughly 10% of the Indigenous population has equity ownership in major projects, and another third is working to secure it.
The data centre industry is making the choice between those two approaches right now.
In March 2026, Kevin O’Leary’s company signed a land sale contract with the Municipal District of Greenview near Grande Prairie, Alta. The project, called Wonder Valley, has been billed by O’Leary as the world’s largest AI data centre campus, with plans for up to 7.5 gigawatts of capacity.
Sturgeon Lake Cree Nation, whose territory borders the site, says it was not consulted on the land sale or on the water permit. The Nation appealed the water licence in April, but the Alberta Environmental Appeals Board ruled it had not shown its members would be directly harmed and threw out the appeal. Sturgeon Lake took the matter to judicial review in Edmonton, where arguments were heard this week.
About 200 kilometres north of Wonder Valley, a different approach is taking shape.
Woodland Cree First Nation holds 51% ownership of Mihta Askiy, a 650-megawatt gas-powered data centre being developed on the Nation’s traditional territory in partnership with Sovereign Digital Infrastructure. Revenue generated will fund education, elder care, housing, and infrastructure in the community, Chief Isaac Laboucan-Avirom told CBC.
The two projects sit in the same province, in the same year, with the same technology, and the operators made different choices about consultation and ownership. Woodland Cree’s approach reflects a framework that has been studied and named for more than a decade.
In a 2011 HBR article called “Creating Shared Value,” Harvard professor Michael Porter and Mark Kramer argued that companies which design societal benefit into the business model from the start build projects that last. The principles behind it have shaped Canadian resource and energy projects over the past decade through Indigenous equity ownership, community benefit agreements, and other models that bring societal benefit into the deal structure from the start.
Some in the energy sector believe the same principles should apply to large infrastructure projects like data centres.
CSV Midstream Solutions, an Alberta natural gas midstream company, recently published an article where its CEO Daniel Clarke argued that, “for good businesses, there’s always capital available, but if you can’t get the community to buy into [a project], then it doesn’t even start.”
Clarke’s point is that the energy sector spent decades learning the cost of skipping that work, and the data centre industry has the chance to use the concept of “creating shared value” as a framework from the start.
Tertzakian made the same point on the Upper Bound stage.
“Get yourself out of this bubble and put yourself in the position of average people that are going to be affected by this,” he said. “Take their perspective and be more sensitive to what they’re thinking, because if you don’t, they have the ability to put up a lot of resistance to what you’re doing.”
Final shots
- The framework for shared value in Canadian infrastructure already exists. The First Nations Major Projects Coalition represents 186 First Nations across the country and advises on equity participation in major projects. The data centre industry can use it or build its own.
- The data centre industry has more capital and more political access than any infrastructure buildout Canada has seen in a generation. What it does with the next two years will set the operating model for the next decade.
- Little’s parting thought to the AI industry: “There is going to be an enormous amount of pressure to go grab the bag of cash. Think about not just the pros associated with it. What are the cons? Because you will live with that for the rest of your journey.”