What If Energy Were Used First to Rebuild Territory?

In Quebec, the question is not new. The agreement reached under Duplessis with the Iron Ore Company on the North Shore remains etched in collective memory as a symbol of resources ceded for very low royalties—often summarized in public debate as the equivalent of roughly one cent per tonne.¹

This memory is not merely historical. It continues to shape how we think about natural resources, energy and regional development. For a long time, much of Quebec’s territory was organized around a simple sequence: extract, transport, export. Ore, timber, hydroelectricity, aluminum and seafood left the regions by rail, through ports, over high-voltage lines or under large supply contracts.

This model was not without benefits. It created jobs, towns, infrastructure and expertise. But it already raised a fundamental question: where is real value created? In the place that supplies the resource, or in the place that transforms, organizes, finances and controls it—and decides how it will be used in the future?

The answer is not to condemn all exports. A resource-rich territory will always sell some of what it produces. The problem arises when exporting becomes not a tool, but an architecture of development. Resources are extracted here, processed elsewhere, and decisions are made elsewhere. The region then becomes a point of departure for other people’s wealth, rather than a place capable of building up its own capacities.

Today’s world requires us to revisit this question in a different way. Electricity is no longer merely an industrial commodity. It is becoming one of the central raw materials of the digital economy, artificial intelligence, automation, research and technological sovereignty. Data centres, especially those related to AI, are becoming strategic infrastructure comparable to the aluminum smelters, railways and ports of the twentieth century.

According to the International Energy Agency, worldwide electricity consumption by data centres could nearly double between 2025 and 2030, rising from approximately 485 TWh to 950 TWh. Consumption by AI-focused data centres would grow even faster, tripling over the same period.² This growth is turning electricity into a critical input for computation, data and the models that will structure an increasing share of the economy.

For Quebec, the question is therefore very concrete: if electricity is one of the raw materials of AI, why continue to think of our regions mainly as sites of energy production or transit? Why not bring part of the digital economy closer to the places where energy, space, cold weather, the need for revitalization and communities seeking a new future already exist?

Consider the example of La Haute-Gaspésie.

Gaspésie–Îles-de-la-Madeleine is one of Quebec’s most rural regions: 96.8% of its population lives in rural areas. It represents just over 1% of Quebec’s population across more than 20,000 km², with fewer than 4.5 inhabitants per square kilometre.³ La Haute-Gaspésie is in a particularly fragile position: data from the Institut de la statistique du Québec ranked it 104th out of 104 regional county municipalities on the economic vitality index in every edition from 2014 to 2022.⁴

In a region like this, the challenge is not simply to create jobs. It is to sustain a year-round economy, support public services, attract or retain families, enable young people to stay, and create something beyond an alternation between tourist season, public-sector jobs, occasional resource extraction and support programs.

Yet the Gaspé Peninsula is also expected to play an important role in Quebec’s wind-power development. Cap-Chat is already part of that history: the first phase of the Énergie Le Nordais project entered service there in 1998, with 76 turbines rated at 750 kW each.⁵ More recently, Hydro-Québec, the Alliance de l’énergie de l’Est and the Mi’gmaq of Gespe’gewa’gi announced a wind-power development partnership for the Gaspé Peninsula and eastern Bas-Saint-Laurent. This territory could host as much as 6,000 MW in the years ahead.⁶

Faced with this potential, the challenge should not be merely to produce more electricity and carry it to major urban centres or external markets. It should also be to ask what activities could be established close to that energy.

An artificial-intelligence data centre in Cap-Chat, L’Anse-Pleureuse or elsewhere in La Haute-Gaspésie would not automatically be a good idea. It would require robust fibre-optic connections, reliable power, redundancy, a cooling strategy, cybersecurity, specialized technicians, clear governance and genuine social acceptance. It would also be necessary to avoid the trap of an energy-intensive facility that created few direct jobs and left most of the value in the hands of outside actors.

But a data centre should not be considered in isolation. On its own, it risks becoming an electrical black box. Integrated into its surroundings, it could become part of a productive ecosystem.

A data centre produces a great deal of heat. That heat is often treated as waste. Yet the Government of Quebec already presents waste-heat recovery as a promising way to heat buildings or greenhouses. It explicitly gives the example of greenhouse vegetables grown using heat emitted by a nearby factory or data-processing centre.⁷ In Lévis, a project involving QScale and Les Fraises de l’Île d’Orléans plans to use heat from a data centre to warm a 12-hectare greenhouse complex, with a projected reduction of more than 23,000 tonnes of CO₂ equivalent per year.⁸

This approach could be adapted to other territories. In La Haute-Gaspésie, one could imagine not a single project, but an interconnected hub: a data centre, heat recovery, year-round heated greenhouses, technical training, remote work, applied research, energy maintenance, digital services, storage, complementary solar generation and integration with the regional renewable grid.

The point is not to present this combination as a magic solution. It raises difficult questions: cost, the jobs actually created, infrastructure ownership, workforce availability, grid capacity, cybersecurity, data governance, environmental impact, and the roles of municipalities and First Nations. Some computing workloads can be located far from major centres; others require greater proximity, low latency or metropolitan redundancy. But this hypothesis has one strategic virtue: it shifts the debate.

Instead of asking only how to export energy, it asks how energy can be used to create local capacity.

This distinction is essential. Exporting a resource can generate revenue. Building an ecosystem around it can generate optionality: skills, businesses, services, food self-sufficiency, digital infrastructure, research opportunities, career paths and reasons to stay.

This is where the idea of bifurcation becomes useful. A territory does not develop simply through the accumulation of projects. It develops through the opening and closing of possible paths. A transmission line, wind farm, data centre, greenhouse, training school or regional interconnection is not merely a piece of equipment. Each is a choice that shapes the territory’s future form.

Over the longer term, we should even be willing to examine ideas that seem unusual today. A submarine interconnection between the Gaspé Peninsula and the North Shore—for example, between Sainte-Anne-des-Monts and Sept-Îles, or along a route yet to be determined—would obviously require rigorous assessment: cost, depth, energy losses, grid connections, environmental impacts, overland alternatives, resilience and its actual role within the system. But simply asking the question has value. It forces us to think of Gaspesian wind power, North Shore hydroelectricity, regional needs and industrial uses as one system rather than as separate files.

Quebec already has a vast electrical grid and exchanges electricity with neighbouring markets. In 2022, its electricity exports reached 36 TWh and its imports 34 TWh.⁹ But the province’s energy picture is more complex than it appears. That same year, only 42% of the energy available in Quebec came from local production, while 58% was imported.¹⁰

In other words, Quebec is highly electrified, but it has not yet achieved full energy sovereignty. It exports electricity while still importing a large share of its total energy in the form of oil and gas. In this context, every new renewable megawatt-hour should be treated as a scarce strategic resource. Its value should be assessed not only by the price it can command on export markets, but also by the dependencies it reduces and the capacities it creates.

This line of thought directly parallels the discussion about raw materials. The Government of Quebec says it wants to create more local value from critical and strategic minerals, through exploration, production, processing and recycling. Its 2025–2031 strategy also aims to foster processing in Quebec and the manufacture of value-added products.¹¹ This language is relevant. But it should be extended to energy itself.

Electricity should not be regarded solely as an exportable raw material. It should be treated as a platform for development.

In the traditional industrial world, processing locally meant going beyond extracting ore to concentrating, smelting and refining it, then integrating it into production chains. In the digital and climate-conscious world ahead, processing locally can also mean converting electricity into computing capacity, converting heat from computation into food production, and converting digital infrastructure into training, research, remote work, specialized businesses and permanent regional services.

This does not mean everything should be done everywhere. Not every region can host the same infrastructure. Not every data centre is desirable. Local processing is not automatically profitable, environmentally sound or socially acceptable. The criteria must be demanding.

A good project should produce more than temporary construction jobs. It should create transferable skills, permanent technical employment, local purchasing, digital infrastructure and regional maintenance capacity.

It should recover heat instead of wasting it. In a cold region, a data centre without heat recovery should be considered an incomplete project.

It should strengthen local autonomy, particularly in food, services, remote work, training and resilience.

It should be compatible with the communities that host it. Municipalities, regional county municipalities, First Nations, citizens and workers should take part in defining the benefits, risks and conditions of acceptability.

Finally, it should retain a meaningful share of the value in Quebec. If public electricity is used mainly to power infrastructure controlled by foreign giants whose profits, data, algorithms and strategic decisions reside elsewhere, Quebec risks repeating an old extractive mistake in digital form.

Hydro-Québec already appears to recognize that data centres are not an electrical load like any other. In February 2026, it proposed to the Régie de l’énergie a new rate for large data centres over 5 MW, corresponding to an average unit price of approximately 13¢/kWh. Its application, planned for the second half of 2026, remains subject to the Régie’s approval. Projects over 5 MW also remain subject to a selection process intended to steer Quebec’s energy toward its most promising uses.¹² This is progress. But the criterion should not be merely financial. It should also be territorial.

A data centre in Montréal, Québec City or a major industrial zone may make sense for certain functions. But we should not assume that metropolitan concentration is always optimal. Where energy, land, cold weather, recoverable heat and revitalization needs converge, other configurations become possible.

Remote work, distance learning, digital laboratories, distributed research and new forms of work organization make this question more realistic than it would have been twenty years ago. Not everything can be relocated. But a significant portion of the knowledge economy no longer needs to be physically concentrated in major centres. If work can travel over fibre, if computation can be distributed, if training can be hybrid and if heat can be recovered, then Quebec’s economic geography can be redrawn.

The question, then, is not simply: should we export our resources or process them here?

It becomes: what local combinations of resources, energy, infrastructure, expertise and social needs can create new collective capacities?

In a region facing economic decline, a megawatt-hour has more than a market value. It can become a technical job, a heated greenhouse, a service kept open, a family that stays, a training program, a business that moves in, a young person who does not have to leave, a researcher able to work remotely, or a village that finds a path forward again.

This is not a guarantee. It is a possibility. But that is precisely the role of intelligent economic policy: not to sell possibilities to the highest bidder before examining what they might open up.

Quebec has long feared a shortage of capital. At times, it has been too quick to cede its resources in exchange for promises of development. Today, the risk is different. It is no longer only a matter of selling ore too cheaply. It is a matter of selling electricity, data, territory and the future without building lasting capacity around them.

Wealth lies not only in ore, wind or the megawatt-hour. It lies in the bifurcations they make possible.

An exported resource often closes an option.

A resource integrated into a local ecosystem can open several.

References