Opportunity Knocks: How Should Saskatchewan Answer? A Critical Assessment of the Regina-Area AI Data Centre

Download the JSGS Policy Paper

Download the Disussion Questions

Bell Canada’s project to build an AI data centre campus south of Regina promises Saskatchewan jobs, research partnerships, and a contribution to Canada’s nation-building objectives. But the project’s demands on the province’s electricity system deserve far more scrutiny than the project has received to date. Decision-makers should insist on answers to basic questions that the project’s supporters have not raised, and the proponent has not yet addressed. More importantly, we should be asking what opportunities this project could create for Saskatchewan.

In early February 2026, reporting on a land use rezoning application revealed that a numbered company linked to Bell Canada’s “AI Fabric” initiative proposed to construct a multi-phase AI data centre campus on approximately 160 acres of agricultural land in the Rural Municipality of Sherwood, just south of Regina.¹ The application has since been approved. This project, when fully built, will see roughly 500,000 square feet of data centre buildings and 50,000 square feet of office space, served by a dedicated SaskPower substation. The first phase involves a 90,000 square foot building with construction beginning in 2026.²

By mid-March, additional details had been clarified: Saskatchewan Premier Scott Moe formally announced the project on March 16, describing it as a $1.7-billion initial investment by Bell with total costs including chip procurement reaching approximately $12 billion.³ At the same event Bell CEO Mirko Bibic called it the company’s “largest ever investment in this province” and emphasized that adequate power supply was crucial to the decision to build in Saskatchewan. The announcement confirmed that the facility will require 300 megawatts of baseload power—equivalent to the electricity demand of roughly 300,000 Saskatchewan households. Bell's group president John Watson also said the centre will use “a water-smart and energy-efficient design, including a closed-loop cooling system that does not draw water from municipal water sources”, a crucial component in such a facility. The first data hall is expected to come online in the first half of 2027, with long-term tenancy agreements already contracted with Cerebras and CoreWeave, two U.S.-based AI compute providers serving global customers.⁴

Also present at the announcement were Crown Investments Corp. Minister Jeremy Harrison and George Gordon First Nation Chief Shawn Longman. The project has been received enthusiastically by local officials and institutional partners. Regina’s mayor, Chad Bachynski, called it an “exciting opportunity.”⁵ The University of Regina’s vice-president research has provided a letter of support,⁶ and Saskatchewan Polytechnic is negotiating a memorandum of understanding with Bell around applied research and workforce development.⁷ The project is framed as essential Canadian digital infrastructure that would support data sovereignty, public-sector computing, and regional innovation.

Through this policy paper, we wish to help illuminate this discussion—not to undermine the project, but simply with the hope that all the potential opportunities and issues attached to such an initiative are considered. While acknowledging the appeal of hosting advanced AI infrastructure in Saskatchewan and the project’s potential benefits to Canada and the province, we argue that a number of opportunities—related to how the project is powered and what it gives back to the province—have not been fully explored. Several design choices and partnership structures could make this project even more beneficial to Saskatchewan than the announcement suggests. As the project is still in its early stages, and the proponent has not yet provided the detailed technical information needed to make a fully-informed assessment, decision-makers should require that information, and signal design criteria, as the project moves forward. More broadly, our aim is to start a conversation about what Sherwood, Regina, and the Government of Saskatchewan should expect in return when a major national corporation asks to consume scarce regional and provincial resources.

Bell Canada and its supporters have emphasized the job-creation potential of the campus, citing hundreds of construction jobs and subsequent technical, operations, and support roles.⁸ Minister Harrison has claimed roughly 1,600 jobs in total: 800+ for construction, 80 permanent on-site, and 750+ in spinoffs and directly associated roles.⁹

Nonetheless, a growing body of evidence from data centre projects elsewhere suggests that permanent operational employment will be modest. The near-term construction employment impacts are more straightforward. As for the ongoing operational employment impacts, modern data centres are highly automated facilities, and the 80 permanent on-site positions are consistent with what we would expect from the evidence: a 2025 employment forecast, from a rigorous field-based analysis of U.S. data centre staffing, found that operations employment runs between 0.15 and 0.35 full-time equivalents per megawatt, and that the most automated hyperscale campuses over 100 MW can operate with as few as 20 to 30 permanent staff.¹⁰ At 300 MW, this benchmark is broadly consistent with the announced figure of 80 permanent on-site staff.

The more uncertain claim is the 750-plus spinoff and associated jobs, which have not been defined or substantiated. Even taken at face value, this level of employment is modest relative to the scale of public resources the project would consume. If employment is a primary objective, other forms of economic development are considerably more labour-intensive per dollar invested.

Two recent developments are worth noting on the construction side. On May 4, 2026, Bell announced its early-phase contractor team: most of the firms engaged are Regina- or Saskatoon-based (Hipperson Construction, Soletanche Bachy Canada, Ardel Steel, Amrize, WaterMark Consulting, Maxie’s Excavating, Red Pelican), with pre-engineered building components supplied from Behlen Industries of Brandon, Manitoba.¹¹ Earlier in April, Bell and George Gordon Developments Ltd.—the economic development arm of George Gordon First Nation—convened the first meeting of a joint Workforce Steering Committee focused on Indigenous procurement participation and workforce development.¹²

The letters of support from the University of Regina⁶ and Saskatchewan Polytechnic⁷ point to opportunities for applied research, student training, and commercialization. These are real possibilities, but they remain speculative. The memoranda of understanding (MoUs) are described as “under negotiation,” with no public details on scope, funding, or governance. The project description mentioned “sustainable greenhouses” using waste heat from the data centre—an intriguing concept, but one that remains wholly aspirational without further details.

One question to ask is what, concretely, University of Regina or Saskatchewan Polytech researchers and students would gain from proximity to a commercial data centre that they could not obtain through cloud computing agreements or partnerships with existing national research computing infrastructure such as the Digital Research Alliance of Canada? Physical co-location with a data centre does not, by itself, provide researchers with access to compute capacity. That said, there are concrete possibilities worth exploring. The University of Regina's existing occupants on Research Drive and at Innovation Place, and the programs in computer science, engineering, and business, could benefit from structured partnerships. Those partnerships should be specified in those MoUs. The question is not whether synergies exist in the abstract, but whether the proponent and the institutions have committed to making them real.

For example, in Utah the University of Utah and Elemental Nuclear are demonstrating a proof-of-concept nuclear microreactor system for powering AI data centres. The initiative aims to advance next generation nuclear microreactors and involves collaboration amongst twelve universities across the United States.¹³ With SaskPower committed to the development of nuclear energy, it raises the question whether the Bell initiative offers potential for such microreactors as part of the project.

Proponents argue that the campus could attract additional technology firms and spinoff ventures to the Regina region, serving as a catalyst for a broader AI ecosystem. And it is recognized that the measurement of productivity spillovers lag behind capital investments.¹⁴ This is plausible in theory but difficult to operationalize. We appreciate the ‘chicken-and-egg’ nature of this problem, but the presence of a facility where data is stored and processed does not inherently create a technology cluster; building this ecosystem requires capital, institutional support, a critical mass of firms, and a long-term investment in talent development.

The project raises a question that has received little attention as far as we can tell: why this site? The campus would consume 160 acres of agricultural land and require new infrastructure connections. Yet Regina already has underutilized institutional and commercial space that might be suitable for at least the initial phases of a data centre operation. The former ISM/IBM building on Research Drive, adjacent to the University of Regina campus, seems largely vacant. Co-locating data centre capacity in existing institutional space would offer proximity to university researchers without consuming agricultural land or requiring new infrastructure from scratch.

We don't imply that this is a simple solution and there are, of course, practical considerations. Retrofitting an existing building for data centre use involves significant HVAC, power distribution, and structural upgrades. The floor loading, ceiling heights, and cooling capacity of a general-purpose office or research building may not suit high-density computing without substantial modification. A purpose-built facility on a greenfield site may ultimately prove more practical and cost-effective. But these are engineering questions that deserve analysis, not assumptions. The proponent and supporting institutions should be asked to explain why existing urban or institutional sites were not considered, or what technical factors ruled them out, even though 160 acres of agricultural land have now been rezoned.

A significant concern is the project’s impact on Saskatchewan’s electricity system. The March 16 announcement confirmed that the facility will require 300 megawatts of continuous baseload power. To put that in perspective: SaskPower's total installed generating capacity is approximately 5,900 megawatts, with average system demand in the range of 2,600 to 3,500 megawatts depending on season.¹⁵ A single 300 MW load is comparable to the BHP Jansen potash mine—one of the largest industrial developments in the province's history. SaskPower is already under significant pressure. The utility is seeking consecutive 3.9 percent rate increases in 2026 and 2027 to finance record capital investments in new generation and infrastructure replacement (e.g., a 370-megawatt Aspen natural gas plant is being built near Lanigan).¹⁶

The March 16 announcement did shed light on how SaskPower proposes to meet this demand. Minister Harrison indicated that the project hinges on the availability of “dispatchable baseload power”—specifically, the refurbishment of Boundary Dam coal-fired Units 4 and 5, which would add approximately 278 megawatts of generating capacity back to the grid. Harrison was quoted as saying that without available coal power, the announcement would not have happened. In addition, Bell is building 300 megawatts of natural gas backup generation on the data centre site itself, with SaskEnergy supplying the gas.

The April 2 consultant’s report adds one further element to the picture. The site plan describes a “dedicated energy generation zone” that will include not only the SaskPower-linked substation and a “future natural gas plant,” but also a “renewable energy system.” The renewable component is new public information.¹⁷ Its scale, technology (solar, wind, geothermal, or other), purpose (on-site supply, back-up, demonstration, marketing), and relationship to the facility’s 300 MW baseload demand have not been specified. A renewable element of any meaningful size would be a welcome design choice; an unspecified element of unknown size is, for now, a category on a site plan.

AI data centres are generally placing significant demands on Canada’s electricity grid. Modelling has shown that adding data centre load to Canadian grids can increase global greenhouse gas emissions even when the host grid is relatively low-carbon.¹⁸ Saskatchewan’s grid is considerably more carbon-intensive than the national average, which would amplify these effects.¹⁹

This raises questions that go beyond the data centre itself. The province is effectively recommitting to coal-fired generation—which it had been winding down—to enable a single commercial project. And the on-site gas plant means Saskatchewan will be importing natural gas from Alberta (which supplies roughly 90 percent of the province's natural gas) to power what is being described as a forward-looking technology investment. These are not arguments against the project, but they are choices with long-term implications that deserve public scrutiny.

There is also a risk question that has not been publicly addressed: who bears the costs if climate policy tightens? If future federal or provincial regulations require a transition away from unabated natural gas generation, the costs of converting or decommissioning those assets will fall on someone—either Bell, SaskPower, or Saskatchewan ratepayers. This is an opportunity for the province to negotiate risk-sharing arrangements that protect the public interest. Models for shared ownership and risk diversification exist in Saskatchewan’s energy sector—the Meridian co-generation facility is one precedent—and could be adapted here.

Elsewhere, Google and Microsoft are investing directly in dedicated nuclear generation to power their data centre operations, and other major technology firms have announced similar commitments.²⁰ While SaskPower has signalled interest in Small Modular Reactors and potentially larger nuclear capacity, requiring Bell to participate in financing or co-developing clean energy sources—rather than simply drawing from the existing grid—would align this project with the direction the industry is moving globally. Bell’s March 16 announcement notes that “advanced discussions are underway” regarding a district energy system that would deliver the facility’s waste heat to nearby university campuses and to a development project led by George Gordon Developments Ltd. directly north of the site.²¹ This is a genuinely encouraging direction. The question is whether “advanced discussions” will result in a binding scheme with specified scale, cost-sharing, infrastructure ownership, and a performance guarantee—or whether, like the early talk of “sustainable greenhouses,” it will recede into aspiration once the construction permits are in hand.

The picture that emerges is of a project that, as currently structured, would be powered substantially by fossil fuels—coal from refurbished Boundary Dam units and natural gas from both the grid and an on-site generating plant. This is not inherently disqualifying, but it does sit uncomfortably alongside the proponent's language about sustainability. And it raises a straightforward question: could this project be structured differently, so that it contributes to Saskatchewan's energy transition rather than reinforcing the status quo?

Data centres can require substantial volumes of water for cooling.²² At the scale of the Regina campus, annual water consumption for such a facility could be significant, depending critically on the cooling technology employed.

On this point, the March 16 announcement brought a significant and welcome disclosure: the facility will use a closed-loop cooling system.⁴ This is the first public commitment on cooling technology and, if implemented as described, would dramatically reduce the facility's water consumption compared to conventional evaporative cooling.

Subsequent disclosures have brought further clarification. A consultant’s report posted by the RM on April 2 describes the design more specifically as “air-cooled, closed-loop chilling technology” requiring only a one-time fill, and Bell’s representatives have stated that the facility “does not draw water from municipal water sources” on an ongoing basis. The facility will, however, be connected to the City of Regina’s municipal water grid for the initial fill and any subsequent top-ups.¹⁷ Several operational questions remain. The volume of the one-time fill for a campus expected to reach 500,000 square feet of data centre buildings has not been disclosed. Neither has the expected frequency and volume of top-ups, nor the contingency plans if a closed-loop system requires unscheduled refilling or partial replacement during maintenance. These are not arguments against the design; they are the operational details that would distinguish a binding closed-loop commitment from a press-release.

The commitment to closed-loop cooling is encouraging, but it needs to be formalized: a press statement is not the same as an enforceable condition. There is precedent in Regina for requiring closed-loop water practices: EVRAZ Regina converted its steel facility to zero liquid effluent, with 100 percent of water treated and recycled on-site. Tracing the origins of that initiative could provide useful guidance for the data centre. A similar standard—expressed as a binding condition of the development permit—would give the public confidence that the water impact will in fact be managed. An industry benchmark of 0.4 litres of water per kilowatt-hour in water-stressed areas, proposed by the DIGITALEUROPE industry association, offers a measurable target against which the facility's performance could be assessed.²³

The RM of Sherwood approved the rezoning of 160 acres from agricultural to light industrial, and subsequently a development agreement with Bell, on the basis of incomplete technical information. The March 16 announcement answered some questions—confirming the 300 MW power requirement, closed-loop cooling, and approximate employment figures—but several important items remain undisclosed:

• the expected electrical load at each phase;
• the greenhouse gas emissions profile of the facility, including indirect emissions from electricity generation and on-site natural gas use for back-up power;
• firm employment projections with timelines (for both the construction phases and long-term operational needs), including a substantiation of the claimed 750-plus spinoff and associated jobs;
• the terms, funding, and governance of the planned research partnerships with post-secondary institutions;
• an explanation of why this greenfield site was selected over existing underutilized space; and
• the risk-allocation framework for energy supply—specifically, who bears the cost of any future transition away from fossil-fuel generation, and what role Bell would play in financing or co-developing clean energy sources.

At the time of the rezoning vote, the responses from regulators and the RM appeared to take a “don’t worry, it’s just rezoning” posture. The RM has indicated that a “fully approved detailed design” will be required before construction permits are issued, and the City of Regina, Water Security Agency, SaskTel, and the Ministry of Highways have raised no objections to the rezoning change itself.¹

Our concern is that a rezoning decision creates expectations, momentum, and sunk costs that make subsequent questions more uncomfortable. We argue that the time to ask hard questions was before the rezoning was approved. Nonetheless, several leverage points remain, including forthcoming construction and operating permits and the binding terms of partnership agreements still to be finalized.

Between mid-March and late April, the local approval process came under significant strain. Two days after the March 16 announcement, four of the seven RM of Sherwood councillors abruptly resigned, including the reeve and deputy reeve. One outgoing councillor wrote that “due to the direction the RM council is going I cannot have my name affiliated with the RM anymore”; another, who had served the RM since 1991, cited an RM email failure that left him without access to materials for the closed-door March 16 special meeting. The resignations cost the council its quorum and halted regular business. On April 10, the Government of Saskatchewan appointed an interim reeve (Ray Orb, former president of the Saskatchewan Association of Rural Municipalities) and three other councillors, none of them residents of the RM of Sherwood, to serve until a November byelection. On April 20, the reconstituted council—three elected members and four provincial appointees—unanimously approved a development agreement with Bell. The meeting was held behind locked doors due to capacity issues, while hundreds of protestors gathered outside the RM office. Earthworks began the following day.²⁴

While the governance situation unfolded, a group of seven families with property immediately adjacent to the site had been seeking binding conditions on noise, lighting, drainage, stormwater, groundwater, and road management. They submitted four formal requests to the RM between January 29 and March 4, only the first of which was posted to the public record. On February 25 they filed a formal complaint with the Saskatchewan Ombudsman.²⁵

None of this is, by itself, an argument against the project. But it is a reminder that the question of how a major investment is approved is not only a question about the investment itself; it is also a question about local democratic processes, and about who is in the room when the decision is made.

This paper does not argue that AI data centres have no place in Saskatchewan. National data sovereignty is a legitimate concern,²⁶ and Saskatchewan’s cold climate, available land, and existing transmission infrastructure give it genuine advantages as a data centre location. However, having genuine advantages as a host jurisdiction also means having leverage—and that leverage should be used to ensure the project delivers more than it currently promises.

We recommend several steps that could help ensure that the public interest is protected.

• The RM of Sherwood should require detailed resource impact assessments before granting further approvals. This should include disclosure of the facility’s expected electricity demand at each phase, the cooling technology and anticipated water consumption, and a greenhouse gas emissions profile (including the emissions from the on-site natural gas plant) that accounts for the carbon intensity of SaskPower’s generation mix. These questions should be addressed by the RM after the November byelection seats a fully-elected council.
• Explore a Bell Canada / SaskPower partnership to develop a clean energy source to power the data centre. The current arrangement—in which the facility would draw from refurbished coal units and on-site natural gas generation—locks in fossil-fuel dependence for a project with a multi-decade lifespan. SaskPower and the provincial government should assess the system-wide implications of adding a 300 MW industrial load to a grid that is already under strain, including the impact on rates for existing customers and the implications for SaskPower's capital investment plan. The assessment should also address who bears the long-term risk if climate policy requires a transition away from unabated fossil-fuel generation. Risk-sharing arrangements—in which Bell co-invests in clean generation or accepts contractual responsibility for future transition costs—would protect Saskatchewan ratepayers and align the project with the direction the global data centre industry is moving.
• Make Bell's commitment to air-cooled, closed-loop chilling technology binding as a condition of the development permit. The facility’s water performance should be benchmarked against international best practice.²³
• The research partnerships should be specified and formalized before they are cited as benefits of the project. If the University of Regina and Saskatchewan Polytechnic are to receive meaningful research access, training opportunities, or compute capacity, those commitments should be made binding and public, not left to future negotiation. Concrete possibilities—such as integrating the data centre's waste heat into the University of Regina's district energy system or providing structured research access for programs in computer science, engineering, and business—should be explored and specified. Bell’s reference to “advanced discussions” on the district energy element should be converted into a public memorandum of understanding with specific scale, cost-sharing, and performance terms before it is treated as a project benefit.
• The proponent should be required to demonstrate that the project’s benefits to the region justify its consumption of scarce public resources. Electricity and water in Saskatchewan are not unlimited; they are limited by their cost, infrastructure capacity, and natural supply. A project that consumes significant quantities of both should be held to a standard of public benefit that goes beyond corporate investment and vague promises of innovation. A useful framework has been proposed by CUPE in its 2025 policy brief on data centre governance in Canada, which calls for no public subsidies without binding benefit agreements, labour standards, and environmental protections.²⁷ If the data centre will serve primarily as commercial infrastructure for Bell's national AI platform—rather than as a shared resource for Saskatchewan's public sector, researchers, and communities—the case for accommodating its resource demands at the expense of other users is considerably weaker. As for the governance challenges that have been witnessed, future major industrial decisions in rural municipalities should require minimum public-consultation thresholds.

Saskatchewan has important motives for wanting AI infrastructure investment in the province. The Bell AI Fabric project presents worthy objectives—data sovereignty, digital innovation, and economic diversification—and the March 16 announcement represents a significant step toward transparency on the project’s scale and design. But scale and ambition are not the same as good policy. The project asks Saskatchewan to provide one of its scarcest resources—electricity from an overburdened, fossil-fuel-dependent grid—in exchange for benefits that have not been fully specified or guaranteed, and opportunities that have not been fully explored.

The enthusiasm of local officials and institutional partners is understandable. What is premature is the assumption that the project as currently stated is the best version attainable. As this project proceeds, decision-makers should ask not only what the project will cost the province, but what it could give back—in clean energy partnerships, in binding research commitments, in risk-sharing arrangements that protect ratepayers, and in enforceable environmental standards. The RM of Sherwood, as the land-use authority, has both the opportunity and the responsibility to insist on those answers. It is important not to settle for less than Saskatchewan deserves.

¹ RM of Sherwood, Notice and Agenda Package of Special Meeting, February 9, 2026. https://rmofsherwood.ca/app/uploads/2026/02/Notice-Agenda-Package-of-Special-Meeting-February.9.2026.pdf

² Data Center Dynamics, “Bell files to develop data center campus in Saskatchewan, Canada,” February 2026. https://www.datacenterdynamics.com/en/news/bell-files-to-develop-data-center-campus-in-saskatchewan-canada/; Data Center Dynamics, “150MW data center proposed in Saskatchewan, Canada.” February 2026. https://www.datacenterdynamics.com/en/news/150mw-data-center-proposed-in-saskatchewan-canada/

³ Premier Scott Moe touts 'Canada's largest AI data centre' as economic boom for Sask. CBC News https://www.cbc.ca/player/play/video/9.7130786

⁴ Silverthorn, C., & Quon, A. (2026, March 16). Bell Canada to begin construction on Canada’s largest AI data centre near Regina this spring: Telecom giant says the centre will generate $12 billion for Saskatchewan’s economy, 80 permanent jobs. CBC News. https://www.cbc.ca/news/canada/saskatchewan/ai-data-rm-of-sherwood-9.7130417

⁵ Yahoo News, “Regina’s mayor expresses excitement over AI data centre proposal,” February 2026. https://ca.news.yahoo.com/reginas-mayor-expresses-excitement-over-203642184.html

⁶ RM of Sherwood, Notice and Agenda Package of Special Meeting, February 9, 2026, pp. 34-35.

⁷ RM of Sherwood, Notice and Agenda Package of Special Meeting, February 9, 2026, pp. 33.

⁸ CKOM, “Jobs, research benefits touted as plans for data centre near Regina take shape,” February 11, 2026. https://www.ckom.com/2026/02/11/jobs-research-benefits-touted-as-plans-for-data-centre-near-regina-take-shape/

⁹ Zinchuk, B. (2026, March 16). Data centre hinges on dispatchable baseload power, including coal refurbishment. Pipeline Online. https://pipelineonline.ca/breaking-data-centre-hinges-on-dispatchable-baseload-power-including-coal-refurbishment.

¹⁰ Hamm Institute, “Data Center Employment Forecast Analysis,” November 2025. https://hamminstitute.org/site-files/documents/data_center_workforce.pdf

¹¹ Bell Canada. (2026, May 4). Bell engages Saskatchewan contractors for early phases of 300MW AI data centre build [Press release]. https://www.bce.ca/news-and-media/newsroom?article=bell-engages-saskatchewan-contractors-for-early-phases-of-300mw-ai-data-centre-build

¹² 650 CKOM, "Several Saskatchewan contractors hired to help build Bell data centre near Regina," May 4, 2026. https://www.ckom.com/2026/05/04/several-saskatchewan-contractors-hired-to-help-build-bell-data-centre-near-regina/; George Gordon First Nation. March 17, 2026. “Bell Canada announces massive AI data centre near Regina in $1.7 billion investment.” https://ggfn.ca/bell-canada-announces-massive-ai-data-centre-near-regina-in-1-7-billion-investment/

¹³ University of Utah, John and Marcia Price College of Engineering. April 28, 2026. Available at: https://www.price.utah.edu/2026/04/28/elemental-nuclear-and-the-university-of-utah-will-demonstrate-proof-of-concept-nuclear-microreactor-system-for-powering-ai-data-centers

¹⁴ Estevao, Marcello. AI Can Lift Global Growth. Point of View. March 2026. Finance and Development. Available at: https://www.imf.org/en/publications/fandd/issues/2026/03/point-of-view-ai-can-lift-global-growth-marcello-estevao

¹⁵ SaskPower, “2024-25 Annual Report,” 2025. https://www.saskpower.com/-/media/saskpower/about-us/reports/report-annualreport-2024-25.pdf

¹⁶ SaskPower, “2026 and 2027 Rate Increases,” 2026. https://www.saskpower.com/accounts/power-rates/2026-and-2027-rate-increases

¹⁷ Spray, H. (2026, April 7). Neighbour wants more than promises on proposed Sask. AI data centre. CBC News. https://www.cbc.ca/news/canada/saskatchewan/ai-data-centre-rm-sherwood-regina-report-9.7155525

¹⁸ Dandres, T. et al. (2016). “Consequences of Future Data Center Deployment in Canada on Electricity Generation and Environmental Impacts: A 2015–2030 Prospective Study.” Journal of Industrial Ecology. https://onlinelibrary.wiley.com/doi/abs/10.1111/jiec.12515

¹⁹ Saskatchewan’s electricity generation mix remains heavily fossil-fuel dependent: natural gas provides roughly 39 percent of the province’s capacity, coal 24 percent, hydro 21 percent, wind 11 percent, and a negligible amount of solar; see Canada Energy Regulator, “Saskatchewan energy profile,” 2026 https://www.cer-rec.gc.ca/en/data-analysis/energy-markets/provincial-territorial-energy-profiles/provincial-territorial-energy-profiles-saskatchewan.html. See also Peter Nicholson, “Implications of the Global Solar Power Revolution for Canada's Energy Future.” JSGS Policy Paper, Johnson Shoyama Graduate School of Public Policy, February 2026. https://www.schoolofpublicpolicy.sk.ca/research-ideas/publications-and-policy-insight/policy-brief/policy-paper-solar-power-revolution.php

²⁰ Rapier, R. (2026, February 19). Why Microsoft and Amazon are turning to nuclear power for AI. Forbes.

https://www.forbes.com/sites/rrapier/2026/02/19/why-microsoft-and-amazon-are-turning-to-nuclear-power-for-ai/

²¹ Bell Canada press release, “Bell AI Fabric Expands National Network with 300 MW Data Centre in Saskatchewan,” March 16, 2026, https://www.bce.ca/news-and-media/newsroom?article=bell-ai-fabric-expands-national-network-with-300-mw-data-centre-in-saskatchewan.

²² UK Department for Science, Innovation and Technology / DEFRA (2025). “Water Use in AI and Data Centres.” https://assets.publishing.service.gov.uk/media/688cb407dc6688ed50878367/Water_use_in_data_centre_and_AI_report.pdf

²³ DIGITALEUROPE (2025). “Enhancing Water Resilience in the Data Centre Industry.” https://cdn.digitaleurope.org/uploads/2025/07/DIGITALEUROPE_Enhancing-water-resilience-in-the-data-centre-industry_July-2025.pdf

²⁴ CKOM, “Council agenda sheds light on RM of Sherwood resignations,” April 20, 2026, https://www.ckom.com/2026/04/20/council-agenda-sheds-light-on-rm-of-sherwood-resignations/; 620 CKRM, “RM of Sherwood names interim council after mass resignations,” April 14, 2026, https://www.620ckrm.com/2026/04/14/rm-of-sherwood-names-interim-council-after-mass-resignations/; CBC News, “RM of Sherwood council approves Bell Canada AI data centre near Regina,” April 21, 2026, https://www.cbc.ca/news/canada/saskatchewan/rm-of-sherwood-sask-council-vote-ai-data-centre-9.7169961.

²⁵ Global News, “Bell holds closed-door meeting with Saskatchewan landowners over AI data centre,” April 14, 2026, https://globalnews.ca/news/11801671/bell-sherwood-rm-ai/

²⁶ Government of Canada, Treasury Board of Canada Secretariat / Shared Services Canada (2018). “White Paper: Data Sovereignty and Public Cloud.” https://www.canada.ca/en/government/system/digital-government/digital-government-innovations/cloud-services/digital-sovereignty/gc-white-paper-data-sovereignty-public-cloud.html

²⁷ Canadian Union of Public Employees. (2025, November 13). Governments must protect the public interest in data centre boom. https://cupe.ca/governments-must-protect-public-interest-data-centre-boom

Justin Longo is an Associate Professor in the Johnson-Shoyama Graduate School of Public Policy at the University of Regina where he directs the Digital Governance Lab. He has a PhD in public policy and public administration from the University of Victoria (2013) and undertook postdoctoral work in open governance at Arizona State University and the GovLab@NYU, joining the University of Regina as the Cisco Research Chair in Digital Governance. His current research focuses on the social, organizational, and political implications of advancing technology.

Margot Hurlbert is a Professor and Canada Research Chair, Tier 1, Climate Change, Energy, and Sustainability Policy of the Johnson Shoyama Graduate School of Public Policy, University of Regina. She explores the gap between what is needed to address climate change and current policy and behaviour.  Margot’s scholarship concerns climate change, energy, Indigenous peoples, sustainability policy, and achieving net zero emissions in the forms of policy briefs, sharing circles, school events, and water walks. She was coordinating lead author of a chapter of the Special Report of the Intergovernmental Panel (“IPCC”) on ‘Climate Change on Land’ and is a Senior Research Fellow of the Earth Systems Governance Project.

 

Please email your feedback or suggestions on the Policy Brief to the editor: dale.eisler@uregina.ca