Meta's groundbreaking approach to powering its massive Canadian AI campus reveals a fundamental shift in how hyperscalers are tackling the electricity crisis threatening data center expansion globally. Rather than relying on existing power grids, Meta is constructing a dedicated 932 megawatt (MW) natural gas generating facility alongside its C$13 billion Alberta data center, illustrating an emerging strategy where companies essentially bring their own power infrastructure to avoid grid constraints.

Why Are Data Centers Running Out of Power?

The explosive growth of artificial intelligence (AI) is creating an unprecedented strain on electrical grids worldwide. Global data center power demand is projected to triple by 2036, with CO2 emissions doubling over the same period. This surge has created a critical bottleneck: electricity availability has become the primary limiting factor for new data center construction, with grid interconnection queues lasting years in several regions.

The problem is especially acute because AI data centers consume vastly more power than traditional cloud facilities. Graphics processing units (GPUs) used for AI training concentrate enormous amounts of power in each rack, creating heat loads that require sophisticated cooling systems and larger electrical infrastructure than conventional data centers.

How Is Meta Solving This Challenge in Alberta?

Meta's Canadian project, announced on July 8, 2026, represents the company's first data center in Canada and the 33rd facility in its global portfolio. The campus is planned initially at 1 gigawatt (GW) of power capacity, with potential to scale to 1.8 GW, making it among the largest data center developments under construction outside the United States.

The key innovation is the integrated approach: Meta coordinated the data center development with a purpose-built 932 MW combined-cycle natural gas generating station called the Greenlight Electricity Centre, grid upgrades, and long-term natural gas transportation agreements. This arrangement, structured as a tolling agreement, allows Meta to make capacity payments while covering fuel, operations, and maintenance costs.

The Greenlight facility is owned by Pembina Pipeline (47.5%), Morgan Stanley Infrastructure Partners (47.5%), and Kineticor (5%), with an estimated capital cost of approximately C$4 billion rising to C$4.6 billion when construction interest and financing costs are included. The plant is scheduled to begin service during the second half of 2030.

The site's location within Alberta's Industrial Heartland, a designated development zone northeast of Edmonton, provides existing infrastructure including pipelines, energy facilities, and petrochemical plants. This industrial setting minimizes community concerns and allows Meta to colocate the data center with the enormous generating capacity required to operate it.

What Makes This Data Center Environmentally Different?

One of the most frequent criticisms of large AI data centers is their water consumption for cooling systems. Meta's Sturgeon County facility addresses this concern through closed-loop liquid cooling supported by dry cooling technology. The company states the cooling system will consume no water during normal operation, with water use limited primarily to domestic requirements, fire protection, and equipment maintenance.

This design choice is significant because it eliminates operational cooling-water consumption, a major environmental concern for AI facilities. While dry cooling can require more equipment and may use more electricity under certain weather conditions than evaporative systems, Alberta's comparatively cool climate allows designers to use low outside-air temperatures to improve heat-rejection efficiency during much of the year.

Ways Data Centers Are Addressing the Global Power and Sustainability Challenge

• Dedicated Power Generation: Hyperscalers like Meta are building purpose-built power plants alongside data centers rather than relying solely on grid connections, ensuring reliable electricity supply and reducing interconnection delays.
• Renewable Energy Integration: The economics of solar power installations paired with battery storage are becoming increasingly competitive, enabling sustainable microgrids that can supplement or replace fossil fuel generation for data center operations.
• Advanced Cooling Technologies: Transitions from air cooling to direct-to-chip liquid cooling and immersion cooling reduce greenhouse gas emissions, water usage, and energy consumption while improving performance and efficiency.
• Small Modular Reactors: Emerging nuclear small modular reactors (SMRs) are attracting hyperscaler investment, with forecasts suggesting SMRs could provide up to 15% of data center power by 2037.
• Component-Level Efficiency: Leading suppliers like NVIDIA, AMD, and Infineon are emphasizing energy efficiency improvements at the GPU, CPU, and power converter levels to reduce overall data center power consumption.

Beyond electricity and cooling, data centers face a broader sustainability challenge. Scope 3 emissions, which represent indirect emissions in a company's value chain, actually constitute the majority of CO2 emissions from data centers. IDTechEx forecasts data center CO2 emissions will exceed 0.8 gigatonnes per year by 2036, with Scope 3 emissions playing a large part.

Companies are addressing these indirect emissions through three primary strategies: purchasing carbon dioxide removal credits to counteract hard-to-avoid emissions, using low-carbon materials in data center construction such as green concrete, green steel, and timber, and choosing IT hardware with lower manufacturing carbon footprints over the lifetime of a data center.

What Does This Mean for the Future of AI Infrastructure?

Meta's Alberta project exemplifies a new model where data centers are paired with dedicated power generation, influencing regional energy and infrastructure policies. The arrangement represents Alberta's emerging "bring your own power" strategy for hyperscale infrastructure development.

The project's scale is substantial: approximately 3,000 construction workers are expected at peak activity, while more than 300 permanent employees will operate the campus after completion. Meta is also committing approximately C$60 million to improvements involving local roads, water systems, and other community infrastructure.

The buildout will occur in phases rather than delivering the entire gigawatt at once, with Alberta's major-project registry estimating a roughly three-year construction period. The work extends far beyond data halls, typically involving multiple computing buildings, substations, utility yards, backup power systems, administrative space, security facilities, equipment staging areas, and miles of internal electrical, mechanical, and fiber infrastructure.

As global data center power demand continues its rapid acceleration, the Meta-Greenlight model demonstrates that hyperscalers are moving beyond waiting for grid capacity and instead building integrated energy ecosystems. This approach may become the template for future AI infrastructure development, particularly in regions where grid constraints threaten to limit expansion. The success of this integrated model could reshape how governments, utilities, and technology companies collaborate to support the infrastructure demands of artificial intelligence.