TLDR
- Small-scale, behind-the-meter natural gas data centers in Alberta offer a lighter-footprint model for AI infrastructure: 57% less CO₂ per megawatt-hour than coal, up to 70% less direct water consumption than traditional evaporative cooling, and zero draw on the public electrical grid.
- With grid interconnection timelines now averaging nearly five years in the United States, behind-the-meter generation in energy-advantaged regions like Alberta is emerging as the most viable near-term path to deploying AI compute capacity without straining public infrastructure.
- Alberta’s cooler climate, abundant natural gas supply, and position on the Western Canadian Sedimentary Basin create a combination of environmental and operational advantages that few North American jurisdictions can match for responsible compute deployment
Artificial intelligence’s rapid growth is driving a similarly rapid expansion in compute infrastructure. AI data centers require large amounts of power, cooling, and physical space, and their environmental footprint has become a subject of legitimate public and regulatory scrutiny.
The answers to most environmental questions about data centers depend on specific design choices, the most critical being fuel source, cooling architecture, relationship to the public electrical grid, and what the local climate and geology make possible.
One configuration that checks those boxes particularly well is small-scale, behind behind-the-meter, natural gas data centers with closed-loop cooling. This data center model’s benefits, both environmentally and in speed-to-power, are complementary: the same design choices that reduce the model’s footprint also enable it to be deployed faster, while larger or grid-dependent projects wait in line for power.
This model has benefits worth exploring in terms of carbon emissions, water consumption, and grid impact, and sustainable opportunities are emerging in Alberta, Canada. Due to the region’s natural resources and geographic advantages, Alberta is emerging as an advantageous region for responsible AI infrastructure development.
Carbon Emissions: Natural Gas, Behind-the-Meter Efficiency, and Alberta’s CCUS Potential
AI data centers require firm, high-availability power. Coal-fired electricity generation produces 2,257 pounds of CO₂ per megawatt-hour. Natural gas-fired generation produces 976 pounds, a 57% reduction, along with substantially lower sulfur dioxide, nitrogen oxide, and particulate emissions.¹ Alberta has seized on these improvements, having already completed most of its coal phase-out.
The province’s grid is now dominated by natural gas, which means on-site behind-the-meter generation operates at essentially the same carbon intensity as grid-delivered electricity, while eliminating the roughly 5% of energy lost in transmission and distribution.² More of the fuel’s energy content reaches the compute load directly.
The behind-the-meter configuration creates opportunities that grid-connected facilities cannot easily access. On-site generation supports a variety of efficiency gains. For example, some natural gas in Western Canada is flared or vented at the wellhead, and a behind-the-meter setup can put a portion of this otherwise-wasted gas to productive use while qualifying under Alberta’s carbon-credit programs.³ A related efficiency for behind-the-meter generation is the integration of combined heat and power (CHP) systems, which capture waste heat that would otherwise be vented and repurpose it for secondary applications. The U.S. Department of Energy documents efficiency improvements of as much as 40% from CHP integration relative to producing heat and power separately.⁴
The longer-term emissions picture in Alberta is even more favorable thanks to its geology. Alberta sits on the Western Canadian Sedimentary Basin (WCSB), which the Government of Alberta identifies as having ideal characteristics for carbon capture, utilization, and storage (CCUS): the reservoir rock, sealing layers, depth, and tectonic stability required for permanent CO₂ containment.⁵
Research published in the Journal of Petroleum Exploration and Production Technology confirms that the WCSB offers excellent conditions for geological CO₂ sequestration.⁶ Norton Rose Fulbright notes that the western Canadian provinces collectively hold a potential CO₂ storage capacity of 385 gigatonnes.⁷
Data center projects in Alberta that incorporate future CCUS integration into their infrastructure planning can reduce effective emissions by as much as 70% compared to unabated natural gas combustion.⁸
Water Consumption: What Closed-Loop Cooling Achieves
Water consumption is among the most closely tracked environmental metrics in data center operations. A 2024 federal report drawing on research from Lawrence Berkeley National Laboratory estimated that U.S. data centers directly consumed 17.4 billion gallons of water in 2023, roughly equivalent to the annual consumption of 160,000 American households.⁹ The International Energy Agency (IEA) estimates that a single 100 MW facility can consume approximately 530,000 gallons per day on average across cooling configurations.¹⁰ Most of this consumption is driven by evaporative cooling towers, which eject heat by evaporating water into the atmosphere.
A closed-loop cooling system operates more efficiently. It recirculates the same water through the facility in a sealed circuit, exchanging heat without evaporating the working fluid. Advanced implementations pair closed-loop chilled water systems with direct liquid cooling at the server level. Water losses are limited to minor makeup quantities for leak replacement. The Environmental and Energy Study Institute documents water consumption reductions of up to 70% compared to evaporative approaches, highlighting the clear benefit of a closed-loop cooling system.¹¹
Alberta’s climate provides an additional advantage. Cooler average ambient temperatures, particularly outside summer months, reduce the thermal differential that cooling systems must overcome. This means facilities can run on economizers or free-cooling rather than active chilling for a longer portion of the year, reducing both energy consumption and any residual water use.
The combination of closed-loop architecture and Alberta’s climate produces one of the lowest water consumption profiles available in AI compute infrastructure today.
Grid Impact: Behind-the-Meter Generation is Better for Everyone
Environmental impact is not the only public issue with the AI infrastructure build-out. Behind-the-meter generation means power is produced on-site, consumed on-site, and never routed through the utility grid. A data center running on behind-the-meter power does not add to grid load.
The current state of North American grid infrastructure makes this feature even more important. As of the end of 2024, approximately 10,300 projects were actively seeking grid interconnection in the United States, representing 1,400 gigawatts of pending capacity.¹² Enverus Intelligence Research found that queue-to-commercial-operation timelines now average over 2,100 days, a 60% increase since 2017.¹³ RMI reports the average time from interconnection request to commercial operation has risen to nearly five years.¹⁴ While the broad generation interconnection queue has begun to contract as speculative projects withdraw,¹⁵ large-load data-center interconnection requests have surged. The Alberta Electric System Operator is fielding more than 20 gigawatts of pending data-center power requests against roughly 12 gigawatts of provincial peak demand.¹⁶
For communities near grid-connected data center developments, interconnection demand can translate into grid strain, rate increases for residential customers, and pressure on utilities to accelerate infrastructure investment ahead of schedule. Behind-the-meter generation eliminates these challenges entirely. The data center neither competes for public grid capacity nor imposes costs on local ratepayers. Furthermore, behind-the-meter generation speeds up the timeline to energize, accelerating progress while at the same time mitigating public burden.
Alberta’s regulatory environment strongly supports this model. Bill 8, the Utilities Statutes Amendment Act of 2025, fast-tracks approvals for projects that bring their own generation, reducing development timelines relative to grid-dependent alternatives. Bill 12, enacted alongside it, levies grid-connected facilities of 75 megawatts or more and eases that levy as a project supplies its own power.¹⁷ The province’s established natural gas pipeline infrastructure provides reliable fuel supply without new buildout requirements.
Natural gas also provides dispatchable baseload power: on-demand generation at consistent output levels regardless of weather or time of day. AI inference workloads, in particular, require high-uptime power. Intermittent renewables are essential to the broader grid, but they cannot carry a continuous inference load on their own. Tier 3 data center specifications call for 99.982% uptime.¹⁸ Natural gas behind-the-meter generation meets this requirement without grid dependency.
Why Alberta’s Combination of Advantages Is Difficult to Replicate
The list of environmental and public advantages described so far stacks on top of each other to make Alberta a very attractive location for responsible AI infrastructure development.
Abundant natural gas supply and established pipeline infrastructure support reliable, low-cost behind-the-meter generation. A cooler climate reduces cooling energy requirements and enables closed-loop systems to operate with minimal water consumption and lower power draw. Geology provides a credible, well-documented pathway to long-term carbon capture that many compute regions lack. And Alberta’s regulatory framework actively supports self-generation projects, compressing development timelines.
AI infrastructure power demand is projected to grow at roughly 15% annually through 2030, according to the IEA.¹⁹ As demand scales, the location and configuration of new compute infrastructure becomes a question of genuine public consequence. Facilities sited in energy-advantaged regions with responsible power and cooling configurations produce superior outcomes than those built elsewhere without the same advantages.
Small-scale, behind-the-meter, natural gas data centers in Alberta provide significant advantages across emissions, water, and grid impact. They benefit both the public and operators who need reliable, high-uptime compute capacity at AI scale today.
One Company Building to This Standard
AVAX One Technology Ltd. (Nasdaq: AVX) is a power-first digital infrastructure company developing data centers to these specifications.
The company’s initial 10 MW Tier 3-ready AI/HPC site in Alberta is built on a fully behind-the-meter natural gas model with closed-loop cooling and the inherited advantages of the Western Canadian Sedimentary Basin. BlueFlare Energy Solutions Inc., a Calgary-based energy infrastructure company with established operations across Western Canada’s oil, gas, and behind-the-meter power sectors, has been engaged as infrastructure development partner. ASCENT Consulting Ltd. has been selected as Owner’s Engineer. The 10 MW facility is targeted for Q1 2027 client deployment readiness.
An AI inference pilot launched June 16, 2026 at the company’s existing Redwater facility will convert approximately 100 kilowatts of existing behind-the-meter power capacity to AI inference workloads as a low-cost, low-risk pilot to validate the company’s enhanced operating model in advance of a full-scale buildout.
AVAX One is responding to the AI infrastructure opportunity, sees the business, public, and environmental advantages of its model, and has chosen Alberta and Western Canada more broadly as its starting point.
More information is available at avax-one.com.
This article contains forward-looking statements within the meaning of applicable securities laws. Actual results may differ materially from those expressed or implied. AVAX One Technology Ltd. does not undertake to update forward-looking statements except as required by law.
REFERENCES
- U.S. Energy Information Administration. “Electric power sector CO2 emissions drop as generation mix shifts from coal to natural gas.” EIA Today in Energy. https://www.eia.gov/todayinenergy/detail.php?id=48296
- U.S. Energy Information Administration. “Frequently Asked Questions: How much electricity is lost in electricity transmission and distribution in the United States?” https://www.eia.gov/tools/faqs/faq.php?id=105
- American Bar Association. “Beyond the Flames: Transforming Wasted Gas into a Power Source for AI Data Centers.” 2025. https://www.americanbar.org/groups/environment_energy_resources/resources/natural-resources-environment/2025-summer/beyond-flames-transforming-wasted-gas-power-source-ai-data-centers/
- U.S. Department of Energy, Office of Energy Efficiency & Renewable Energy. “Combined Heat and Power Technology Fact Sheet: Overview.” https://betterbuildingssolutioncenter.energy.gov/sites/default/files/attachments/Overview_of_CHP_Technologies.pdf
- Government of Alberta. “Carbon Capture, Utilization and Storage: How It Works and Benefits.” https://www.alberta.ca/carbon-capture-utilization-and-storage-how-it-works-and-benefits
- Journal of Petroleum Exploration and Production Technology. “CO2 retention mechanism storage in Alberta tight oil and gas reservoirs at Western Canadian Sedimentary Basin, Canada.” Springer, 2021. https://link.springer.com/article/10.1007/s13202-020-01052-7
- Norton Rose Fulbright. “Understanding carbon capture and storage in Canada.” 2025. https://www.nortonrosefulbright.com/en/knowledge/publications/91fbfcdf/understanding-carbon-capture-and-storage-in-canada
- Carbon Direct. “Carbon capture for natural gas-fired power generation: an opportunity for hyperscalers.” 2024. https://www.carbon-direct.com/insights/carbon-capture-for-natural-gas-fired-power-generation-an-opportunity-for-hyperscalers
- MOST Policy Initiative / Lawrence Berkeley National Laboratory (Shehabi et al. 2024). “Data Center Water Use.” https://mostpolicyinitiative.org/science-note/data-center-water-use/
- Congressional Research Service, citing International Energy Agency. “Data Centers and Their Energy Consumption: Frequently Asked Questions.” R48646. https://www.congress.gov/crs-product/R48646
- Environmental and Energy Study Institute (EESI). “Data Centers and Water Consumption.” https://www.eesi.org/articles/view/data-centers-and-water-consumption
- Lawrence Berkeley National Laboratory. “Queued Up: 2025 Edition, Characteristics of Power Plants Seeking Transmission Interconnection As of the End of 2024.” https://emp.lbl.gov/publications/queued-2025-edition-characteristics
- Enverus Intelligence Research. “Why Data Centers Are Looking to Natural Gas for Behind-the-Meter Power.” June 2026. https://www.enverus.com/blog/why-data-centers-are-looking-to-natural-gas-for-behind-the-meter-power/
- RMI. “The Interconnection Queue Continues to Be a Barrier to US Economic Competitiveness.” March 2026. https://rmi.org/interconnection-reform-ai-data-centers-generator-queues/
- Lawrence Berkeley National Laboratory. “Queued Up: 2025 Edition, Characteristics of Power Plants Seeking Transmission Interconnection As of the End of 2024.” https://emp.lbl.gov/publications/queued-2025-edition-characteristics
- Alberta Electric System Operator. “Large Load Projects.” https://www.aeso.ca/grid/connecting-to-the-grid/large-load-projects/. Demand and peak figures per McCarthy Tétrault, “Alberta Faces a Surge in AI Data Centre Power Demand: AESO Responds with Phased Connection Plan,” 2025. https://www.mccarthy.ca/en/insights/blogs/canadian-energy-perspectives/alberta-faces-a-surge-in-ai-data-centre-power-demand-aeso-responds-with-phased-connection-plan
- Gowling WLG. “Alberta is reshaping its electricity system for data centres.” 2026. https://gowlingwlg.com/en/insights-resources/articles/2026/alberta-reshaping-electricity-system-for-data-centres
- PowerMag. “Understanding Uptime Institute’s Tier III Standard: A Guide to Data Center Electrical System Design.” https://www.powermag.com/understanding-uptime-institutes-tier-iii-standard-a-guide-to-data-center-electrical-system-design/
- International Energy Agency. “Energy and AI.” 2025. https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai
