A company can have the latest AI processors, millions of dollars in computing equipment, and a fully designed data center. But without electricity, it's just an expensive building filled with idle technology.
That's becoming a serious problem for the artificial intelligence industry. As demand for computing capacity accelerates, traditional electrical infrastructure is struggling to meet the enormous power needs of modern data centers. Utility connections can take years, transmission capacity is limited in many markets, and developers are increasingly searching for alternatives.
One approach is gaining attention: generate electricity where the computing takes place.
Known as behind-the-meter power generation, it's at the center of a new infrastructure strategy being pursued by FingerMotion, Inc. (NASDAQ: FNGR) and its Canadian development partner, BlueFlare.
Their initial focus is Alberta, where natural gas resources and established industrial infrastructure offer an opportunity to build computing facilities with their own dedicated power generation. The concept isn't particularly complicated. The engineering required to make it reliable, scalable, and commercially practical is another matter.
Why the Grid Is Becoming a Bottleneck
Artificial intelligence is changing the electrical demands of data centers. Traditional facilities were built to accommodate relatively predictable computing workloads. Today's AI systems can require dense clusters of advanced processors operating continuously, producing substantial heat and consuming large amounts of electricity.
The International Energy Agency has projected that worldwide data center electricity consumption could roughly double by 2030.
Alberta provides an interesting example of the pressure facing energy infrastructure. Provincial figures cited in recent industry reporting showed approximately 19.6 gigawatts of proposed data center demand seeking grid connections in July 2026. That's considerably greater than Alberta's historical peak internal electricity demand of roughly 12.8 gigawatts.
Not all proposed projects will necessarily be built, but the numbers illustrate the scale of interest. For developers, the challenge extends beyond securing an adequate supply of electricity. It's also about timing.
A company ready to deploy AI computing equipment may not be prepared to wait several years for the necessary utility infrastructure. That creates an opening for facilities that can produce their own power.
How Behind-the-Meter Generation Works
The conventional data center model depends heavily on electricity delivered through the utility grid. Behind-the-meter generation changes that relationship by producing electricity directly at or near the facility where it's consumed.
FingerMotion and BlueFlare intend to use natural gas-powered generators at their Alberta campuses, providing dedicated electricity for computing equipment, cooling systems, and supporting infrastructure.
This approach can reduce exposure to lengthy grid-interconnection queues and potentially provide developers with greater control over power availability.
There are engineering considerations, of course. Fuel supply must be dependable, generators require maintenance, and facilities need sufficient backup capacity to keep critical equipment operating.
But the underlying advantage is significant. A data center with its own reliable electricity supply may be able to advance on a timetable driven more directly by construction and customer requirements than by utility expansion schedules.
Brooks Campus #1 Provides the First Engineering Blueprint
FingerMotion's first major project is Brooks Campus #1 in Alberta's County of Newell. On November 18, the company announced that BlueFlare had completed the facility's Design Basis Memorandum, establishing the engineering specifications for its first planned colocation campus.
The design calls for 9.6 megawatts of installed natural gas generation supporting approximately 7.2 megawatts of critical IT capacity. The difference between those figures reflects the electricity required to operate cooling, power distribution, and other supporting systems, as well as the facility's design requirements.
The generators are configured with N+1 redundancy, meaning the system is designed with additional generation capacity to support reliability if a unit becomes unavailable.
Cooling is another important consideration. The Brooks design specifies air-cooled IT equipment in standard 19-inch racks, supported by rear-door heat exchangers and associated cooling infrastructure.
BlueFlare is targeting a power usage effectiveness, or PUE, of 1.20. PUE measures how efficiently a data center uses electricity. A value of 1.20 means the facility is designed to consume approximately 1.2 units of total electricity for every unit delivered to computing equipment. Lower values generally indicate more efficient infrastructure.
That 1.20 figure is a design target, not an operating measurement from a completed facility.
The engineering memorandum also establishes a phased development schedule, with initial capacity targeted for delivery approximately six months after a binding customer agreement becomes effective. Full capacity is targeted within 12 to 14 months.
Smaller Facilities, More Flexible Deployment
FingerMotion's approach differs from the enormous hyperscale campuses being developed for the world's largest technology companies. Its proposed facilities are designed for enterprise customers requiring several megawatts of computing capacity.
These customers may need more electricity and physical space than conventional retail colocation facilities can provide, without requiring hundreds of megawatts at a single location.
The company's announced Alberta program encompasses ten proposed sites, each designed around approximately 9.9 megawatts of gross generation capacity. Together, they represent 99 megawatts across the Newell, Coronation, Fox Creek and Vulcan regions.
Rather than designing each facility entirely from scratch, BlueFlare intends to use Brooks Campus #1 as a repeatable engineering template. Standardizing power generation, cooling, and facility design could reduce engineering time and simplify construction planning across multiple locations.
Customers would also have options. Some may prefer powered land on which to install their own modular computing facilities. Others may want a complete colocation environment, with electricity, cooling, and computing-ready space already provided.
In either case, customers would generally supply their own servers and processors. FingerMotion focuses on the infrastructure surrounding that equipment.
Turning Engineering Into Operating Capacity
The company has already reported prospective customer interest at Brooks Campus #1 involving up to seven megawatts of critical computing capacity. The contemplated arrangement could represent more than $100 million in contracted revenue over five years if definitive agreements are completed.
That commercial interest gives the engineering work a practical purpose. A customer needs computing capacity, the developer designs the power and physical infrastructure to deliver it, and construction can proceed as financing and contractual commitments are secured.
The approach remains in development. Brooks Campus #1 still requires binding customer agreements, financing, construction, and commissioning before it can begin operating.
But its recently completed engineering memorandum establishes a more detailed technical foundation for that process, with implications extending beyond one facility.
AI infrastructure development increasingly depends on coordinating electricity, computing equipment, cooling, and construction schedules. A shortage in any one of those areas can delay an entire project.
FingerMotion and BlueFlare are pursuing a model that addresses the electricity problem at the source.
The next generation of AI may depend on increasingly sophisticated computing technology. Getting that technology running could depend just as much on something far more fundamental: having the power ready when the equipment arrives.
Sources and References
International Energy Agency — Energy and AI, 2025–2026 research and electricity-demand outlook; Government of Alberta — Data Centre Development and Electricity Demand Statistics, 2026; FingerMotion, Inc. (NASDAQ: FNGR) — September 15, 2026, 99 MW Alberta Behind-the-Meter Development Program; November 4, 2026, Brooks Campus #1 Customer Expression of Interest; November 18, 2026, Brooks Campus #1 Design Basis Memorandum. Additional Sources: https://ir.fingermotion.com/newsroom/press-releases; https://www.iea.org; https://www.alberta.ca/datacentres/numbers
