Alberta’s Data Center Boom: Economic Opportunity or Electricity Affordability Crisis?

What is true for Alberta, is true for many jurisdictions in North America that remain conflicted over investing in weather dependent electricity or embracing in full, the 4th Industrial Revolution.

Alberta’s hyperscale data center market is booming, with investments potentially hitting $100 billion, fueled by low taxes, cheap natural gas, and a robust fiber optic network.

However, this economic opportunity comes with risks.

Premier Danielle Smith’s mandate that data centers bring their own power conflicts with approvals like eStruxture’s $750 million Calgary project, reliant on a power purchase agreement. Alberta’s history of electricity inflation, driven by mandated renewable transitions, raises concerns about grid reliability and affordability.

As AI-driven data centers demand massive amounts of high quality power, Alberta must balance economic diversification with protecting ratepayers from soaring costs and an unstable grid.

McKinsey estimates that by 2030, the growth of AI focused data centers in United States may well consume as much as 100 TWh more power as Canada consumed in 2023 if sufficient power generation can be built fast enough.

"If" is the key word to use in this context, especially in view of the Electrification Transition that we have heard about non-stop since 2020 here in North America and in my home province of Alberta, Canada.

Here in this Substack article, I call into question whether Alberta will manage this opportunity properly and ask whether the monstrous appetite that investors have for AI focused data centers will simply cause more energy inflation?

The Alberta hyperscale data center market has of late been receiving high profile accolades as an optimal investment destination in North America due to its low income tax, cheap natural gas and large fiber optic network.

According to Invest Alberta’s CEO Rick Christiaanse, data center investment in Alberta represents a $75 to $100 billion opportunity.

However, Premier Danielle Smith has recently stated that investors must bring their own power if they want to bring data centers to the province.

The reason for this red-line is that Alberta’s power market experienced mass inflation following the Alberta NDP’s mandate in 2016 that the provincial grid achieve 30% wind plus solar capacity by 2030.

However, in spite of this red-line, Montreal based eStruxture just announced it was proceeding with a $750 million hyperscale data center north of Calgary and that its 90 MW grid load has been contracted.

Meaning, eStruxture engaged in a power purchase agreement with an independent power producer and will not be bringing its own power.

This is important, given eStruxture would not have announced its investment decision unless it had received regulatory approval from the Alberta Utilities Board (AUB) and the Alberta Electric System Operator (AESO) for their grid connection permit.

In other words, the Premier drew a red-line in the sand over data centers bringing their own power and then within a matter of a few months, the AUB and AESO approved a large hyperscale data center without its own power source.

This is not the sign of effective governance.

Figure 1 shows the geographical distribution of Alberta’s fiber optic network, as well as the project connection queue list with AESO for both new large loads and power supply within the province as of the end of 2024.

Note that loads are categorized as data, industrial, distribution and contract, while supply is categorized by fuel or energy source (e.g., solar, gas).

Figure 1. Alberta’s fiber optic network and load versus power supply queue.

The main take-away points from Figure 1 are two fold.

First, there are over 16,000 MW of wind and solar PV projects in queue and only 3,500 MW of gas powered thermal generation. Note that 16,000 MW of wind and solar PV is 160% larger than the total capacity of commissioned and under construction projects in the province.

Second, 89% of the new load in queue to connect to the grid is data centers and there is 2x more load from data centers proposed for connection to Alberta’s grid than there is new gas powered generation seeking grid connection.

You can not run a billion dollar data center off a wind or solar PV farm.

Know that not all proposed projects are approved, so we have no idea of which projects will receive grid connection permits.

Note that I have sent multiple requests for information and dialogue from both the Ministries of Utilities and Affordability and Technology & Innovation, which have overlapping mandates on the file of power generation and data center roll-out in the province.

Neither ministry responded.

My concerns here are that Premier Smith’s red-line appears to be falling on deaf ears and the carrot to diversify the economy will over power concerns over electricity affordability.

The need for a hard red-line on hyperscale data centers is entirely justified in view of the following facts regarding Alberta's recent bitter experience with electricity inflation:

1. The compound annual growth rate (CAGR) for transmission and distribution infrastructure in Alberta between 2004 and 2023 was 13% and 2%, respectively;

2. The 6-fold higher CAGR for transmission versus distribution is almost entirely due to the massive public investment in dedicated high voltage transmission lines, in large part to support wind & solar PV infrastructure in the south regions of the province while load centers are in the central to northern regions; and

3. The CAGR for Alberta's pool price was 40% between 2016 and 2023 (Figure 2) and was largely driven by the mandated early retirement of almost half of Alberta's coal power (2885 MW) and the conversion of the remaining half of the fleet (2,934 MW) to function largely as natural gas fueled peaking power generators.

Figure 2 is a time series demonstrating the impact of decommissioning and converting Alberta's coal fleet on the monthly average power pool price.

Figure 2. Ten year Alberta power pool price history.

Those who are not familiar with Alberta history, it is important to note that the Alberta NDP (socialist) party legislated that independent power producers implement 30% wind and solar in 2016, which triggered the end of our stable power purchase agreement era and the transition to an energy based market.

Another compounding fact at play was the explosive growth in population (almost entirely due to immigration).

Figure 3 shows the growth in power generation in GWh versus population between 2017 and 2023. When we divide the change in power generation (AIL) relative to the growth in population, the ratio of the two is a mere 7 MWh per capita.

Figure 3. The Alberta Internal Load (AIL) versus population time series.

Putting this in perspective, in 2017 the per capita electricity consumption was 19.4 MWh and the Canadian average was 13.7 MWh.

Had generation expanded consistent with historical trends and within increasing demand, approximately 10,200 GWh of consumption growth would have occurred versus the dismal 3,700 GWh.

The perfect storm for electricity inflation.

Those who will argue that wind, solar and batteries are the low cost and most affordable path forward, ignore the fact that the retail consumer pays for the full system cost of the grid, while independent power producers simply pay for their facilities and grid tie-in.

Full system = Generation + Transmission + Distribution + Storage.

Table 1 shows the basis of my estimated capital cost for the roll out of solar PV, wind, batteries and dedicated new transmission in the province, as well as the $1.36 billion carbon tax break the Alberta NDP gave coal power producers in compensation for breaking their capacity contracts and forcing them to decommission their facilities.

Table 1. Estimated capital costs for producers and rate payers for Alberta’s current renewable infrastructure

As shown, this estimate includes both commissioned facilities as well as those under construction.

Of this estimated $26 billion, at least $7.5 billion was directly paid for by rate (tax) payers; unlike pipelines, transmission infrastructure is paid for by Albertan rate payers.

I have not yet been able to account for new transmission infrastructure projects commissioned between 2019 and 2024.

Likewise, note that I used $1,500, $1,700 and $600 per kW specific cost for solar PV, wind and utility scale batteries, respectively, in Table 1, which are representative values taken from recent projects in Alberta.

However, as these projects wouldn’t exist without billions in publicly funded dedicated transmission infrastructure, the total full system specific cost is at least $2,470 per kW. Once I find public records on the capital cost of transmission infrastructure between 2019 and 2024, this full system capital cost will rise beyond $2,470 per kW.

It is fair to say the full system specific costs, ultimately borne by the retail consumer, varies between $2,500 to $3,000 per kW.

Compare this to a 900 MW combined cycle gas turbine (CCGT) system was commissioned west of Edmonton in April 2024 that cost $1.5 billion or $1,660 per kW or to a CCGT + carbon capture and storage (CCS) in queue to be built in the Red Deer Alberta region for an estimated specific capital cost of $2,940 per kW.

Note that while I do not endorse CCS technology, it is difficult to not acknowledge the fact that the full system cost of wind, solar PV, batteries and long haul transmission in Alberta appears to be on par with CCGT + CCS.

Both systems would have comparable CO2 intensities, but it is the CCGT system that has the power quality (i.e., high frequency inertia) that a multi-billion dollar AI focused data center would require.

Further, when we consider that CCGT systems will not need dedicated long haul transmission built on the public dime, can be co-located near load centers and that its capacity factor will likely be 3 to 4 times that of utility scale solar PV and wind units, it is clear which technology mix is optimal from a rate payer's perspective.

I conclude this article by predicting that as long as politicians and regulators remain conflicted over non-dispatchable wind, solar, batteries and their back-up natural gas turbines versus higher capacity factor dispatchable technologies like coal power, nuclear and CCGT systems, hyperscale data center investments represent a clear and present danger for rate payers.