The province is making a multibillion-dollar wager that demand for massive AI data centres will keep growing

We’re all familiar with it: that quiet little voice that whispers an offer is too good to be true. Our Spidey sense tells us that there has to be a catch, even if we don’t yet know what it is.

I had that feeling when I learned of plans to build several power-gobbling data centres in Alberta. Proponents promise prosperity, jobs and an unprecedented economic boost to regions hungry for investment. It’s the sort of boosterism that has brought huge smiles to Alberta Premier Danielle Smith, whose government envisions turning the province into an artificial intelligence and data hub.

But there is good reason to question not only whether these promises will be fulfilled, but also at what cost. Before we rush headlong into the AI data-centre business, we need a much clearer idea of what we’re getting into.

Consider the scale of what is being proposed.

The Mihta Askiy Data Center Project, about 40 kilometres northeast of Peace River, would include a 650-megawatt natural gas-fired power plant.

Then there’s the Wonder Valley project in the Municipal District of Greenview, south of Grande Prairie. Promoted by “Mr. Wonderful” Kevin O’Leary through O’Leary Ventures, the proposed AI data centre campus could, at full buildout, reach 7.5 gigawatts and span thousands of acres, relying on natural gas and geothermal infrastructure.

Meta, the parent company of Facebook and Instagram, meanwhile, is investing more than $13 billion in an AI data centre north of Edmonton. The company says the project will support more than 3,000 construction jobs, but about 300 permanent positions once it is operating. The facility is expected to consume about one gigawatt of electricity.

And in Olds, the proposed $10-billion Synapse AI data centre and gas-fired power plant has generated extraordinary opposition, with more than 1,100 statements of intent to participate filed in the regulatory process.

These are enormous projects, amounting to a multibillion-dollar wager that demand for AI computing will keep growing for years. Maybe that will prove to be a very good bet. But Alberta seems awfully eager to push all its chips onto the table before some pretty basic questions have been answered.

These facilities can consume staggering amounts of electricity, and some require significant amounts of water. Residents worry about industrial noise and development.

And there is another question that deserves more attention: what happens to electricity and natural gas prices when enormous new industrial users start competing for energy?

It also strikes me as more than a little ironic that Alberta has restricted the development of solar projects, in part because of concerns about taking agricultural land out of production, while welcoming data centres that could occupy thousands of acres.

Apparently, acres are not all created equal. None of those concerns alone is necessarily a reason to reject a data centre. Taken together, though, they deserve a lot more scrutiny than they appear to be getting.

Maybe those concerns would be easier to swallow if the economic promise of these projects were as certain as their promoters suggest.

But there’s another reason my Spidey sense is tingling: what if Alberta is making this enormous bet just as the technology begins to change?

Robert Diab, a Thompson Rivers University law professor who writes about technology, recently argued in The Globe and Mail that AI data centres could be overbuilt or even obsolete by the time some are completed.

That’s because some AI processing is moving away from giant cloud-based centres and onto the devices sitting on our desks and in our pockets. A friend recently showed me an app called ChatJimmy, which uses a specialized computer chip designed specifically to run an AI model. Apple and other technology companies are also moving more AI processing directly onto phones and computers.

No one can say with certainty where that technological shift will lead. Training massive AI models and handling complex applications still require enormous computing resources.

But when companies are proposing facilities costing many billions of dollars, the possibility that computing needs could change dramatically deserves more than a shrug.

I’m also not convinced the province is taking the potential environmental consequences seriously enough.

I recently received a form letter from Environment and Protected Areas Minister Grant Hunter explaining when Alberta requires an environmental impact assessment.

Hunter said an assessment is required where potential environmental effects are uncertain, unproven technologies are involved or cumulative effects may significantly affect the environment. He added that data centres use established technologies for power generation and water infrastructure that are already understood and regulated in Alberta.

If I’m reading that correctly, the argument is that because we understand the technologies involved, a data centre doesn’t necessarily require an environmental impact assessment.

I’m pretty sure that response won’t satisfy many of the people who will have to live beside these projects.

And it doesn’t really answer the bigger question.

The concern isn’t whether Alberta understands how a natural gas power plant works. It is whether anyone is stepping back and considering what happens when numerous massive data centres, power plants and water users begin operating across the province.

To be fair, Alberta has some obvious advantages in attracting data centres: abundant natural gas, available land, engineering expertise and communities hungry for investment. There may well be a significant economic opportunity here.

It’s too soon to declare that Alberta’s proposed data centres will become a collection of white elephants. But there is good reason to wonder whether some will live up to promises of massive economic benefits and long-term job creation.

That’s why some basic questions still need answers. What will these projects do to energy prices? How many permanent jobs will they actually create? And what happens if technological change makes today’s extraordinary demand forecasts look wildly optimistic?

Maybe there are convincing answers to those questions. If so, Albertans deserve to hear them.

There is widely accepted wisdom that if your friend at the coffee shop gives you a hot tip on a stock, it’s probably already too late to buy. Alberta should keep that in mind.

Because that little voice is still whispering: if something sounds too good to be true, there is usually a catch.

The data-centre boom may turn out to be every bit as transformative as its promoters claim. Or Alberta may be arriving at the party just as everyone else starts looking for their coats.

Either way, the government needs to smarten up and slow down.

Doug Firby is an award-winning editorial writer with over four decades of experience working for newspapers, magazines and online publications in Ontario and western Canada. Previously, he served as Editorial Page Editor at the Calgary Herald.