The Objections That Refute Themselves

Alberta's data centre critics claim the province regulates nothing, that investment is fleeing, and that a bust would ruin Albertans.

Alberta’s data centre debate runs on three claims: that the province regulates nothing, that political turbulence is driving investment away, and that an AI bust would leave Albertans holding the wreckage. Each claim collapses on inspection. Each collapse is instructive.

One of the largest private investments in Canadian history landed this summer in a rural county north of Edmonton. The people who spent the past decade demanding exactly this kind of investment now want it stopped.

Progressives and the Alberta NDP have opened a campaign to torpedo the data centre projects taking shape across the province. Their objections repay a close reading because the arguments tell us a great deal about the movement’s neuroses. Less than a decade ago, these same voices clamoured incessantly for exactly what has now arrived: diversify the economy, attract technology, end the dependence on hydrocarbons.

Three claims carry their campaign: Alberta regulates nothing, political turbulence is scaring capital away, and an AI bust would leave Albertans holding the wreckage. Each claim refutes itself. Watch.

AB NDP social media post

I. The demand that dared not come true

What does a political movement do when its demand comes true? Naheed Nenshi delivered the demand personally. As mayor, he flew to Toronto’s Collision conference in 2019 to pitch Calgary to the industry: “If we actually get the next great big app in Calgary, we’d be thrilled,” he said. “But we are actually looking at tech, AI and machine learning to enable other sectors in the economy.”

He advertised the city’s “tons of engineers, tons of data scientists.” A collection of people in what is now his party matched him. Under Rachel Notley’s name, the Alberta NDP caucus declared that “Alberta’s tech and AI sectors have enormous potential to create jobs and grow our economy” and demanded a stable funding environment to “further diversify Alberta’s economy.” The party even tabled legislation for a public technology venture fund, lamenting that Ontario drew $7.4 billion in venture investment in a single year while Alberta managed only $561 million.

Then the investment came. On July 8 this year, Meta committed more than $13 billion to a one-gigawatt campus in Sturgeon County, its first Canadian data centre and its largest anywhere outside the United States. It comes paired with a $4.6 billion privately financed power plant. Premier Danielle Smith stated the terms in four short sentences: “No subsidies. No grants. No discounted power. No taxpayer backstop.” No public money for it. The project promises roughly 3,000 construction jobs, 300 permanent positions and about $250 million a year in royalties, taxes, levies and fees.

It is clear to everyone that this project brings diversification in the exact sector progressives had specified, on terms no subsidy-hunting battery plant, gaming company or film studio has ever offered this country.

But the movement that demanded it, if you ask progressives, seems to have vanished. Nenshi now calls the file a “Wild West free-for-all” with “no regulations in place to protect Albertans.” His technology critic, Nathan Ip, describes a “concierge service for foreign billionaires.” They want Meta and all similar or related projects to pause, and eventually scrapped.

Let’s test the free-for-all Nenshi claim against the statute book first. The Alberta Electric System Operator capped large-load connections at 1,200 megawatts in June 2025 precisely because it refused to let data centres become a risk to grid reliability. Bill 8, the Utilities Statutes Amendment Act, 2025, built a tailor-made regulatory framework for the industry and makes every data centre pay for its own transmission upgrades rather than pushing the cost onto the province’s ratepayers.

The Data Centre Regulation, Alberta Regulation 117/2026, in force since June this year, prioritizes projects that bring their own generation ability and leaves the rest waiting. Bill 12, the Financial Statutes Amendment Act, 2025 (No. 2), imposed a levy on any facility drawing 75 megawatts or more from the grid.

Proposed power plants still face full Alberta Utilities Commission hearings; the Indus project spent this very August before the commission. Municipal permits, Water Act licences and Environmental Protection and Enhancement Act approvals all continue to apply. A politician may argue these instruments are miscalibrated. But Nenshi cannot credibly say they do not exist. His claims of a free-for-all fail miserably when they meet the work of the legislature, where he also now sits.

Nenshi has long sold himself as the professor, the policy wonk, the smartest man in the room. Yet the regulation he says does not exist is numbered, published and already in force. He sits in the Legislature whose laws created the framework. Either he has not read it, or he has read it and chooses to describe Alberta as having “no regulations in place.” Neither explanation flatters the wonk.

A harder question follows. Set the gas plant aside for a moment and ask what remains of the opposition. The official objection is environmental: data centres burn natural gas, and gas means emissions. If that were the true complaint, it would vanish the moment the fuel changed. So apply a simple test. Imagine every data hall in Alberta running on some emissions-free miracle, on unicorn flatulence if you like. Would the NDP stand down?

Their own town halls answer the question. Participants cited “the disruptive cultural effect of generative AI” as a reason to oppose the projects. Others objected to foreign control of data, or to the personalities of the billionaires involved. They never mentioned these when they championed the AI tech they wanted to bring. And notice that none of these grievances burns a single molecule of gas. The objections would survive the greenest power source on earth, which means the power source was never the true objection. The emissions argument is the respectable costume. Underneath it stands a plain dislike of the industry and of the people building it.

---

Haultain Research is a reader-supported publication. To receive new posts and support our work, consider becoming a free or paid subscriber. You may also support us by sharing our work with like-minded friends and family, or making a modest donation at www.haultain.org

---

II. The exodus that keeps not happening

A second inconsistency concerns political predictions of failure, even advanced by some conservatives. Through the spring, three choirs sang the same warning: Alberta’s political turbulence was strangling investment. Jason Kenney called the referendum question “kryptonite for investor confidence” and accused the government of “playing with fire,” warning that Quebec “has paid the price for the uncertainty created by separation for 50 years.” By May he had escalated: separation would be “the single most dramatic act of economic suicide in modern history.”

Nenshi wrote that “just the talk of separation is putting a freeze on investment, scaring businesses away,” and accused Premier Smith of “doing generations of economic harm to Alberta.” The Calgary Chamber of Commerce supplied some numbers: 28 per cent of surveyed businesses said separation talk was already affecting them, and a commissioned analysis by UofC economist Trevor Tombe modelled that Alberta could forgo $10 billion to $15 billion in investment in the first year if investment responded to a separation process as it did after Brexit.

Note the grammar of the warnings. Could forgo. Would be. A freeze that an unnamed someone is putting on. It is all done in the conditional, sometimes with great hyperbolic drama.

Now set the speculation beside the ledger. The referendum date, October 19, was public knowledge by early June. On July 8, five weeks later, Meta announced. A corporation whose planning horizon spans decades, whose lawyers even price the political risk of eating a bagel for breakfast, and which had never built in Canada before chose Alberta this year, with that question on the ballot. Its dedicated power plant enters service in 2030 and will run for decades past it. Every dollar of the commitment assumes Alberta remains a jurisdiction where property is secure, contracts hold and electricity flows, straight through the vote Kenney and the autonomy alarmists said would freeze capital in its tracks.

The private conversation matches the public ledger. When The Logic canvassed Alberta’s corporate heavyweights this summer, Ensign Energy chief executive Bob Geddes reported: “Not one single person said, ‘Oh, I’m worried about investing in Calgary, or in Advantage, or Alberta, because of the separatist thing.’” The S&P/TSX Capped Energy Index climbed more than 30 percent between late August 2025 and late February 2026, and the foreign capital that stayed away when PM Trudeau wanted to leave it all in the ground now has started to arrive in the oilpatch.

We can grant the alarmists their strongest ground. One hyperscale commitment cannot disprove hesitation at the margin, and Tombe’s figure describes a counterfactual nobody can observe. Fair enough. But the burden now runs the other way. The forecasters predicted a freeze but one of the largest private investments in Canadian history arrived in the middle of it. Surveys measure an ethereal anxiety. Final investment decisions measure judgment. A poll asks a business owner how the headlines make him feel; a long-term tolling agreement asks him to commit for decades. Between the Chamber’s forgone billions and Meta’s committed billions, only one is a fact.

Kenney governed this province for a bit. Nenshi wants to. Both told Albertans capital was fleeing while capital was signing in. When the announcement came, neither said a word. A man who diagnoses economic suicide owes the province a note when the patient walks out of the hospital alive.

One more plank of the alarm needs some scrutiny: the Quebec analogy itself. Yes, capital fled Montreal after 1976. Sun Life announced it would move its head office to Toronto in 1978, and dozens of firms followed. But those head offices fled a hostile language law that imposed significant burdens and liabilities, as much as they fled a referendum. They also fled a government: the Parti Québécois came to office as the most left-wing socialist party ever to govern a Laurentian province, and business was as genuinely concerned about its socialism as about its separatism. A head office, besides, is the most portable asset a province holds. Alberta’s wealth does not fit in a moving van. Nobody relocates an oilsands mine, a gas field or a petrochemical complex to Bay Street.

History elsewhere is just as unkind to Kenney’s analogy. Czechs and Slovaks parted in 1993 and both economies grew. Singapore left its federation in 1965 and became one of the richest places on earth. Norway left its union with Sweden in 1905, and nobody in Oslo mourns. Separation carries real costs and real risks, and this piece takes no position on the October question at the moment (it deserves a full discussion that would only sidetrack the argument here). My point is narrower: economic ruin follows bad policy, not the mere act of a people voting to govern itself.

III. Heads Alberta wins, tails Alberta keeps the turbines

The naysayers may hold one strong card. No one can guarantee business success. Artificial intelligence may be a bubble. Meta will spend between US$130 billion and US$145 billion on capital projects in 2026 alone, nearly double last year’s outlay; its free cash flow shrank below one billion dollars last quarter, and Wall Street has punished the stock, down roughly 12 per cent on the year and nearly 30 per cent from its peak. If model progress stalls, or the revenue never matches the compute, the hyperscalers will slash their spending. Alberta, the argument goes, will then inherit ghost warehouses on 1,750 acres. A monument to a gold rush that ended the way gold rushes end.

So suppose it all happens. Now inventory what Alberta holds on the morning after.

A data centre is two assets wearing one name. The first is the computing hardware: servers that depreciate in three to five years, belong to Meta, and become Meta’s write-down in any bust. The second is the power plant, which endures. Greenlight will deliver 932 megawatts of modern combined-cycle generation, expandable to 1,864 megawatts, built by Pembina, Morgan Stanley Infrastructure Partners and Kineticor for roughly $4.6 billion of private money, under fixed-price contracts, with Siemens equipment reserved years in advance. Combined-cycle plants run for roughly forty years. Servers come and go. Turbines remain.

Recall January 2024. Emergency alerts lit up phones across the province on a brutally cold night because the grid stood within a rounding error of rolling blackouts. The system lacked one thing: power plants that run on demand, in any weather. Alberta’s rules now make the data centres solve that problem at their own expense. Before a large project can jump the connection queue, it must bring its own generation that covers everything it consumes. The rule works as a priority system rather than a ban, but the grid’s spare room is already spoken for, so self-supply is in practice the only door in. Every campus that walks through it adds more reliable power to Alberta than it takes out.

The AESO is also designing contracts that let it pause a data centre for an hour on the coldest evening of the year, and here the computers hold an advantage no heavy industry can match. Pause a smelter, and the metal freezes in the pots. Pause a training run and the machines simply wait. The bottom line for anyone paying a power bill is worth repeating: each data centre arrives carrying its own power plant, the province gains the right to interrupt the computers in an emergency, and the lights in Alberta homes sit more securely than they did before the industry came. That is a net benefit in potentia.

So run both branches of the future. If AI succeeds, Alberta hosts a new industry, collects its levies and taxes, sells its gas into a domestic market and pockets transmission relief. If AI busts, the shareholders of Meta, Pembina and Morgan Stanley absorb the losses, exactly as the premier framed it: “If it fails, the investors lose the money, not you.” And Alberta’s market keeps a modern generating fleet the public purse never paid for, wired into a grid that nearly failed without those very machines. Name another industrial policy in this country’s history whose failure mode is surplus electricity. I can’t think of one.

The critics demanded diversification and, when it arrived, called it a menace. They predicted flight and watched one of the largest arrivals in the country’s history. They warned of a bust whose principal legacy would be cheap, abundant, dispatchable power. Three objections, three self-refutations. The debate Alberta deserves begins where these end.

One question remains for Nenshi and the progressives: What happened?

---