The Answer Is in the Parenthesis

Commentary on "Is it worthwhile courting data centres?" · Winnipeg Free Press

An editorial board asking whether an industry is worth having is doing something useful, and the question in the Free Press headline is the right one. The evidence assembled underneath it is almost entirely American, and the piece contains, in brackets, the reason that matters.

The editorial catalogues four failures. Lake Tahoe, where a utility serving 49,000 people learned that the supplier providing three quarters of its electricity will stop next May, because data centres east of Reno want the power instead, and where the supplier reports being approached for 22,000 megawatts, roughly 150 times what Tahoe consumes in a year.

Fayette County, Georgia, where a water utility discovered two industrial-scale connections to a data centre it had no record of, unbilled, delivering more than 30 million gallons while other customers were under drought conservation orders. A proposed Utah project across 40,000 acres that would consume more power than the entire state does now, self-supplied by natural gas generation. And American residential rates up 9.5 per cent year over year, to a national average that converts to nearly 24 cents Canadian.

Against that, Manitoba households pay 9.97 cents.

The editorial notes the gap and moves on. It is the whole answer.

The parenthesis is the argument

Midway through, the piece observes in brackets that under Canadian public utility regulation this is far less of a concern in most provinces, Manitoba included, before returning to the American material. That bracketed sentence is not an aside. It is the finding.

Work through what actually happened in Lake Tahoe. A merchant generator with no obligation to serve a particular community reallocated supply from a distribution utility to industrial customers willing to pay more. That transaction requires a deregulated wholesale market, a supplier free to decline renewal, and no regulator standing between the two. Manitoba Hydro is a Crown corporation with a statutory obligation to serve, selling at rates set by the Public Utilities Board. There is no counterparty that can outbid a Manitoban for their own electricity, because the structure that would permit it does not exist.

The Georgia case is not a data centre failure either. A utility did not know about two industrial connections on its own system and did not bill for what went through them. Any large industrial customer could have done that, and the fact that this one did says more about the utility's asset records and metering practice than about server halls. The Utah project describes a jurisdiction willing to permit nine gigawatts of behind-the-meter gas generation without meaningful emissions accounting.

Four cases, four different regulatory failures, none of which is a property of the technology. Importing the conclusion without importing the conditions produces the wrong policy. The correct lesson from the American experience is that market structure determines who absorbs the cost of new load, and that Canada's structure is different in ways that are load-bearing.

Manitoba has already asked the better question

Here is what makes the editorial's framing frustrating. Manitoba has already produced the most sophisticated answer in the country, and the piece does not mention it.

When the province turned down the proposed facility at Île-des-Chênes, the stated reason was not water and not power. It was that the employment created did not justify the resources the project would consume. That is a benefits test, and it is a different instrument entirely from a moratorium or an environmental objection. It concedes that the province could supply the project and declines anyway, because the trade was poor.

No other Canadian jurisdiction has held that line. Alberta courts the investment. Ontario promises guardrails it has not written. British Columbia's largest city defeated a disclosure requirement and then could not run its own hearing. Manitoba looked at a specific proposal, priced what it would give up, and said no. It also told CBC it has no intention of imposing a moratorium, which is the correct pairing. A benefits test needs the door open to work at all.

This reframes the headline's question. Cheap hydroelectric power is not a reason to build in Manitoba. It is Manitoba's competitive advantage, held on behalf of its households and its existing industry, and selling it at industrial rates to a facility employing a few dozen people is a bad trade regardless of how clean the electricity is. The question was never whether to court these projects. It is what a province should require in exchange, and Manitoba is the only place asking it properly.

Two arguments wearing the same clothes

The final paragraph changes subject without marking the change. It argues that firms increase profits by replacing workers with AI, raise everyone's costs, damage collective health, and eliminate the incomes of their own future customers, and asks whether any of it makes economic sense.

That is an argument about artificial intelligence and labour. It is a serious argument and people are entitled to make it. It is not an argument about infrastructure siting, utility regulation, or water metering, which is what the preceding 800 words were about.

This conflation is now the defining feature of the debate. A placard at a protest in Vancouver made the same move. So does an editorial board. The physical building has become the available proxy for a set of anxieties about the technology itself, and because the building is the thing that requires a permit, that is where the argument lands. Operators who respond only to the infrastructure claims will keep finding that answering them changes nothing, and critics who lead with infrastructure claims will keep finding their real objection unaddressed.

The editorial calls the buildout a boom that "outstrips fiscal logic." On the American evidence it presents, that is a fair description. On Manitoba's own evidence, the province has been applying fiscal logic more rigorously than anyone else in the country, and got a defensible answer by doing so.

VOLTEDGE

Reference: "Is it worthwhile courting data centres?" · Winnipeg Free Press · read the article