Ontario Solved the Wrong Half

Commentary on "Data centre regulation in Ontario: Grid access rules" · Borden Ladner Gervais

Three months ago the complaint across Ontario municipalities was that no technical standard existed for connecting very large computational loads, and that councils were therefore improvising with zoning bylaws. The standard now exists in draft. It is a serious piece of work, it addresses the reliability risk the North American Electric Reliability Corporation flagged, and it does not touch the thing the public is actually angry about.

BLG's summary of Ontario's grid access regime sets out the machinery. On May 14 the system operator posted draft technical requirements for large computational loads, consolidating existing load requirements and adding new provisions aimed specifically at the behaviours that make these facilities different. Applicants must now supply a project development plan, primary equipment data, load composition, and voltage and frequency operating ranges. A Major Projects Identification Committee has been established to flag large new demand early. A public engagement session was held on July 23. The firm's assessment is that compliance and disclosure obligations rise materially.

That is the connection performance standard the municipal argument has been calling for. It arrived with no press coverage at all.

What the draft actually fixes

The requirements are aimed at reliability, and on that they are responsive. The behaviours in question are steep ramp rates, load intermittency, and protective equipment that disconnects a facility within milliseconds of a disturbance. Requiring an applicant to declare its operating ranges and load composition before connection is exactly how a system operator prevents a cluster of facilities from tripping in unison. Alberta's operator has done the same, publishing its own connection guide for transmission-connected data centres in June, explicitly ahead of formal continental standards.

Two provinces have now written the rule. Neither of them wrote it because a council asked. They wrote it because engineers did.

The cost question is a market design question

Here is why the standard will not settle the public argument. Ontario allocates a large share of system fixed costs through the global adjustment, and for large industrial customers that allocation is determined by their share of demand during the five highest-demand hours of the year. Predict those five hours, reduce consumption through them, and your share of the fixed costs falls. It does not disappear. It moves to everyone who did not reduce.

There is an established consulting industry built around forecasting those hours. It exists because the incentive is legal, substantial, and working as designed.

Now consider what happens when facilities of this scale enter that system. A training cluster is one of the few very large loads that can, in principle, be curtailed for a few hours without damaging anything, because the work is deferrable in a way that a smelter's is not. An operator that shifts load out of those five hours is behaving exactly as the rule intends and reducing its contribution to costs that other ratepayers then carry.

This is the mechanism underneath the bill increases people have been protesting, and it has almost nothing to do with connection standards, water, or zoning. It is a cost allocation rule written for an industrial economy that did not include flexible gigawatt-scale loads. No technical requirement in the draft addresses it, because that is not what a technical requirement is for.

Something else follows from the same observation, and it points the other way. If these loads can shift out of peak hours, that flexibility is genuinely valuable to a system operator, and it is the kind of thing a jurisdiction could require or pay for rather than leave to each operator's own cost optimisation. A curtailment obligation is a connection term. It could be written. It has not been.

Discretion is not a framework

The remaining gap is the one this series has flagged before, and it has not closed. Legislation passed in December gives the province authority to decide which very large loads may connect on the basis of whether they serve Ontario's economic interests. The regulations defining the criteria, the community obligations, and the reporting requirements have still not been published. Consultation closed in November.

So Ontario now has a detailed technical standard governing how a facility must behave electrically, and no published standard governing whether it should be permitted at all. A developer can determine precisely what voltage ride-through it must demonstrate and cannot determine what economic contribution the province will require. That is an odd asymmetry, and it is the reverse of what municipalities need. Councils are not equipped to assess ride-through performance and are perfectly equipped to weigh local benefit. They have been handed the question they cannot answer and denied the one they can.

What this means going forward

The connection standard is a real advance and the sector should say so plainly rather than treat every new requirement as friction. Predictable technical rules are worth more to a serious developer than the absence of rules, because the absence is what produces moratoriums.

But nobody should expect it to quiet the argument. The public objection was never about frequency response. It is about who pays, what the community gets, and whether anyone asked. Ontario has answered a question the engineers were asking. The two questions the residents were asking are still sitting in an unpublished regulation.

VOLTEDGE

Reference: "Data centre regulation in Ontario: Grid access rules" · Borden Ladner Gervais · read the article