Headline megawatts describe ambition. Rank describes evidence. The DC Intel tracker as of 11 September 2026 lists 33 Alberta data-centre projects. Six are Highly Credible, seven Probable, twelve Early-Stage, and eight Speculative. That distribution is the 2026 pipeline.

It is not a list of facilities that will be built.

What the ranks actually mean

Scores use the scoring method: sponsor quality, site and approvals, power certainty, commercial commitment, and execution evidence. Power is the heaviest weight, at 25%. Inactive projects never sit in Highly Credible; the inactive override puts them in Speculative.

Highly Credible (75-100) is for independent delivery signals. In this snapshot, that set is:

Probable (55-74) still has at least one major gate open. Keephills leads that band on a signed AESO Phase I load contract, and TransAlta still says no specific data centre is yet being built. Beacon, Teton, Diode Ventures and Mihta Askiy sit in the same band as proposed campuses.

Early-Stage names have real activity. Requested MW is not committed capacity. Speculative names have thin public evidence, or are inactive.

Status is the other axis

Five projects are under construction. One is approved. Twenty-four are proposed. Three are inactive. The Highly Credible set and the construction/approved set are the same six names. That is a fact about this snapshot. It is not a rule that every future Highly Credible name will already be in the ground.

The three inactive names are cautionary, not a rounding error. Kineticor’s Rocky View concept had meaningful power work; the county rejected the required area structure plan in September 2025. Crusoe’s Crossfield / Lone Pine proposal was withdrawn and related AESO requests were cancelled. Synapse Olds had a 1 GW campus concept; the associated power plant was denied by the AUC in August 2026 and the proponent confirmed it would not proceed as proposed.

Do not add the megawatts

Announced capacity across the 31 projects with a published figure is 17,665.8 MW. Two projects have no published MW. Those figures mix facility power, computing load, on-site generation and connection requests. Methodology is explicit: do not add MW across rows. A request is not an allocation.

Some Highly Credible scores already refuse to treat the headline campus as the ranked object. AHI is ranked on Phase 1. K-Tech is ranked on a 5 MW proof-of-concept. Prairie Sky’s 50 MW is a staged roadmap. Treating those rows as if the full announced stack were committed would misread the tracker.

How to use the ranking

If the question is what is most likely to be built next, start with Highly Credible, then read projects most likely to proceed. If the question is whether the power queue can clear, read how much of the power queue is likely to be built and why power will decide the winners.

This article does not forecast a count of final investment decisions, a year when the proposed stack converts, or a single Alberta MW total that will reach commercial operation.