A source-backed research brief and dataset: grid emissions, the 2026 large-load connection process, and the interconnection bottleneck across five provinces — alongside the playable model of the same trade-offs.
On paper, several provinces plainly have enough electricity for AI. The finding is that in every one of the five examined, a new large compute load in 2026 has to get through a rationing mechanism — a selection process (Québec), a competitive call (British Columbia), an interim allocation block already fully committed (Alberta), or a system study with no published duration (Ontario). Manitoba has the most near-term room, and a forecast capacity shortfall around 2029/30. How clean or how large the provincial grid is does not change this.
Between a province's installed generation and an energized compute load sit five steps, in series. A weakness at any one of them stops the project.
This release has primary-source data for the first three at the province level. Land and fibre matter and are named here, but standardized Canadian data for them was not sourced for this release. The structural point: a province being rich in clean generation tells you almost nothing about whether the interconnection step is open.
Grid GHG intensity is observed (ECCC 2026 consumption-intensity factors, Table 5.3). "Access" is this project's categorical assessment against a published three-part rubric (rationing mechanism, interim-block status, formal study process) — not a queue-time measurement and deliberately not a 0–100 score. The connection description is quoted from each system operator.
| Region | Grid GHG gCO₂e/kWh | Access | Large-load connection in 2026 | Primary bottleneck |
|---|---|---|---|---|
| QuebecHydro-Quebec | 1.9 | HIGH | Data centres above 5 MW are subject to a selection process; Hydro-Quebec has proposed a dedicated large-data-centre rate near 13 cents/kWh, roughly double current large-power customers. | Administrative access: connection is allocated via a selection process, not available on demand. |
| OntarioIESO | 59 | MODERATE-HIGH | IESO is developing specific technical and reliability requirements for large computational loads; connection feasibility is determined by a system study and is location-specific. | Location-specific grid capacity and a system-study process with no published timeline. |
| AlbertaAESO | 438 | MODERATE-HIGH | AESO operates an interim Phase-1 large-load framework with a 1,200 MW connection block (970 MW Genesee/GLDC + 230 MW Keephills); the block is fully allocated and further large loads move to a subsequent connection phase. | The interim large-load allocation is exhausted; Alberta also has the highest grid emissions of the five regions. |
| British ColumbiaBC Hydro | 18 | MODERATE | BC Hydro connects emerging-industry large loads through a competitive process; its 2026 Call for Demand received 15 applications representing roughly 800 MW. | Connections are rationed through a competitive call cycle and demand exceeds available awards. |
| ManitobaManitoba Hydro | 2.5 | MODERATE | Connection capability is assessed project by project; Manitoba Hydro's Integrated Resource Plan context projects that peak demand will exceed capacity around 2029/30 and identifies roughly 1,760 MW of additional capacity needed by 2035. | A provincial capacity shortfall is forecast for around 2029/30 - near-term room but a closing window. |
Full rows, with the scenario timeline field and per-value source ids, are in canada_compute_regions.csv. Every value maps to a source in source_manifest.csv; claims.csv tags each headline statement SOURCE / DERIVED / ASSESSMENT.
If a region needs new generation to host compute, its cost and speed become the constraint. Overnight capex and lifecycle GHG below; capacity factor is a 2025 U.S. EIA empirical proxy. Capex per average-MW delivered = overnight capex ÷ capacity factor — it normalizes for how much of the year a plant runs, but not for firmness.
| Technology | Overnight capex $/kW | Capacity factor | Lifecycle GHG gCO₂e/kWh | Capex per avg-MW delivered (derived) |
|---|---|---|---|---|
| Natural gas combined cycle firm | 1,450 | 58% | 423 | 2,483 |
| Utility-scale solar PV variable | 1,700 | 24% | 48 | 6,967 |
| Onshore wind variable | 1,900 | 34% | 11 | 5,556 |
| Natural gas combined cycle + CCS firm | 3,705 | 58% | 106 | 6,344 |
| Large hydro (>100 MW) firm | 4,500 | 35% | 24 | 12,748 |
| Geothermal firm | 7,100 | 66% | 38 | 10,774 |
| Small modular reactor firm | 9,262 | 91% | 12 | 10,178 |
Overnight capex and fixed O&M: CER Canada Energy Future 2023, Electricity Supply Model, Table E.1, verified exact. Full table: technology_costs.csv.
Research release 2026-09 (v1.0). The brief, the dataset, and the live model read the same sourced values. The dataset is regenerated from a documented pipeline and passes a schema / source-id / rubric validation gate with zero warnings.