Ontario’s Electricity Sector Carries the EPS, and Cannot Create a Credit

At a Glance — Electricity under the Ontario EPS

Electricity was 999,224 tonnes short in the 2022 compliance period, 1,736,266 tonnes in 2023 and 2,167,436 tonnes in 2024. That is the largest position of any sector in every year the programme has published, and its share of provincial compliance demand rose from 32.2% to 44.9% across the three years.

Over the same period the sector issued no emissions performance units at all. Not a reduced number — zero, in 2022, 2023 and 2024. Every other covered sector has created at least some credits; cement alone supplied 335,616 tonnes in 2024. Electricity is the one activity in the Ontario EPS that has only ever been a buyer.

The two facts are connected, and the connection is structural rather than behavioural. Understanding it requires looking at how the sector’s limit is set.

Fossil electricity generation is the one activity in the EPS where the limit is stated directly as an output benchmark. Method B sets an emissions limit of 310 tonnes of CO₂e per gigawatt hour of electricity produced, and section 4.2(c) of the TAEL Methodology fixes the stringency factor at 1.0 in every compliance period. No glide, no annual tightening, no phase-down.

The consequence is that a facility’s annual limit equals 310 multiplied by its generation. The compliance position is therefore the gap between the fleet’s actual emissions intensity and the benchmark, multiplied by output:

Ontario’s covered fossil fleet ran at roughly 391 tonnes per gigawatt hour in 2024, against a benchmark of 310. On approximately 26.8 terawatt hours of covered generation, that 81-tonne gap produced the 2.17 Mt shortfall. The arithmetic reproduces the published figure directly.

Two things follow. First, tightening elsewhere in the programme does not reach electricity — the stringency factor that squeezes cement and steel is frozen at 1.0 here. Second, the sector cannot create a credit until the fleet operates below 310 tonnes per gigawatt hour. A modern combined-cycle plant can approach that figure; a simple-cycle peaker run for a few hundred hours a year cannot come close. Ontario runs both, and the fleet average is what the benchmark meets.

The sector’s position tripled between 2022 and 2024 without any change in how the plants operate. Output rose, because the gas fleet was covering for nuclear units that were out for refurbishment.

Across the 23 electricity facilities that reported in every year, emissions went from 4.43 Mt in 2019 to 8.81 Mt in 2024. Greenfield rose from 0.94 to 1.72 Mt, Halton Hills from 0.74 to 1.21 Mt and Goreway from 0.54 to 1.17 Mt. None of these is a new plant. They are the same units, run harder.

Statistics Canada’s generation series shows the same movement on the system side. Ontario non-renewable combustible generation went 17.49 TWh in 2022, 21.37 in 2023, 27.45 in 2024 and 34.26 TWh in 2025 — a 24.8% increase in a single year, against total provincial generation growth of 3.3%. Gas took share. Nuclear output fell from 90.45 TWh in 2019 to 78.28 TWh in 2025 while total generation rose 13.4 TWh, and gas absorbed both movements. Its share of the province went from 7.7% in 2020 to 20.2% in 2025.

The refurbishment programme is now turning. Darlington’s four-unit refurbishment finished in February 2026, four months ahead of schedule and $150 million under budget. Pickering units 5–8 came off in September 2026 for a refurbishment expected to extend their life by up to 30 years, with the execution phase starting in early 2027. The gas fleet covers that gap the same way it covered Darlington’s.

The IESO’s 2026 Annual Planning Outlook puts natural gas generation at 38.75 TWh in 2027, rising to a peak of 44.08 TWh in 2031 while Pickering B is out, then falling by half to 22.11 TWh by 2035 as the units return and procured capacity arrives.

Because Method B holds intensity constant by construction, that path maps directly onto the compliance position. Our facility-level model has the electricity shortfall peaking at 3.44 Mt in 2031 and easing to 1.73 Mt by 2035. The sector’s share of provincial compliance demand follows: 47.4% in 2030, falling to 25.7% by 2035. Electricity stops being the dominant obligation only when the reactors come back.

That forecast carries a specific and well-documented uncertainty. The IESO’s own view of 2031 gas output has moved across four consecutive vintages — 19.8 TWh in the 2022 outlook, 46.73 TWh in 2024, 55.9 TWh in 2025 and 44.08 TWh in the 2026 edition. The same year has spanned a 182% range in four publications. The near-term record runs one way: the 2022 outlook put 2024 gas at 17.9 TWh against 27.45 TWh actually generated, and 2025 at 20.3 TWh against 34.26 TWh. Part of that gap is coverage, because Statistics Canada counts embedded and industrial generation the IESO adequacy tables do not. The rest is forecast error, and it has been in one direction.

Every terawatt hour of that error is worth roughly 81,000 tonnes of additional provincial obligation.

The federal Clean Electricity Regulations (SOR/2024-263) govern whether and how long fossil units may run. They are separate from the EPS, which governs what a unit pays when it does run. The distinction matters for how the two interact.

Under the Canada–Alberta Memorandum of Understanding signed on 27 November 2025, Canada committed to “suspend immediately the Clean Electricity Regulations (CER) in Alberta pending a new carbon pricing agreement, which includes the electricity sector, administered through Alberta’s TIER programme,” and, once that agreement is complete, to place the CER in Alberta in abeyance. That suspension is Alberta-specific. It does not lift the CER in Ontario.

What does reach Ontario is what came next. On 14 May 2026 the federal government released a national electricity strategy that identified a strategic role for natural gas generation and proposed to adjust the CER to allow more flexibility, including offsetting emissions elsewhere and adding new units in the near term, followed by a four-month consultation. Ontario has argued the regulations as drafted would add $35 billion in system costs by 2050, on an IESO analysis, and has been adding gas capacity to hold the system together through the refurbishment period.

For an Ontario EPS participant the direction of that change is worth stating plainly. A weakened CER removes a constraint on how much the gas fleet may generate. Under Method B, more generation is more obligation, not less. Relief on the operating side of the rule book increases the compliance bill on the pricing side. The two instruments push in opposite directions, and only one of them is being loosened.

Method B leaves exactly two levers. Output, and intensity. Nothing else in the sector’s EPS position responds to anything.

Output is being addressed, and not for carbon reasons. Darlington is back. Pickering B returns through the early 2030s. Ontario is procuring storage and has committed to new nuclear at Darlington and is examining a large-scale expansion at Bruce. Each of those displaces gas generation and reduces the obligation mechanically. This is the whole of the improvement between 2031 and 2035 in our forecast, and it is the only lever with committed projects behind it.

Intensity is the lever nobody is pulling. The covered fleet would need to fall from roughly 391 to below 310 tonnes per gigawatt hour — a 21% improvement — before electricity generates a single credit. Carbon capture on gas generation, hydrogen co-firing, or a structural shift from peaking to combined-cycle operation would each move it. No announced Ontario project does so at fleet scale, and our forecast holds intensity flat at the 2024 level for that reason rather than as a judgement about what is technically possible.

That assumption is where the forecast is most exposed. If a meaningful share of the fleet were to move below the benchmark, electricity would flip from the programme’s largest buyer to a credit source, and the provincial balance would change character rather than degree. Nothing currently under construction does that before 2035.

Electricity carries the Ontario EPS because a fixed 310-tonne benchmark meets a fleet running near 391, and because the province has been leaning on that fleet to cover a nuclear refurbishment programme. Neither condition is a function of how well the sector is managed, and neither responds to the stringency schedule that tightens the rest of the programme.

The position improves when the reactors come back, and on the IESO’s own path that is a 2032–35 story, not a 2027 one. Until then the sector sets the provincial shortfall, absorbs the EEU charge at whatever rate applies, and returns nothing to the credit market. Participants sizing an EPS exposure through the early 2030s should treat electricity output as the single most important input, and should treat the IESO gas path as a forecast with a demonstrated downward bias in its near years.