How Flexibility, Not Nuclear, Can Secure Ontario’s Electricity Future

Support CleanTechnica’s work through a Substack subscription or on Stripe.

Or support our Kickstarter campaign!

Ontario is moving forward with planning for an entirely new nuclear generation site in Port Hope, 100 km east of Toronto, at a moment when its electricity system is already one of the most nuclear-heavy in the world. Nuclear power today provides roughly 55% of Ontario’s electricity, with hydro adding another 25%. Wind, solar, batteries, and demand-side resources together account for a much smaller share, having been cut off at the knees in 2018 when the provincial conservative party took power and summarily cut 758 contracts for renewable generation. Advancing a new site signals how the province understands its future electricity challenge. It reflects an expectation that Ontario will require another large block of firm, always-available capacity to remain reliable as demand grows, particularly during the most constrained hours of the year.

Ontario’s electricity planners, primarily through the Independent Electricity System Operator, frame the case for new nuclear around long-term reliability rather than annual energy supply. Their planning outlook projects electricity demand rising by about 65–75% by 2050—a low energy value not aligned with actual climate or competitiveness goals—with a projected winter peak reaching roughly 36–37 GW. Summer peaks are also expected to rise, but they remain slightly lower, in the range of about 35–36 GW by mid-century. The winter peak, not the summer peak, is treated as the binding constraint, and it is that single cold, dark evening hour that underpins the justification for new nuclear capacity.

This framing matters because of how nuclear is treated in planning models. Nuclear plants supply energy year-round, but the decision to build new nuclear capacity is driven mainly by how much firm capacity planners believe is needed to meet future peak demand. Nuclear units are counted as fully available during peak hours, even though they operate continuously, do not follow demand and are not available when down for maintenance, refueling or refurbishment for months or years. From a reliability perspective, this approach is understandable. System operators are rewarded for avoiding shortages and penalized heavily for blackouts, while overbuilding capacity carries fewer immediate consequences.

The IESO forecast assumes substantial electrification of transport and some electrification of buildings, with electric vehicle charging energy growing to roughly 40 TWh per year by mid-century and millions of EVs connected to the grid. It also assumes that a meaningful portion of this new load remains coincident with existing peaks, particularly during winter evenings when heating demand is high. Demand response, smart appliances, and storage are included in the modeling, but they are treated as supporting measures rather than as structural features that fundamentally reshape the load curve.

The distinction between energy growth and peak growth is critical here. Energy demand, measured in TWh, reflects how much electricity the system produces over a year. Peak demand, measured in GW, reflects the single hardest hour the system must meet. Nuclear plants are not built to follow peaks, but they are sized to peaks. If peaks remain sharp and high, nuclear looks attractive in planning models. If peaks flatten or decline due to significant system component flexiblity, the value of adding large, inflexible, always-on generation falls quickly, even if total energy demand continues to rise.

Electrification without flexibility is genuinely concerning, and planners are right to worry about it. A maximally electrified Ontario with unmanaged EV charging, heat pumps running flat out during cold evenings, and limited storage would see peaks well above today’s levels. Simple stacking shows how this happens. Start with a base load of around 25 GW, add roughly 15 GW of electrified space and water heating at a winter peak, and add 5 GW of coincident EV charging as drivers plug in after work. The result is a peak around 45 GW. In that world, the winter peak is dramatically higher than the summer peak, and new nuclear capacity looks necessary because planners must size the system to cover that extreme hour.

That stress case, however, is not an argument against electrification. It is an argument against unmanaged electrification. Electrification done well looks very different. Electric vehicles are digital devices with batteries attached. Charging behavior is software-defined and responsive to price signals. Today, more than 80% of EV charging can be shifted in time with no impact on mobility, and by 2050 unmanaged charging would represent a policy failure rather than a technical constraint. Smart charging, time-of-use rates, and fleet scheduling all move load away from peak hours and into overnight or midday periods.

The same logic applies to appliances and buildings. Heat pumps, water heaters, and commercial HVAC systems are increasingly grid-interactive. Buildings have thermal mass that allows heating to be shifted by hours without affecting comfort. Hot water tanks act as simple thermal batteries. In a digitized system, these loads respond automatically to price and control signals. Treating them as static demand in 2050 planning implicitly assumes that the electricity system will fail to adopt the same digital control capabilities that are already standard in other sectors.

The reason winter dominates the planning outlook is heating. Winter peak demand in Ontario is primarily a heating problem, not an electricity problem. Space and water heating dominate cold-day peaks, while summer peaks driven by air conditioning are already more elastic and easier to manage. Cooling demand aligns better with solar output, can be shifted through pre-cooling, and is readily served by batteries. That is why summer peaks, even as they grow, do not drive the case for new nuclear in the same way.

Seasonal thermal storage and district energy change the winter picture directly. Aquifer thermal energy storage and other seasonal heat storage options allow heat to be generated when electricity is cheap and stored for later use. District systems using large heat pumps operate at higher coefficients of performance than individual air-source units, often closer to 4 rather than 2–3 during cold periods. Even modest deployment has material effects. If 25% of winter peak heating demand were served by district energy and 60% of that were supplied from seasonal storage, roughly 9 GWth of heat would be delivered during peak hours without drawing electricity at that time. At a peak COP of 2.5, that reduces electric peak demand by about 3.5–4 GW. That reduction applies precisely to the winter hours that justify new nuclear capacity.

Batteries then act on what remains. Grid-scale batteries and behind-the-meter batteries respond automatically to price spreads, charging when electricity is abundant and discharging during peak hours. They flatten load curves as a natural consequence of market behavior. A few gigawatts of batteries are enough to shave residual peaks once heating and EV charging have already been shifted. In this role, batteries are not backup generation. They are load-shaping infrastructure that reduces the height of the peak planners are trying to insure against.

Renewables complete the picture. Wind and solar are modular, fast to deploy, and continue to fall in cost. Overbuilding renewables combined with storage is now cheaper than building generation sized to meet rare peak hours. A system optimized around renewables prefers flexible demand because inflexible baseload forces curtailment during periods of high production and creates operational challenges. Adding more nuclear into an already nuclear-heavy system increases those challenges rather than resolving them, because nuclear output cannot easily adjust to changing system conditions.

When these elements are combined, the seasonal distinction that drives the planning narrative erodes. The IESO planning case projects a 2050 winter peak around 36–37 GW and a slightly lower summer peak. A maximally electrified Ontario—the realistic end state, not the IESO scenario—with conservative IESO assumptions about flexibility produces a winter peak around 45 GW, reinforcing fears about reliability. A maximally electrified, cost-optimized Ontario using smart charging, seasonal thermal storage, batteries, and flexible demand produces a peak closer to 33–34 GW. That peak is lower than the IESO planning winter and summer peaks in a much less electrified, less likely economy and closer to what planners already consider manageable.

The difference is visible when hourly load curves are compared. The current system’s winter peak day is relatively flat, with a trough around 17.7 GW and a peak around 21.9 GW, a swing of about 4.2 GW. Summer peak days are more pronounced, with larger trough-to-crest swings, yet they are already managed without nuclear expansion. The IESO 2050 winter planning curve shows a sharper evening spike lasting two to three hours. The optimized electrified curve shows a broader, flatter plateau with higher overnight load due to EV charging and thermal storage charging, but a lower evening crest. This matters because new nuclear capacity is sized to that crest, even though the plants themselves run continuously regardless of whether the system needs the energy at that moment.

At this point in the argument, it is useful to show how today’s energy services translate into electricity under full electrification, something I explored for Ontario recently. A maximally electrified Ontario energy system, with fossil fuel inputs removed, primary energy reduced through efficiency, and electricity supplying transport, heating, and industry is a much more efficient and flexible Ontario energy system. The diagram makes clear that electrification collapses wasted energy while increasing electricity’s role, and it provides context for why energy growth does not automatically imply peak growth.

None of this suggests that nuclear power has no role in Ontario. Refurbishing existing nuclear plants makes sense. What it does suggest is that adding another 10 GW-scale nuclear site is a high-risk response to a winter peak reliability problem that becomes much smaller once heating is treated as flexible rather than fixed. Nuclear projects have long lead times, high capital costs, and limited ability to adapt once built. They crowd out faster investments that reduce winter peaks directly, which is what planners are actually trying to insure against.

The planning incentives faced by system operators help explain the gap. Conservative assumptions protect against worst-case outcomes but do not describe a cost-optimized future. A planning framework that assumes flexibility underperforms will naturally overbuild firm generation. A planning framework that assumes flexibility succeeds, as digital control systems already demonstrate in practice, produces a very different investment pathway and sharply reduces the need for new firm capacity.

A more coherent hierarchy for Ontario’s electricity future follows directly from this analysis. Electrify aggressively to reduce overall energy demand. Make flexibility the default through pricing, controls, and aggregation. Build renewables and batteries at scale. Preserve and refurbish existing nuclear capacity. Only then assess whether additional firm generation is required. When this sequence is followed, the winter peak that justifies new nuclear begins to look more like a summer peak, and the case for a large new nuclear site weakens from necessity to optional insurance at very high cost.

Ontario does not lack clean electricity. It lacks a planning framework that fully reflects how electricity systems are changing, why winter peaks appear hard only under outdated assumptions, and how firm capacity is actually used in a flexible, digitized grid. The choice facing the province is not between reliability and decarbonization, but between building infrastructure sized for a winter peak that no longer needs to exist and building a system designed to avoid creating that peak in the first place.