Pull up an industrial electricity bill in much of Africa or South Asia and the real cost of power is hidden—buried in diesel, in downtime, in damaged motors that never appear as a line item. North America is the opposite. Here the bill is candid. It tells an operator, in dollars and cents, exactly what poor power factor and a spiky load profile are costing—because demand and reactive power are printed on the invoice as their own charges.
That candour is the story. The United States and Canada run comparatively reliable grids, so the case for power quality is not led by keeping the lights on. It is led by the bill: a demand charge that can rival the energy charge for size, and a reactive-power penalty that activates the moment power factor drifts below a threshold the utility has written down. Both are recoverable. Both respond directly to correction at the switchboard.
A second force is now reshaping the picture. An AI and electrification demand surge—data centres above all—is tightening grid capacity across both countries for the first time in a generation. Where spare capacity was once assumed, it is now scarce and valuable. That turns the headroom freed inside an existing building from an engineering nicety into deferred capital. This article maps both threads—the billed cost of bad power, and the rising worth of released capacity—across the United States and Canada.
Section 01
The defining feature of a North American industrial electricity bill is that it is rarely a single number multiplied by kilowatt-hours. For a commercial or industrial account, the energy charge—the kWh consumed—is only part of the total. Sitting alongside it is the demand charge: a separate fee based on the highest rate of power draw, in kilowatts, recorded in any short interval during the billing period. One fifteen-minute spike can set the demand charge for the whole month.
This is where power factor stops being abstract. On many tariffs the billed demand is not simply the metered kW; it is the metered kW divided by the power factor—effectively billing the apparent power (kVA) the utility’s network actually has to carry. A site running at 0.80 power factor is billed as though it drew 25% more demand than its real load. On top of that, many utilities levy a direct reactive-power charge—typically in the range of US$2–8 per kVAR—whenever power factor falls below a stated threshold, commonly between 0.90 and 0.95. Poor power factor, in other words, is charged twice: once by inflating billed demand, and again as an explicit penalty.
There is no single national electricity regulator in either country. In the United States, rates and demand-charge structures are set utility by utility and state by state; in Canada, they are set province by province. The mechanism is consistent even where the numbers are not: where the bill separates demand and reactive power from energy, correcting power factor cuts the bill directly—independent of the energy rate, and without changing a single kilowatt-hour of production.
Section 02
Neither “the United States” nor “Canada” is a single power market. US commercial rates average roughly US$0.13/kWh nationally, but that average conceals a wide spread: around US$0.18–0.22 in California and much of the Northeast, against roughly US$0.08–0.10 across the Southeast and Texas. Canada’s range is just as wide, set by each province’s generation mix. The exhibit below summarises the picture across the markets in scope.
| Market | Commercial rate (indicative) | What drives the bill | Lead lever |
|---|---|---|---|
| US — California / Northeast | ~$0.18–0.22/kWh | High energy rate plus demand charges billed on kW÷PF | Demand + energy savings |
| US — national average | ~$0.13/kWh | Demand charges and kVAR penalties dominate the bill | Demand + reactive penalty |
| US — Southeast / Texas | ~$0.08–0.10/kWh | Low energy rate; demand charges and capacity still bite | Demand + capacity |
| Canada — Ontario | ~$0.10–0.15/kWh | Demand billed on greater of kW or 90% of kVA; Global Adjustment peaks | Reactive (kVA) + peak demand |
| Canada — Alberta | ~$0.10–0.12/kWh, volatile | Volatile market pricing; a fast-filling data-centre queue | Demand + capacity |
| Canada — Quebec / BC / Manitoba | ~$0.05–0.08/kWh (hydro) | Cheap hydro; the energy rate is a weak savings lever | Capacity + harmonics |
| Canada — Maritimes (NS) | ~$0.19+/kWh | An expensive grid where every billed kVA counts | Demand + reactive penalty |
The heaviest North American loads—aluminium smelting, oil refining, large pulp and paper mills—sit on medium- and high-voltage connections outside the scope of low-voltage power-quality equipment. The opportunity this article describes is the broad low-voltage base: light and medium manufacturing (much of it reshoring under the IRA and CHIPS Act), commercial real estate and offices, big-box and general retail, healthcare, colocation and edge data centres, cold chain and food & beverage logistics, hotels, university and K-12 campuses, water and wastewater, and airports. These are the sites where a switchboard-level system goes to work.
Two patterns cut across the table. First, in the higher-rate markets—California, the Northeast, Nova Scotia—the energy charge is large enough that demand-side savings and reactive correction carry real weight on their own. Second, in the cheap-hydro provinces—Quebec, British Columbia, Manitoba—the energy rate is a weak lever, so the value shifts to releasing capacity and meeting harmonic standards. Ontario sits in a category of its own, for reasons worth a closer look.
Section 03
Ontario offers the clearest illustration on the continent of power factor as a direct monthly charge. Many Ontario utilities bill demand on the greater of metered kW or 90% of metered kVA. The mechanism is deliberate: once a site’s power factor falls below about 0.90, the 90%-of-kVA figure overtakes the real-power figure, and the customer is billed on apparent power rather than the work actually done. Low power factor is not a hidden inefficiency in Ontario—it is a number that appears, every month, on the demand line of the bill.
Layered on top is the Global Adjustment, and for the largest consumers the Industrial Conservation Initiative—a mechanism under which a large share of annual transmission and generation cost is allocated according to a site’s demand during the province’s five highest peak hours of the year. A facility that can hold its load down during those coincident peaks pays dramatically less; one that spikes through them pays more. It is one of the strongest peak-demand signals anywhere in North America.
In Ontario, the same underlying behaviour—an uncorrected, peaky load—is penalised through two distinct channels: a demand charge billed on apparent power (kVA) whenever power factor sags, and a Global Adjustment allocation driven by demand during a handful of system peaks. Correcting power factor and shaving peak draw therefore work against both at once. The physics is ordinary; what makes Ontario instructive is how plainly the tariff puts a price on it.
The lesson generalises beyond Ontario. Wherever a tariff bills on kVA, applies a reactive-energy penalty, or allocates cost by coincident peak, the bad-power problem has already been quantified by the utility and printed on the invoice. The operator’s task is not to discover a hidden cost—it is to remove one that is already itemised.
Section 04
For two decades, US electricity demand was essentially flat. That era has ended. The combination of AI and data-centre build-out, manufacturing reshoring, and the electrification of transport and heating has put load growth back on the system at a pace utilities have not planned around in a generation. The US Energy Information Administration projects national electricity demand to rise on the order of 17% by 2030—a step change after years of stagnation.
The clearest market signal is in capacity prices. In the PJM Interconnection—the operator coordinating the grid across the Mid-Atlantic and parts of the Midwest—the capacity-auction clearing price rose roughly ten-fold over two years, as the value of firm, available capacity caught up with how scarce it had suddenly become. Canada is seeing the same pressure concentrated regionally: Alberta alone is reported to have more than 10 GW of data-centre projects in its interconnection queue—a figure that dwarfs the province’s incremental supply.
“For twenty years, spare grid capacity in North America was assumed. The AI demand surge has made it scarce—and scarce capacity is valuable capacity. The headroom inside an existing building is now worth releasing.”
This reframes the second value lever. Correcting power factor and cleaning harmonics typically frees 15–20% of headroom on an existing transformer and switchgear—capacity that was always there, locked up by reactive current and distortion rather than useful load. In a flat-demand world that headroom was a curiosity. In a capacity-constrained one, it is deferred capital: a way for a growing site to add load, a new production line, or an EV-charging bank without triggering a service upgrade and the months-long interconnection wait that now comes with it.
Section 05
The same loads driving the bill are also degrading the waveform. Variable-speed drives, electronic power supplies, LED lighting, EV chargers, and the inverters behind on-site solar and storage are all non-linear—they draw current in pulses that distort the clean 60 Hz sine wave the network expects. The result is harmonic distortion: extra heating in transformers and conductors, nuisance tripping, and accelerated ageing of the very equipment a site depends on.
North America’s common reference for this is IEEE 519, the recommended practice for harmonic control on power systems, increasingly written into utility interconnection requirements as a condition of connecting non-linear load or embedded generation. In Canada the same role is played by CSA standards alongside IEEE 519. As more drives, chargers, and inverters arrive on commercial and industrial sites, holding total harmonic distortion within those limits—and protecting equipment from it—increasingly calls for active filtering rather than a one-off survey or a passive trap tuned for yesterday’s load.
| Value lever | High-rate markets (CA, NE, NS) | Cheap-hydro provinces (QC, BC, MB) |
|---|---|---|
| Demand-charge reduction | Lead—large share of the bill | Strong—billed even where energy is cheap |
| Reactive-power penalty removal | Lead—billed kVA / kVAR charges | Strong where kVA billing applies |
| Energy savings | Strong—high $/kWh | Modest (hydro rate is low) |
| Capacity release | Rising—defers service upgrades | Lead—the main growth enabler |
| Harmonics & protection | Strong—IEEE 519 at interconnection | Strong—IEEE 519 / CSA |
Rates shown are indicative commercial averages and exclude the demand and reactive-power charges layered on top of them; verify current figures with the relevant utility.
Section 06
For an operator in the United States or Canada, the order in which power quality pays back is the inverse of the weak-grid world. The grid is comparatively reliable, so reliability is not the headline—the bill is. Demand-charge and reactive-penalty removal lead. Capacity release is the rising second lever, lifted by the data-centre surge. Reliability and equipment protection ride alongside, valuable but rarely the reason a project starts.
HarmoniQ installs a coordinated, solid-state system at the low-voltage switchboard, as a parallel, removable retrofit—sized to the site and commissioned without breaking circuits or stopping production. Three products are deployed as the site requires. The HarmoniQ Booster corrects power factor in real time, which on a North American tariff means fewer billed kVA, the direct removal of reactive-power penalties, and freed transformer capacity—the lever that maps most directly onto the bill. The HarmoniQ Filter (HPF) holds harmonics within IEEE 519 and the relevant CSA limits, the standard now increasingly required at interconnection. HarmoniQ Alpha integrates the system and gives operators visibility—metered, switchboard-level insight into power factor, demand, and distortion as they are billed.
On a North American bill, better power factor is not an abstraction—it is a smaller number on the demand line and the disappearance of the reactive-power charge. Because demand is so often billed on kW÷PF or on a percentage of kVA, lifting power factor from 0.80 toward unity cuts billed demand directly, before a kilowatt-hour of consumption changes. Where capacity is tight, the same correction frees headroom on the existing connection—letting a growing site add load without paying for a service upgrade it can defer.
Every installation is held to the same standard HarmoniQ applies worldwide: a minimum performance guarantee, with results proven at the customer’s own meter and switchable on and off so the difference can be confirmed in metered results in real time. In a market where the bill itemises the problem, the meter is also where the solution is verified—the same instrument that prices bad power confirms its removal.
North America’s industrial power problem is not that the grid fails. It is that the grid charges, plainly and precisely, for power factor and peak demand—and that, for the first time in a generation, the spare capacity inside a building is worth more than the cost of releasing it. Both of those are within an individual operator’s control. What an operator cannot control is the surge tightening the grid around them; what they can control is how much of their own connection is wasted on reactive current and distortion, and how much of their bill is paying for it.
References
- U.S. Energy Information Administration (EIA), Electric Power Monthly and Annual Energy Outlook 2025—commercial and industrial average retail prices by state, and projected demand growth to 2030.
- IEEE Std 519-2022, “IEEE Standard for Harmonic Control in Electric Power Systems,” Institute of Electrical and Electronics Engineers.
- PJM Interconnection, Base Residual Auction results and capacity-market reports, 2024–2025.
- Ontario Energy Board (OEB) rate handbooks and distributor tariff schedules; Independent Electricity System Operator (IESO), Global Adjustment and Industrial Conservation Initiative documentation, 2025–2026.
- Alberta Electric System Operator (AESO), connection-queue and load-forecast reporting on data-centre interest, 2025.
- Individual investor-owned utility tariff schedules (demand and reactive-power / power-factor charge provisions). Verify current figures with each utility.
- CSA Group standards for power quality and electrical installations, applied alongside IEEE 519 in Canada.
- Provincial utilities—Hydro-Québec, BC Hydro, Manitoba Hydro, Nova Scotia Power—published commercial rate schedules, 2025–2026.
Figures in this article are drawn from public regulatory and market sources and are indicative of conditions in 2025–2026; rates, demand-charge structures, and reactive-power penalties vary by utility and province and change over time. Verify any figure against the relevant regulator or utility before relying on it.