Why punta-hour demand charges distort resort operating costs

For most hotel groups, electricity is filed as a utility: a fixed cost of keeping rooms cool and kitchens running, reconciled property by property and rarely questioned above the site level. At portfolio scale, that reading is expensive. A third or more of a resort's bill is often not driven by how much power the property consumed. It is driven by the 15 minutes each month when the property drew the most power. That single distinction determines both how much cost is recoverable and how much volatility the energy line injects into the P&L, and neither shows up when the bill is treated as a flat monthly expense.

For a portfolio CFO or an asset manager overseeing several properties, this is worth understanding precisely, because the mechanism is invisible at exactly the altitude where the capital decisions get made.

Two meters, and finance is usually watching the wrong one

Under GDMTH, the CFE tariff for customers supplied at medium voltage with 100 kW or more of demand, the bill separates into distinct charge families, and two of them dominate a resort's cost. The first is the energy charge, measured in kilowatt-hours: consumption in the intuitive sense, the total electricity drawn over the period. The second is demand, measured in kilowatts: the highest average power the property drew in any 15-minute interval. GDMTH bills it twice, through a capacity charge that counts only intervals inside the punta window and a distribution charge that counts the highest interval of the month. Kilowatt-hours are the odometer. Kilowatts are the speedometer. The bill charges for both, and in a property that does not manage its demand, the speedometer can account for 30% to 40% of the invoice, by our estimate at CFE's August 2026 rates.

The reason this matters is that the figure finance tends to watch, consumption, is not the figure that moves the cost. A property can run a disciplined kilowatt-hour reduction program and see its demand charge barely shift, because a single 15-minute peak, one evening when chillers, kitchens, laundry, and guest cooling align, sets the demand CFE bills for the entire month.

Why hospitality load lands in the most expensive hour of the day

GDMTH prices energy across three time-of-use periods: base overnight, intermediate, and punta, the highest-priced window. On the national grid, which serves Cancún, the Riviera Maya, Puerto Vallarta, Acapulco, and Huatulco, weekday punta runs from 20:00 to 22:00 in summer, from the first Sunday of April to late October, and from 18:00 to 22:00 in winter. Baja California Sur runs its own calendar: in Los Cabos, summer weekday punta runs from 12:00 to 22:00, and winter has no punta period at all.

Consider what a resort is doing in those hours. Guests are back in their rooms with the air conditioning on, restaurants and bars are full, and kitchens are in service. Cooling, the dominant electrical load in most coastal properties, stays high after sunset because coastal nights stay hot and room occupancy peaks. That puts a large share of the property's daily demand inside the most expensive window the tariff contains, and every kilowatt of punta demand is billed through the capacity charge. In Los Cabos, the summer punta window covers the afternoon cooling peak as well. Elsewhere, the afternoon chiller peak falls outside punta but still sets the month's distribution charge.

The order of magnitude is unforgiving. At CFE's August 2026 GDMTH rates, each kilowatt of punta demand costs MXN 377.17 a month in capacity charges before IVA in the Peninsular division, which bills Cancún and the Riviera Maya, and in the Jalisco division. That is about USD 21 at the Banco de México FIX rate of 17.8413 pesos per dollar on September 28, 2026. In Baja California Sur the charge is MXN 246.78, about USD 14. The charge recurs every cycle until the load shape changes. A 400-room resort that draws 1,000 kW during punta pays about MXN 4.5 million a year in capacity charges alone, before distribution and IVA, and every 100 kW taken off that peak saves about MXN 450,000 a year. Actual figures depend on each property's load shape and CFE division, but the pattern holds across the sector.

The problem finance actually cares about is variance, not level

There is a second cost to the demand charge that rarely gets named, and it matters more to a finance director than the absolute number. Because a single peak event sets the charge, monthly energy cost becomes unpredictable. A quiet month and a busy month can produce very different demand charges for reasons no one at the property can reconstruct after the fact, and budgets built on an annual average overrun without explanation. For a CFO, unpredictable opex is often a worse problem than high opex, because it undermines forecasting and forces conservative reserves. Managing peak demand does two things at once: it lowers the line and it flattens it. The second benefit is the one that shows up in a cleaner variance-to-budget report.

If the demand charge swings month to month and nobody at the property can name the evening that set it, that is the first gap worth closing. Book a Quick Scan.

The variance a portfolio hides from itself

The sharpest opportunity in a hotel group is usually not inside any single property. It is in the spread between them. Two resorts of comparable size, class, and occupancy will often carry materially different demand exposure and different power factor penalties, driven by nothing more than local load discipline, tariff classification, and equipment scheduling. That variance is real money, and almost no group sees it, because each property's bills sit in a separate regional accounts-payable queue and are never laid side by side on a common baseline.

A portfolio screen closes that gap. A group that collects twelve months of bills across every property and normalizes them onto comparable baselines can rank its own sites by recoverable opportunity. The outliers surface. Attention and capital then go where the return is highest, instead of spreading evenly across a set of disconnected site expenses. This is the difference between a facilities errand and a capital-allocation framework, and it is a decision that belongs to finance.

The line no one reads

One smaller charge deserves mention because it is almost always present and almost always fixable: the power factor penalty. When a property's power factor falls below 0.90, CFE applies a surcharge, a standing tax on electrical inefficiency. Because it appears as a modest recurring line and never as a spike, it survives for years unexamined, and across a portfolio it compounds. Correction is typically capacitor installation, measured in weeks, with no operational disruption and no effect on guest experience.

A measurement problem before an engineering one

Demand exposure persists not because the remedies are exotic. Load scheduling, pre-cooling, power factor correction, and, where the economics support it, battery storage sized for peak shaving are all well understood. It persists because most properties cannot see their own demand curve. The monthly bill reports peak demand as one number per period after the fact, with no indication of which loads coincided to produce it, on which day, in which window. Without that resolution, an operator is managing a cost it cannot locate.

Independence is a structural advantage here, not a slogan. A firm that sells chillers will find a chiller solution. A firm that sells batteries will find a battery solution. A property's actual demand profile may call for neither, or for a sequencing change that costs almost nothing. Because Mexico Energy Partners sells no equipment and is compensated only by the client, the recommendation follows the data rather than an inventory.

Where to start

The first step is deliberately small. A Tariff and Peak Demand Quick Scan starts with a 30 to 45 minute conversation with a manager, with nothing to send in advance. We ask for bills only after that initial assessment. From there, the scan establishes whether the property sits on the correct tariff, where in the punta window its demand concentrates, and whether it is paying an avoidable power factor penalty. For a group, the same review feeds a portfolio screen that ranks every property on a common basis. The scan attaches no equipment and no obligation. For a portfolio spending materially on electricity, that is an inexpensive way to convert an unexamined expense into a managed line.

Sources

• GDMTH time-of-use periods by grid and season, and the 0.57 load factor used in the capacity and distribution demand formulas. Comisión Nacional de Energía, “Acuerdo por el que la Comisión Nacional de Energía autoriza el cálculo y ajuste de las tarifas finales del suministro básico que debe aplicar de manera individual a la Comisión Federal de Electricidad,” for January 1 to December 31, 2026, Diario Oficial de la Federación.
• GDMTH capacity charges of MXN 377.17 per kW per month in the Peninsular and Jalisco divisions and MXN 246.78 in Baja California Sur, before IVA. Comisión Federal de Electricidad, “Tarifas finales del suministro básico,” August 2026, Diario Oficial de la Federación.
• Exchange rate of 17.8413 pesos per US dollar. Banco de México, FIX rate, September 28, 2026.
• Demand share of 30% to 40% of the invoice. Mexico Energy Partners estimate from August 2026 GDMTH rates for the Peninsular division at load factors of 0.55 to 0.75.

---

Figures above are illustrative and reflect typical hospitality load profiles under the CFE GDMTH tariff. Actual results vary by property, tariff classification, load profile, and baseline conditions. Mexico Energy Partners is an independent, vendor-neutral advisory firm compensated exclusively by the client, and does not guarantee specific savings outcomes.