What is pre-cooling and how does it cut costs on time-of-use plans?

By Michael Fitzgerald, Founder & Editor, Phoenix Cooling Guide
Technically reviewed by Tom Mattingly, Owner & President, AZ Trusted Air LLC (ROC 317368)
Published 2026-09-18

Pre-cooling is one of the few bill strategies in the Phoenix metro that both major utilities publish themselves. APS lists pre-cooling the home before 4pm on weekdays as a savings tactic on its demand-charge time-of-use plan, and SRP tells time-of-day customers to pre-cool the house before peak hours start. It is not a contractor upsell, and it requires no new equipment — only a thermostat schedule and the correct plan name read off the bill.

Where Valley homeowners go wrong is copying a schedule from a neighbour. A household on the current APS residential time-of-use plan has to coast through 3 hours of on-peak pricing. A household on SRP's Manage Demand 5-10 p.m. and Save has to coast through 5 hours, starting an hour later in the evening. The same thermostat program lowers one bill and raises the other. What follows is what pre-cooling does, what the published research says it does not do, and how to build a schedule around the window that actually applies to a given meter.

What is a good pre-cooling schedule for APS and SRP peak windows?

APS is the simpler of the two to schedule for, provided the plan name is right. On the current residential plans — Time-of-Use 4pm-7pm Weekdays and Time-of-Use 4pm-7pm with Demand Charge — on-peak runs 4pm to 7pm Monday through Friday year-round, and every other hour of the week is off-peak. That gives a pre-cool run ending at 4pm and a coast of 3 hours. APS publishes the tactic itself: the demand-charge plan page lists pre-cooling or pre-heating the home during lower-cost off-peak hours before 4pm on weekdays among the ways to save on that plan.

Older APS plans did not move. The Arizona Corporation Commission ordered the residential on-peak window cut from five hours to three, and APS states that removing two hours required re-spreading the cost across the remaining three on-peak hours and all of the off-peak hours — so a shorter window is not automatically a cheaper one. Frozen plans were left alone. Saver Choice Plus keeps on-peak at 3pm to 8pm weekdays, a coast of 5 hours, and APS business rate plans remain 3pm to 8pm as well. A schedule built for a 4pm start on a Saver Choice Plus meter misses the first hour of peak entirely.

SRP differs in two ways. Its time-of-day plans peak later: Conserve 6-9 p.m. and Save is on-peak 6pm to 9pm weekdays year-round, and Manage Demand 5-10 p.m. and Save is on-peak 5pm to 10pm weekdays — a coast of 5 hours, through the hottest evenings of July. Both plans also carry a super off-peak band, the cheapest energy of the day, from 8am to 3pm every day, weekends and observed holidays included. The cheap window therefore closes two to three hours before peak opens, so the pre-cool run is best finished by 3pm and the house asked to drift through the gap.

Two housekeeping points. SRP's legacy Time-of-Use price plan is closed to new enrollment and will be discontinued as of or before the November 2029 billing cycle, a retirement approved by SRP's publicly elected board on February 27, 2025, with at least 120 days' notice to customers still on it. And on any demand-charge plan, APS warns that smart-thermostat geofencing and occupancy sensing work against the customer, because a hard recovery run triggered by someone arriving home can create a high demand spike inside the on-peak window. APS's guidance is to switch those features off and schedule pre-cooling instead. A thermostat set up for Arizona rate plans is a scheduled thermostat, not a reactive one.

Does pre-cooling really save money or just shift the cost?

The research is not ambiguous, and it is worth knowing before spending a summer on this. Modelling by the Residential Building Systems Group at Lawrence Berkeley National Laboratory found that mechanical pre-cooling always increases a home's total annual cooling energy. It reduces and shifts peak load; it does not reduce kilowatt-hours. The same work found that the most aggressive deep pre-cooling could eliminate up to 97% of annual peak cooling load, but carried the largest off-peak energy penalty of any strategy tested. The extreme setting buys the most peak reduction at the worst energy cost.

Phoenix is, on the record, one of the better places to try it. The LBNL work identified the hot-dry climate zones 2B and 3B — naming Phoenix — as where pre-cooling and related strategies were most effective, and found the high-cooling zones from 1A through 3B returned the best peak savings for each unit of extra energy spent. A homeowner in a mild coastal climate buys very little for the energy penalty. A Valley homeowner on a plan with a wide spread between on-peak and off-peak prices buys a great deal.

Because the whole case rests on that spread, the spread is what has to be checked. APS re-spread its residential pricing when the on-peak window shrank from five hours to three. SRP applied a temporary price reduction of $0.0038 per kWh across the May through October 2026 billing cycles, with prices returning to posted levels with the November 2026 cycle. SRP also bills on three seasons rather than two: Summer covers the May, June, September and October billing cycles, Summer Peak covers July and August, and Winter begins with November. Summer Peak is a season label, not a daily time block — a common misreading of an SRP bill.

One SRP rule catches people out in the shoulder months. The seasonal price in effect on the meter-read date is applied to all usage on that bill, so a bill spanning a season change is charged entirely at the new rate. A pre-cooling experiment run across a May or November boundary produces a result that has nothing to do with the thermostat. The clean way to test is hourly usage data, available in SRP My Account and in an APS online account: compare one pre-cooled weekday against a non-pre-cooled weekday with a similar outdoor high. The APS and SRP plan comparison and the demand charge explainer cover how each bill is built.

For Phoenix-area homeowners who want this looked at on their own system, Phoenix Cooling Guide recommends AZ Trusted Air LLC (ROC 317368), whose owner Tom Mattingly reviews this guide. (623) 252-6403 · second opinion on a quote.

How cool should I drive the house before peak hours start?

The LBNL pre-cooling studies tested setpoints of 72 F and 74 F against pre-cool windows of 8 hours, 5 hours and 3 hours. The finding that matters to a homeowner is which combination won: the short window at the higher setpoint gave the best trade-off in most climates, not the long deep freeze. The reasoning is straightforward. Every degree below the normal setting widens the gap between indoor and outdoor temperature, so the house sheds its stored cooling faster, and the extra kilowatt-hours are burned in the early afternoon when the equipment is already working against the hardest conditions of the day.

Only one utility-published recipe carrying both a depth and an end time turned up in the research for this page, and it is written for business customers: APS advises Cool Rewards business participants who want to pre-cool themselves to lower the thermostat a few degrees on an event day until 3pm, then return it to the normal setpoint. The depth — a few degrees, not ten — is the transferable part. The 3pm end time is not, because it belongs to the 3pm to 8pm business window rather than a residential window opening at 4pm, 5pm or 6pm.

How deep a Phoenix house can usefully go is set by whether it can hold the cooling, and Valley construction is where this breaks down. A Phoenix attic runs 150 to 170 degrees in July, and a great deal of Valley ductwork sits in it. A home whose supply ducts leak into that attic gives back a two-degree pre-cool before the peak window is half over, because the stored cool air is being delivered to the attic rather than to the bedrooms. The same applies with force to the rooftop package units common in Valley housing stock, where the entire air handler and every cabinet seam sits in full sun.

No utility and no laboratory publishes the correct number of degrees for one particular house, and anyone quoting one has guessed. The way to find it is to run the test and read the meter. Pre-cool a couple of degrees below the normal setting on one weekday, leave the next comparable weekday alone, and compare hourly usage. Then watch the recovery. If indoor temperature climbs past the normal setpoint before the on-peak window closes, the pre-cool was too shallow or the envelope too leaky, and the fix is duct sealing and attic work rather than a colder thermostat.

Does pre-cooling stress the AC or shorten its life?

This is the one question on the page with no primary source behind it, and saying so is more useful than picking a side. Contractor blogs assert both positions confidently: one camp claims an air conditioner is most efficient at a constant temperature and that swings beyond roughly five degrees drive bills up, while the other calls the extra-wear argument a myth outright. Neither camp is citing equipment-longevity research, because none was located for this page. Any percentage attached to the claim, in either direction, should be read as an opinion wearing a number.

What is not in dispute is the effect of outdoor temperature. Cooling capacity and efficiency both fall as outdoor air gets hotter, which is why manufacturers rate equipment at fixed outdoor conditions and why a Valley system labours hardest through the late afternoon in July. A pre-cooling schedule moves compressor hours out of that band and into the late morning, when outdoor air is cooler and the same compressor delivers more cooling per kilowatt-hour. That is the strongest defensible statement available in this direction, and it deliberately stops short of a lifespan claim.

The Phoenix context matters more than the scheduling question. The cooling season here runs from April into October, and design conditions reach 115 F, which is why national lifespan tables built on fifteen to twenty years of service mislead Valley homeowners. Because pre-cooling raises total annual cooling energy only modestly, any wear argument has to rest on that small increment — set against a system already accumulating run hours far faster than the same equipment would in a moderate climate. The Arizona lifespan page sets out the local numbers.

The documented life-shorteners in the Valley sit elsewhere. Monsoon dust and haboob debris load a condenser coil until the system cannot reject heat, high duct static from undersized returns keeps the blower and compressor working against pressure they were never selected for, and deferred maintenance compounds both. One practical caution does apply to pre-cooling: a system that is marginally sized or short on airflow may never recover after the on-peak window closes. That is a capacity or duct problem being exposed by the schedule, not created by it.