AC running all day: normal in a heat wave, or a real problem?

Set the thermostat and leave it alone for an hour during the hottest part of the day. Then compare the thermostat's reading against a separate thermometer in the room — not because thermostats are commonly wrong, but because this is the check that actually answers the question. If the room is at or within a degree or two of the setpoint, the system is doing its job; continuous running on a 95° afternoon is exactly what a correctly sized system is supposed to do, since the cooling load on the hottest hour of the day is the load the system was sized against in the first place.

If the room is still meaningfully above setpoint after an hour of continuous running — four, five, six degrees off — the system is running but not winning, and that's the version worth calling about.

Extended runtime on the hottest days of the year is expected, not a symptom. A system sized correctly for the house's peak cooling load is built to run close to continuously right at that peak — that's what "sized for the load" means. The mild-weather cycling pattern most people expect is what happens well below the system's peak demand, not a universal baseline the system should hold to on every day of the summer.

If the house never reaches setpoint even in moderate heat — not just on the single worst day of the year — the system is likely undersized for the house rather than just working hard on a hard day. This is common after a replacement matched to the old unit's tonnage instead of an actual load calculation, or in a house that's had insulation, windows, or an addition changed since the AC was last sized. A Manual J load calculation is the fix, sizing the system to the house's real cooling load rather than a guess.

A system low on refrigerant, or one with a dirty condenser or evaporator coil, loses real cooling capacity — it's not moving as much heat per hour as it's rated for, so it runs longer to try to compensate and often still doesn't reach setpoint. Both show up the same way from the thermostat's perspective: long runtime, weak results. A refrigerant leak needs to be found and sealed, not just topped off, and a dirty coil is exactly what an annual tune-up catches before it becomes a summer no-cool call.

The practical difference from an undersized system: this version is usually a change from how the system used to perform, not how it's always run. If the AC used to keep up fine on hot days and now doesn't, refrigerant or a dirty coil is the more likely cause. If it's never kept up, sizing is the more likely one.

A correctly sized system running nearly nonstop on a genuine heat wave is running at its normal, rated efficiency the whole time — long runtime on the hottest day isn't wasted energy, it's the system doing the job it was sized for. The expensive version is different: a system that's undersized, low on refrigerant, or running dirty coils is often running just as long, or longer, while moving less heat per hour than it should. That's the real cost of ignoring "it never shuts off, but the house never quite gets there" — the system keeps burning power at a worse ratio of cooling delivered per dollar spent, every hour it runs that way, not just on the one bad afternoon.

This is also why the setpoint test matters more than watching the electric bill directly. A higher bill during a heat wave is expected either way, correctly sized or not, so the bill alone won't tell the two situations apart — whether the house is actually reaching temperature is the signal that does.

Run the setpoint test first — it's free and takes an hour. If the house is holding near setpoint, there's nothing to fix; that's a system working correctly under real load. If it isn't, and the gap has shown up recently on a system that used to keep up, a diagnostic visit checking refrigerant charge and coil condition is the faster, cheaper first step. If the system has never kept up even in ordinary summer heat, the more useful conversation is whether it was sized correctly for the house in the first place.

This is close to the opposite problem of AC short cycling, and the two share a cause in common. Short cycling is a system stopping too soon, before finishing a real cycle; running constantly without result is a system that keeps going but never gets there. Refrigerant charge and coil condition show up in both — the difference is usually whether a safety switch is tripping (short cycling) or whether the system just doesn't have the capacity to catch up (running constantly). An annual tune-up catches both the same way, since static pressure, refrigerant charge, and coil condition are exactly the readings that predict either failure mode before it becomes a mid-summer call.