Buyer's guide

Furnaces: staging is the specification that matters, not AFUE

Two furnaces can both be rated 96% AFUE and behave completely differently in your house. What actually separates them is whether they can throttle down to match a mild day, or only run flat-out until the thermostat is satisfied.

Why AFUE tells you less than it sounds like it does

AFUE (Annual Fuel Utilization Efficiency) is measured under one fixed lab condition. It answers "how much of the fuel becomes heat," which matters, but it says nothing about how the furnace behaves across an actual heating season -- mostly mild days where the house needs a fraction of the unit's full output.

That's controlled by staging: whether the furnace is single-stage (full blast or off), two-stage (a high and a low fire), or variable-speed/modulating (throttles continuously to match the load). A single-stage furnace satisfies the thermostat quickly on a 45 F day, shuts off, and starts again soon after -- short-cycling. A modulating furnace just runs quietly at low fire and holds temperature steady. Two furnaces at the identical AFUE rating can differ enormously here, and the AFUE number doesn't distinguish them at all.

Sizing is where the real regret is

Furnace research keeps turning up the same failure mode this project has now found in boilers, heat pumps, and central air: units sized by matching what was already there, not by an actual heat-loss calculation.

Two concrete patterns are worth knowing before you get a quote. First, a furnace can "pass" a sizing calculation that looks close on paper -- five percent over the calculated load, say -- and still short-cycle badly in practice, because the on/off pattern a homeowner actually observes can work out to real-world oversizing well over 40%, once the paper number meets an actual house. Second, and more subtle: replacing an old furnace with a new one at the same input BTU/h rating is not a like-for-like swap if the new unit has meaningfully higher AFUE -- more of that same fuel input becomes usable heat, so the real output can jump by a third or more even though the "size" on the spec sheet looks unchanged. Ask for a Manual J calculation on the house as it is now, not a match to the old equipment's nameplate.

An oversized single-stage furnace is the worst combination -- no way to throttle down at all. A modulating furnace tolerates some oversizing better, because it can back off, but it isn't a substitute for sizing correctly.

Condensing efficiency creates a new siting problem

A 90%+ AFUE "condensing" furnace gets there by pulling extra heat out of the flue gas, which produces acidic condensate that has to drain somewhere. That's a genuinely new failure mode a conventional 80% AFUE furnace simply doesn't have: in a cold climate, a condensing furnace sited in an unconditioned attic, or venting through an exterior wall with a long condensate run, can freeze that condensate solid during a cold snap -- shutting the furnace down on a safety lockout at the worst possible time. This isn't a defect in the unit; it's a siting and installation-design question that has nothing to do with which brand you buy.

If your furnace lives in an unconditioned space and your winters get seriously cold, ask your installer directly how the condensate line is insulated and routed, and whether that installation has actually held up through a real cold snap locally -- not just whether the furnace is rated to.

"Communicating" thermostats can lock you into one brand

Some higher-end systems use a proprietary communicating control protocol between the thermostat, furnace, and outdoor unit rather than the old universal low-voltage wiring. That can genuinely improve staging and diagnostics -- but it also means a future thermostat, or a future outdoor unit if you later add a heat pump or AC, may need to come from the same manufacturer's matched lineup, not whatever you'd otherwise choose. This is the kind of thing an installer rarely volunteers, because it only shows up in the fine print of the installation manual, not in sales material. Ask directly: is this a communicating system, and what does that lock me into later?

Don't take a "leaking furnace" claim on faith

A cracked heat exchanger is a real, recurring failure mode and a legitimate reason to replace a furnace -- carbon monoxide can enter the airflow through a crack. It is also, unfortunately, a documented sales-pressure tactic in the replacement trade: some technicians use an unverified "you have a CO leak" claim to push a same-day replacement sale. Both things are true at once, which is exactly why you shouldn't take either the "everything's fine" or the "you need a new furnace today" answer purely on the technician's word. Ask to see the actual reading from a combustion analyzer, not just a verbal claim, before authorizing a full system replacement on safety grounds.

The 95% AFUE minimum is genuinely unsettled

For years, federal minimum efficiency for furnaces has been 80% AFUE. DOE finalized a rule raising that regional minimum toward 95% for most of the country, but as of mid-2026 that rule's legal status is unresolved after a Supreme Court remand. If you're being quoted a non-condensing 80% AFUE furnace right now, that's still a legal purchase today -- but ask your contractor directly what happens if the rule is reinstated during the equipment's install window, and don't assume the current legal minimum will still be the legal minimum in a year.

The installer matters as much as the furnace

Ductwork quality -- sealing, sizing, static pressure -- determines how much of a furnace's rated output the house actually feels, independent of the equipment itself. A well-installed mid-tier furnace on a properly sealed duct system will outperform a premium modulating furnace bolted onto leaky, undersized ducts. If your ducts have never been tested or sealed, that's worth doing at the same time as a furnace replacement, not treated as a separate, optional project later.

---

Equipment in this category