A remaining-life range for an AC, heat pump or furnace, from published service-life figures. Never a single year.
15 yearsNAHB figure for central AC and heat pumpsNAHB / Bank of America Home Equity, Study of Life Expectancy of Home Components, February 2007: a 2006 telephone survey, meant as a general guide. The 70–130% band and condition factor are ours.
The short answer
Central air conditioners and heat pumps last about 15 years and furnaces 15 to 20, per NAHB’s 2007 study.
So a 12-year-old AC in average condition has roughly 2 to 4 years left, and it is time to price a replacement alongside the next repair.
That figure is a guide from a survey, not a field study, so this tool gives a range, never one year.
Estimate the years left
Service-life figure minus age, shown as a range. Nothing is emailed.
The estimator gives a range from published service-life figures. This section shows what is actually installed, from the EIA Residential Energy Consumption Survey, a federal survey representing 123,529,025 households.
Among homes with cooling equipment, 9.6% run equipment 20 years old or more and 38.4% run equipment at least 10 years old. Only 14.3% have equipment under two years old.
HyreHVAC analysis: that is a spread of installed ages, not a lifespan, and it is often misused. An old stock can mean durable equipment, delayed replacement, or both. A one-time survey cannot tell which, so we do not turn it into an average lifespan.
Does maintenance extend HVAC life?
Nobody has shown it. Almost every lifespan estimate says servicing extends it. We looked for the evidence and did not find it.
DOE reviewed 44 home HVAC field studies and concluded: "More research is needed to determine the relative impact of faults on HVAC system performance beyond energy (e.g., occupant comfort, indoor air quality, and equipment durability). Most studies have focused primarily on energy impacts of improper system design and installation practices."
No national dataset links maintenance to lifespan either. RECS records equipment age in every home it surveys but never asks about servicing. We checked all 799 variables in the public file.
The claim is plausible, just not proven, and nobody can evidence a number of extra years. What DOE does show is that 70-90% of installed systems carry a performance fault. A real diagnostic visit on a system that has never had one will likely find something worth fixing.
What shortens HVAC equipment life most?
Factor
How it causes damage
How common
Can you still change it?
Oversizing
Short cycling. The system hits the set temperature fast, shuts off, restarts. Starting is the hardest thing a compressor does
The coil runs too cold and sends liquid refrigerant back to the compressor, which compressors do not survive
more than 50% of systems
Yes: ducts, returns, filter, blower setting
Wrong refrigerant charge
Too little overheats the compressor; too much floods it. Both are silent until failure
Undercharge 29-78%, overcharge 4-50%
Yes: it can be measured and corrected
Duct leaks
The system runs longer for the same comfort, and can pull dirty unconditioned air across the coil
90-100% need duct work
Yes: sealing is one of the best-value fixes
A dirty filter left too long
It restricts airflow, which is the fault two rows up
Fully in your control
Yes: the cheapest fix there is
Coastal salt air
Corrodes the outdoor coil and cabinet from outside in
Depends on location
Partly: coated coils and rinsing help
Run hours
The common explanation. Systems in hot climates run far more but do not fail proportionally sooner
Varies by climate; see the map below
No, and it matters less than the rows above
Ranked by how much field studies suggest each matters, not by how often it gets blamed.
The top three are all about how the system was installed. Equipment life depends more on the install than on maintenance.
Does a hot climate wear out HVAC equipment faster?
Cooling degree days by state, 2020: weighted by population from each surveyed household’s location.HyreHVAC calculation from EIA RECS 2020 microdata.
Not as much as you might think. Cooling degree days are the best public stand-in for run hours. The map shows the spread: single digits in Alaska, over 4,000 in Hawaii and on the Gulf Coast.
We have no data linking climate to equipment failure, and run hours seem to matter less than install quality. A Phoenix system running four times the hours of a Seattle one does not fail four times sooner.
HyreHVAC analysis: climate clearly matters on the coast, where salt air corrodes the outdoor unit whatever its run hours. That is corrosion, not wear.
What are the signs an HVAC system is near the end?
A second refrigerant leak
A pattern, not an event. Each repair buys less time, and running low on charge between them strains the compressor.
Compressor failure out of warranty
The most expensive single part. On an older system this is usually where the math stops. Check warranty registration first: a covered part changes the answer.
Coil replacement on an aging system
Expensive, and usually a sign corrosion or age has reached the rest of the equipment.
Rising bills with no change in use
Often a developing fault, not age. Get a diagnostic visit first; it is often fixable.
It uses R-410A and needs a big repair
No rule requires replacing it. But R-410A supply tightens as HFC production is phased down, so refrigerant cost is a fair factor in a large repair on old equipment.
Age alone is not a sign
A 20-year-old system that works, has no fault and costs little to run does not need replacing because a calculator says so. Replace for a reason, not a birthday.
What can this lifespan estimator not see?
How well it was installed
The biggest factor, and invisible to any age-based estimate. A well-installed system at 15 may outlast a badly installed one at 8.
Whether it was oversized
Oversized systems rack up compressor starts fast. A third to a half of systems are oversized, and few owners know which they have.
Its actual condition
This is math on a date, not a diagnosis. It cannot spot the fault that ends the system next summer.
Service history
Nobody has measured how much it matters, so we leave it out rather than guess.
Two numbers people mix up
Is a warranty the same as a service life?
No. A warranty is a promise about parts. A service life is an estimate of how long the machine runs. Neither predicts the other, and mixing them up is how repairable systems get replaced.
A parts warranty is a contract term: if a listed part fails within the period, the maker supplies a new one. A warranty running out does not mean the machine is near the end.
What it usually does not cover: the labor to fit the part, the refrigerant to recharge afterward, or the diagnostic. So a covered failure can still bring a big bill. And a repair quote that looks like a reason to replace may shrink to a labor bill once the part is claimed.
The registration trap: longer parts warranties often require registering the equipment with the maker soon after install. If nobody did, the term drops to the shorter base period. The homeowner is rarely the one meant to register, and always the one who finds out.
What to do: before treating a big repair as the end, find the model and serial number on the outdoor unit and call the maker yourself. Ask: is it registered, what is the parts term, and is this part covered? One call can change the answer more than any input here.
We do not publish typical warranty lengths. They vary by maker and product line, and change. The one that matters is tied to your serial number.
What does each “it’s finished” signal really mean?
These five reasons are what people hear when told a system is done. Only some are about the machine. The third column says what each one actually tests.
The signal
How it is presented
What it really tests
What it should change
Age past the published figure
Proof the equipment is worn out
Whether a 2007 phone survey put this kind of equipment at 15 years. It tests a survey, not your machine
Nothing on its own. It raises the question; it does not answer it
A compressor failure
The end of the system
The most expensive part, but often the result of an airflow or charge fault nobody fixed
Ask why it failed first. Undercharge shows up in 29-78% of systems and overcharge in 4-50%; both kill compressors
A second refrigerant leak
The system is leaking everywhere
Whether the first repair fixed the cause or the symptom, and how long it ran low in between
A real reason to replace. Each repair buys less time, and running low strains the compressor
Rising bills, same use
Efficiency lost to age
Whether a fault has developed. DOE found installed systems delivering 59-84% of rated SEER; most of that gap is fixable faults, not age
A diagnostic visit, not a replacement quote. Often a cheap fix sold as an expensive one
“It uses the old refrigerant”
A legal deadline
Nothing about your system. No installation compliance date for residential and light commercial systems where all specified components were manufactured or imported before January 1, 2025. No rule requires replacing working equipment
Get a second quote. Refrigerant cost is fair to weigh on a big repair; a deadline is not, because there is none
Five reasons a system gets called finished, and what each is really evidence of.
HyreHVAC analysis: three of the five should send you to a diagnostic, not a quote. Most systems declared dead on these signals were never measured, and measuring costs less than replacing.
A deadline that no longer exists
Is there a deadline to replace R-410A equipment?
No. Through 2025, a lot of buying advice was built on an install deadline for R-410A equipment. That provision is not in force as reported, so treat older advice on this point as wrong.
EPA’s Technology Transitions program sets a global warming potential limit of 700 for new home AC and heat pump equipment, with a manufacture and import cutoff of January 1, 2025.
On installation, EPA’s current sector table states: No installation compliance date for residential and light commercial systems where all specified components were manufactured or imported before January 1, 2025.
Variable refrigerant flow systems are handled separately, with an install date of January 1, 2027, extended to January 1, 2028 for buildings with permits approved before October 5, 2023.
HyreHVAC analysis: your existing R-410A system is not covered. The rule governs new equipment. It does not require replacing a working machine, does not make servicing illegal, and sets no date for you to act.
What it does change: HFC production is being phased down, so R-410A supply tightens over time. That is a fair cost to weigh on a big leak repair, as a price, not a deadline. The difference is whether you are being sold urgency.
This federal program has already moved once. We read these figures from EPA’s pages on September 5, 2026. Check again before relying on them.
How did the refrigerant rule change?
The order of events explains the confusion. A rule was published, widely reported, then reconsidered.
2023
The final rule is published
EPA finalizes the Technology Transitions limits for home and light commercial AC and heat pumps: a GWP limit of 700 and a manufacture cutoff of January 1, 2025. It also carried an install deadline.
January 1, 2025
Making or importing higher-GWP systems stops
The cutoff takes effect. Makers have already moved home lines to lower-GWP refrigerants. Stock built before the date stays in the supply chain.
2025
The install deadline drives sales
The reported install deadline is used widely in sales talks all year, and much consumer content assumes it will hold.
September 5, 2026
The install date is not in force as reported
EPA’s sector table, read on this date: No installation compliance date for residential and light commercial systems where all specified components were manufactured or imported before January 1, 2025. Older advice is likely wrong on this point.
HyreHVAC analysis: when a deadline is used to close a sale, check it on the agency’s own current page. Sales scripts can lag a rule change by years.
Why not work out an average lifespan from equipment ages?
Because the survey does not hold the data that math needs. It is the most requested figure on this subject, and we will not invent it.
9.6% of US households with cooling equipment run it at 20 years or older. One square per thousand households. A spread of installed ages, not a lifespan.HyreHVAC calculation from EIA RECS 2020 public use microdata.
9.6% of US households with cooling equipment run equipment 20 years old or more, per the EIA Residential Energy Consumption Survey.
What that figure is: a snapshot of what is installed now. It records each machine’s age, not when anything was replaced, or why.
Why it cannot give a lifespan: you would need a replacement rate, and the survey does not measure one. An old stock fits three stories equally well: long-lasting equipment, owners who cannot afford to replace, or slow home turnover.
A second problem: the systems still installed are, by definition, the ones that have not failed. Reading life expectancy off survivors overstates it. It is the same trap behind “they don’t make them like they used to.”
HyreHVAC analysis: so the estimator uses a published figure, widens it into a range, and tells you the range is mostly our judgment. Less satisfying than “4.2 more years”, but honest. The chart after this section shows how much is ours.
How much of this estimate is sourced?
Only one input. A calculator that shows its math looks authoritative whether or not the inputs are, so here is the split.
Illustrative: how far each adjustment moves the published figure. Only the center line comes from a source; every bar either side is our choice, shown so you can discount it.HyreHVAC presentation of this tool’s own numbers. Illustrative only.
Sourced: the central service-life figure, from the NAHB study cited on this page. That study was a phone survey of manufacturers, trade groups and researchers, and it says to use it as a general guide only.
Our judgment: everything else. The 70–130% band is our choice of how wide the uncertainty is. The condition factors, 0.7 for poor and 1.1 for good, are estimates, and the calculator labels them that way.
Why publish them: the usual alternative is one number with no band and no disclosure. A range that says it is a range fails safely.
What to do: look at the width of the band. If even the low end sits well in the future, age is not your problem, and something else is.
What should you ask before accepting a system is finished?
Each question has a paper answer. None needs you to understand how refrigeration works.
Get the model and serial number, then call the maker yourself
The maker, not the contractor quoting the repair, holds registration status and the parts term. One call can turn a replacement back into a labor bill.
A number from a manometer. Low airflow shows up in more than 50% of systems and is tied to early compressor failures. If nobody measured it, nobody knows why the machine is failing.
These show whether the refrigerant charge is right. Any competent diagnostic takes them, and the technician will have written them down.
Ask why the part failed, not just what failed
Most HVAC parts do not fail at random. Replacing the part without fixing the cause buys the same failure again, just later.
Ask whether the diagnostic fee counts toward the work
It often does if you go ahead with the same company. Knowing that changes the real cost of a second opinion.
Do not accept age as a diagnosis
A system that works, has no fault and costs what it always has does not need replacing because it passed a published figure. Those figures are survey results, not expiry dates.
Do not let a refrigerant deadline rush you
No rule requires replacing working equipment, and the install deadline widely reported through 2025 is not in force as reported. Urgency built on it is a sales tactic.
What can this tool not do?
The estimator is math on a date and a dropdown. These are the things that math cannot reach.
It cannot examine your equipment
It has not heard the compressor, seen the coil or measured anything. It cannot spot the fault that ends the system next summer.
It cannot weigh install quality, probably the biggest factor
DOE’s review found 70-90% of installed systems with at least one performance fault. A well-installed machine at 15 may outlast a badly installed one at 8.
It cannot credit maintenance, because nobody has measured it
That servicing extends life is plausible but not shown in the studies we found. We leave it out rather than guess.
It cannot tell you what your warranty covers
Terms vary by maker and product line and depend on registration. That is tied to your serial number, and it often matters more than the range here.
It cannot price anything
HyreHVAC does not install, service or sell HVAC equipment, and holds no data on quotes, invoices or finished jobs. Nothing here is a price we observed. That is also why these tools can tell you to keep the system you have.
It can tell you to do nothing
If the range says the equipment is past its figure but it works and has no fault, keep it and reassess at the next failure.
If you replace, how have efficiency ratings changed?
Since 2023, equipment is tested differently, and the old and new numbers do not convert. That changes what “the same as I had” means.
Since January 1, 2023, home equipment is rated under 10 CFR part 430, subpart B, appendix M1, using SEER2, EER2 and HSPF2 instead of SEER, EER and HSPF. The same machine scores differently under the two tests.
We could not retrieve the appendix text itself, so we quote no conversion factor. The safe rule needs none: never compare a SEER figure with a SEER2 figure.
The federal minimum for a split heat pump is now 14.3 SEER2 and 7.5 HSPF2. ENERGY STAR asks for 7.8 HSPF2, 15.2 SEER2 and 11 EER2 on a split system.
HyreHVAC analysis: “it was a 16 SEER before” tells you little about what to buy now. Ask for the SEER2 of the exact proposed pairing and compare it to the minimum and the ENERGY STAR line. The SEER2 savings calculator shows what the difference is worth.
Questions this calculator answers
How long does a furnace last?
15 to 20 years, per the NAHB 2007 study. This tool uses 18, the middle of that band, then returns a range. The study was a telephone survey of manufacturers and trade groups, not a field measurement.
About 15 years for central AC and heat pumps, per the same study. How it was installed and its condition move that a lot. Nothing here is a warranty.
Why not give a single year?
Because no field study measures it. A calculator that says “your AC will last 4 more years” is dressing up a survey as math. A range is the honest answer.
Does regular maintenance make HVAC equipment last longer?
Nobody has shown it. DOE’s review of field studies found no evidence linking maintenance to longer life. It is plausible, and a first real diagnostic visit often finds a fault worth fixing, but no one can prove a number of extra years.
Should I replace my HVAC system just because it is old?
No. A system that works, has no fault and costs what it always has does not need replacing because it passed 15 years. Replace for a reason: a major failure, a repeat refrigerant leak, or rising bills a diagnostic cannot fix.
Do I have to replace my R-410A system?
No. EPA’s refrigerant rule covers new equipment. No rule requires you to replace a working R-410A system, and servicing it is legal. R-410A may get more expensive over time, which matters for a big leak repair.
Sources and methodology
Figures dated August 23, 2026. Last reviewed .
Study of Life Expectancy of Home Components, February 2007 (NAHB / Bank of America Home Equity, retrieved 2026-08-09. Public mirror; no longer hosted on nahb.org. A telephone survey of manufacturers, trade associations and researchers.)