What It Costs to Run an Air Conditioner in 2026

The real price of cooling is not the sticker. It is the decade of electricity that follows.

Among the ENERGY STAR certified window air conditioners we track, the least efficient burns about a third more electricity than the most efficient for the same rated cooling. At 12,000 BTU that is roughly $105 versus $140 a year, and you will not find the gap on either price tag. We know this because we track the running cost of every certified room air conditioner we list, computed the same way for all of them, and the spread is not a rounding error. It is the difference between a smart buy and a quiet, recurring tax on your electricity bill.

Most air-conditioner shopping goes wrong at the same place. A buyer lines up two units, sees $499 next to $599, and picks the cheaper one. That comparison feels rigorous, but it measures the one cost that matters least. The purchase price is a single payment. The running cost is paid every summer for as long as you own the machine, and over a typical window-unit lifetime it usually adds up to more than the unit itself. A cheap air conditioner that is expensive to run is not cheap. It is a loan with the interest hidden in your utility bill.

This is not a small decision hiding in a small appliance. About 88% of US households run air conditioning, according to the U.S. Energy Information Administration, and cooling is one of the largest lines on a summer electricity bill. The EIA estimates air conditioning accounts for about 17% of home electricity use, and the average air-conditioning household spends about $265 a year just to stay cool. Those numbers are climbing, not falling. July 2026 was the hottest month ever recorded in the contiguous United States, NOAA reported, cooling demand across US cities is up more than a third since the 1970s, and residential electricity prices rose about 13% from 2022 to 2025 with further increases expected. The machine you buy this summer will run in a hotter, pricier decade than the one that shaped your expectations, which makes the running-cost number more important every year, not less.

This guide is the honest version of the math. We pulled the numbers from our own live catalog: 459 air conditioners, real retailer prices refreshed daily, and the ENERGY STAR annual-energy figures behind every certified room unit. We are going to show you exactly what a year of cooling costs by size and efficiency, why the cheapest unit to buy is rarely the cheapest to own, where the mini-split math flips, and how to read a spec sheet so the lifetime cost is visible before you pay. Everything here is priced at one standardized rate so the comparisons are fair, and we will be clear about the places where the honest answer is "we cannot give you a certified number," because those places matter as much as the ones where we can.

Contents

• 1. The three numbers that decide what an AC costs you
• 2. What a year of cooling actually costs
• 3. The efficiency gap: same cooling, different bill
• 4. Why we will not quote a running cost for a portable
• 5. What you actually pay upfront, by type
• 6. The install cost nobody prices in
• 7. Total cost of ownership over ten years
• 8. Your electricity rate and climate change the number
• 9. How to read a spec sheet for lifetime cost
• 10. The running-cost mistakes buyers make
• 11. Your decision framework

1. The Three Numbers That Decide What an AC Costs You

Every air conditioner has three costs, and only one of them is printed where you can see it. There is the purchase price, the number on the tag. There is the installation cost, which is zero for a window unit you lift in yourself and thousands of dollars for a mini-split that needs a certified technician. And there is the running cost, the electricity the machine draws every hour it is on, summer after summer. The reason so many air-conditioner purchases feel like a mistake a year later is that shoppers optimize the first number and ignore the other two, when the other two usually dominate the total.

Think about what you are actually buying when you buy cooling. You are not buying a box. You are buying a certain number of degrees of temperature drop, delivered over a certain number of hot hours, for a certain number of years. The box is just the delivery mechanism. The right frame is total cost of ownership, the sum of what you pay to acquire the machine, what you pay to install it, and what you pay to feed it electricity across its whole life. When you add those together, the sticker price often turns out to be a minority of the bill. For a typical window unit in our catalog, purchase price is roughly a third of the ten-year total and electricity is the other two-thirds. The number you fixate on in the store is the smaller one.

The components behave very differently over time, which is the whole point:

• Purchase price is paid once, up front, and never again. It is the most visible cost and the easiest to compare, which is exactly why it gets too much weight.
• Installation is also paid once, but it is invisible on the product page and can rival or exceed the hardware for a fixed system.
• Running cost is paid continuously, scales with how efficient the machine is, and compounds over the years you own it.

That third cost is the one this guide exists to make visible. It is the hardest to see at the moment of purchase because no price tag shows it, but it is the most predictable to estimate, because a machine's efficiency is a published, testable fact. An air conditioner's efficiency does not change with your mood or the weather; it is a fixed property of the hardware, so once you know the efficiency rating and the size, the annual electricity cost is arithmetic, not a guess. That is what lets us put a real dollar figure next to every certified unit we list, and it is what lets you compare two machines on the cost that actually adds up.

It is worth asking why the running cost is hidden at all, because the answer tells you how to shop. The market is structured to compete on the sticker, not the total. A retailer's product page exists to win the click and the sale, and the sticker price is the number that does that, so it gets the large type, the strikethrough "was" price, and the urgency banner. The running cost, by contrast, is paid to a completely different company, your utility, months and years later, so no retailer has any incentive to feature it. The cost that hurts you most is the one nobody in the transaction is paid to show you. That is not a conspiracy; it is just where the incentives point, and it is exactly the gap an independent comparison site exists to fill. When we put a running-cost figure next to a price, we are surfacing the number the sales page is structurally motivated to omit.

You can feel the effect in a single side-by-side. Two certified window units that ENERGY STAR rates at the same cooling capacity can sit three or four CEER points apart, and that alone moves the bill from about $105 to $140 a year at 12,000 BTU, a $350 difference over a decade for the same cooling. The cheaper-to-run one is often a mainstream inverter unit that costs no more to buy, so that $350 is simply waste a sticker-only comparison never surfaces. The pattern is sharper across types: a portable air conditioner usually carries a lower sticker than a window unit yet costs more to run and delivers less real cooling, so the cheaper-looking box is the more expensive one to own. This is not a rare edge case. It is the normal shape of the market, and it is why the rest of this guide fills each of those three numbers with real data so you can make the comparison the product page will not make for you.

The rest of this guide takes those three numbers one at a time, fills each with real data from our catalog, and then stacks them into a single ten-year figure so you can see which type of air conditioner wins for your situation. We build every number from first principles and from primary sources, the way we do across the whole of compare.ac, because a buying decision this expensive deserves better than a sticker-price reflex. If you want the short version of how we compute any figure you see here, it lives on our methodology page.

2. What a Year of Cooling Actually Costs

Here is the number almost no product page will give you: the typical window air conditioner in our catalog costs about $110 a year to run, and the full range runs from $70 a year at the efficient, small end to $385 a year for the largest, thirstiest commercial unit we track. Those are not invented figures. They come from the ENERGY STAR annual-energy rating published for each certified room unit, priced at a single national electricity rate so every unit sits on the same basis. We compute this for every US window unit in the catalog that carries an official ENERGY STAR certification (60 window air conditioners at the time of writing), and it is the single most useful number for deciding what an air conditioner will actually cost you.

The method matters, because a running-cost figure is only trustworthy if you know how it was built. We start with the machine's certified annual energy use in kilowatt-hours, the figure ENERGY STAR publishes from the Department of Energy test procedure. That test assumes 750 cooling hours a year, the national standard written into federal law at 10 CFR 430, Appendix F, which stands in for an average cooling season so every unit is measured against the same yardstick. We multiply that annual energy use by the US average residential electricity price of 18.44 cents per kilowatt-hour, the figure the U.S. Energy Information Administration reports for residential customers (May 2026, Table 5.6.A), and round to the nearest five dollars so the number never claims a precision it does not have. The identity is simple enough to check by hand: annual kWh equals the unit's BTU divided by its efficiency rating, times 750 hours, divided by 1,000.

This is not our invention; it is the government's own arithmetic. The Department of Energy's Federal Energy Management Program publishes exactly this calculation for a 12,000 BTU room unit run 750 hours a year, and its worked figures land right on the identity: a CEER 14.7 unit uses 612 kWh a year, a CEER 15.4 unit uses 584 kWh, and a wasteful CEER 10.9 unit uses 826 kWh. Run our formula on any certified room unit and you reproduce the published figure to within a fraction of a percent, which is why we trust it enough to print a dollar figure next to every certified machine. One detail matters when you check it by hand: the capacity in that identity is the one ENERGY STAR rated the machine at, which is not always the round number on the box. Friedrich's Kuhl and WallMaster lines, for example, are sold as 12,000 BTU units but rated at 12,500 to 13,600 BTU, and their published annual energy reflects the rated figure. So we always use the published number, and in the rare case a certified unit lacks one we derive it from the identity only where the nameplate and rated capacities are known to agree, and suppress it otherwise, because a standardized method is only honest when you also honor its limits.

It helps to know what 750 hours a year actually represents, so you can scale our figure to your own life. It is roughly a three-month cooling season at about eight hours a day, which is a reasonable national average and a good midpoint. But it is only a midpoint. A household in a mild Northern climate that runs the unit a month a year will spend well under our figure, while a household in Phoenix or Houston running the machine twelve hours a day for six months is closer to 2,000 hours, nearly three times the standard, and should mentally triple the number. Independent usage-based estimates capture this range: one calculator pegs a mid-sized window unit at about $138 across a four-month summer at eight hours a day, close to our standardized figure, while the EIA's average across all cooling households is about $265 a year once heavier users and central systems are folded in. The point of a standardized number is not to predict your exact bill; it is to let you compare two machines fairly, then scale the winner to your own climate and habits.

The distribution of those costs across sizes is steep, which is why size discipline saves real money. Here is the typical annual running cost by cooling capacity, using the median for each size bracket in our data:

The shape of that chart carries a lesson that our sizing guide makes in detail: cooling costs scale with capacity, so buying more BTU than your room needs is a permanent surcharge. An 8,000 BTU unit sized for a bedroom costs about $75 a year. Jump to a 24,000 BTU unit because "bigger is safer" and you are paying $250 a year, more than three times as much, to cool the same small room worse, because an oversized air conditioner short-cycles and leaves the room clammy. If you are not certain what size you need, work through our BTU sizing guide before you spend a dollar, because getting the size right is the largest single lever on your running cost after efficiency. The running-cost data and the sizing data point at the same conclusion from two directions: match the machine to the room.

Our median of about $110 a year deserves one clarification, because ENERGY STAR itself publishes a lower headline. The agency's own room air conditioner fact sheet states that certified units use about 10% less energy and cost less than $70 a year to run, on average. Both figures are correct; they simply describe different populations. ENERGY STAR's average is weighted toward the small bedroom-sized units most people buy, where our data agrees exactly: our cheapest class, the 8,000 BTU inverters, runs $70 to $75 a year. Our $110 median is higher only because our catalog carries the full size range up to 34,800 BTU commercial machines, which pull the middle up. The agreement at the small end, where the two datasets overlap, is the real check on our method, and it holds. It also puts a national number on the stakes: ENERGY STAR estimates that if every room AC sold in the US were certified, buyers would save more than $350 million a year.

There is a second, quieter lesson in that range. A 34,800 BTU commercial window unit costs about $385 a year to run, roughly $3,850 over a decade, which is more than the purchase price of most units in this catalog. At the large end of the size scale, the machine is no longer the main expense; the electricity is. This is exactly the point at which the running cost stops being a footnote and becomes the headline, and it is why the number belongs on the product page even though almost no retailer puts it there. Our per-unit pages and our cheapest-to-run guide exist to put it there.

3. The Efficiency Gap: Same Cooling, Different Bill

The single most valuable thing our running-cost data reveals is what happens when you hold size constant and let efficiency vary. Two air conditioners rated at the same BTU deliver the same cooling by definition, because BTU is a measure of cooling output. What separates them is how much electricity they burn to produce that output, and that is captured by a single number: the Combined Energy Efficiency Ratio, or CEER, the standardized efficiency rating that US room air conditioners have carried since the Department of Energy consolidated the older EER and standby measures. A higher CEER means more cooling per watt, which means a lower bill for the same comfort.

It is worth understanding what CEER actually measures, because the "combined" in the name is the whole point. The older EER (Energy Efficiency Ratio) counted only the electricity a unit drew while actively cooling, which quietly ignored the power a modern air conditioner sips around the clock through its electronics, its display, and its standby mode. The Department of Energy replaced EER with CEER precisely to close that gap: CEER divides cooling output by the total power the unit uses, cooling plus standby and off-mode, so it reflects the real draw of a machine that sits plugged in all season. That makes CEER the honest single number for running cost, and it is the number the DOE test procedure and the ENERGY STAR criteria are both built on. When you compare two units on CEER, you are comparing the whole electrical footprint, not just the part that happens while the compressor is spinning.

Because the annual figure is rated capacity divided by CEER times 750 hours, CEER converts directly into dollars once you fix the size. Per 1,000 BTU of rated cooling, a CEER 16 unit uses about 47 kWh a year and a CEER 12 unit about 63 kWh, a third more electricity for the same cooling. Our certified window units span exactly that range, from CEER 12.0 on a Friedrich WallMaster to CEER 16.0 on the GE Profile inverters, the Midea U-shaped inverter and the Friedrich Kuhl 8,000. At a rated 12,000 BTU, that spread is about $105 a year at CEER 16 against $140 at CEER 12, and the room reaches the same temperature either way. The only thing that changes is how much of your money leaves through the meter to get there. Here is what each CEER point is worth at that size, computed with the same identity we use on every unit page:

Two things about that chart deserve care, because honesty about the caveats is what makes the headline trustworthy. First, the low-CEER end of our catalog is almost entirely through-the-wall and heat-pump variants, which trade some efficiency for a different installation or for winter heating; they are not simply "worse" versions of the same product, and their higher sticker price reflects those features rather than laziness. They are also, as a rule, bigger than the box says: the Friedrich WallMaster sold as a 12,000 BTU unit is rated at 13,500 BTU by ENERGY STAR, so when it publishes 767 kWh a year (about $140) against the 600 kWh (about $110) of a GE Profile 12,000 BTU inverter, roughly half of that gap is extra cooling and the other half is its lower CEER of 13.2 against 15. Compare on CEER and on the capacity ENERGY STAR rated, not the nameplate, and the numbers stop surprising you. Second, and more importantly, the efficiency gap does not track the price gap. At a rated 14,000 BTU, the GE Profile inverter (CEER 16) uses 656 kWh and costs about $120 a year to run at a $394 sticker, while the LG LW1522IVSM dual inverter at the same rated size (CEER 14.7) uses 714 kWh for about $130 a year at $498. And at a 10,000 BTU nameplate, the GE Profile inverter (CEER 16, 469 kWh, about $85 a year, $298) sits next to a Friedrich WallMaster through-the-wall model that ENERGY STAR rates at 11,100 BTU and CEER 13, which draws about 640 kWh and costs $120 a year to run at $1,199. In both cases the efficient machine is both cheaper to own and cheaper to buy, which is common at the top of the efficiency scale because inverter compressors have also become the mainstream mid-price choice.

The dollar figure over time is what makes this worth caring about. A $30 to $35 annual gap looks trivial next to a $500 purchase, which is exactly why buyers ignore it. But a window air conditioner is not a one-season appliance. Run that gap across a ten-year service life and it is $300 to $350, spent quietly, in small summer increments, for cooling you could have had for less. That is real money attached to a decision most people make in thirty seconds by comparing stickers. The fix is not complicated. It is to read the CEER rating alongside the price, which we surface on every unit page and rank directly in our most energy-efficient guide, so the cheaper-to-own machine is easy to find rather than buried.

Efficiency is also where the newest hardware genuinely earns its price. Inverter air conditioners, which vary compressor speed instead of slamming between full-on and off, dominate the efficient end of our data. A conventional single-speed unit runs at full power until the thermostat is satisfied, then shuts off, then lurches back to full power, and those startups are where a disproportionate share of the electricity goes. An inverter unit modulates, holding the room steady at a fraction of full draw once it has caught up. That is why the CEER 15 and CEER 16 units in our catalog are almost all inverters, and it is why the efficiency premium is not marketing: it is a different way of moving heat that costs less to run every hour it operates.

4. Why We Will Not Quote a Running Cost for a Portable

There is a conspicuous absence in everything above. We have given running costs for window units and said nothing about portable air conditioners, even though we track 70 of them in the catalog. That silence is deliberate, and it is one of the most important honesty decisions on the whole site, so it deserves a full explanation rather than a footnote. The short version is that not one portable air conditioner in our catalog carries an ENERGY STAR certification with a published annual-energy figure, so any running cost we printed for a portable would be a number we invented rather than a number we verified. We would rather show you nothing than show you a guess dressed up as a fact.

That is not an accident of our catalog. Portable air conditioners are formally excluded from ENERGY STAR certification altogether: the program's room air conditioner specification lists portables as an ineligible product type, so there is no ENERGY STAR portable anywhere, in our catalog or any other. The certified annual-energy figure that anchors every window-unit cost in this guide simply does not exist for the portable category. So when a review site hands you a tidy "costs $X a year" for a portable, ask where the number came from, because it did not come from the standardized test that makes such numbers comparable.

The deeper reason is a genuine technical problem with how most portables work, and understanding it will save you from a common and expensive mistake. A single-hose portable air conditioner, which is the overwhelming majority of the category, blows room air across its coil and pushes the resulting hot exhaust out through one hose to the window. The trouble is conservation of air: every cubic foot the unit blows outside has to be replaced, and it gets replaced by hot, unconditioned air pulled in through every gap in the room, around doors, through the window kit, under the floor. The machine is quite literally fighting its own exhaust, creating a negative pressure that drags the outside heat back in. The Department of Energy documented this exact effect when it wrote the test procedure: a single-duct unit draws all of its condenser air from the conditioned space, so the infiltration penalty hits it hardest, and the standard now subtracts that infiltration heat from the rating.

The result is that a portable's real-world cooling is meaningfully below its headline rating, which is why the industry rates them on a special standard, SACC (Seasonally Adjusted Cooling Capacity), that runs well below the older ASHRAE number still plastered on the box. The gap is not small. A single-hose portable's true DOE cooling often lands 30% to 50% below its advertised BTU, so a box marketed as "10,000 BTU" may deliver closer to 5,500 BTU of real cooling, while a dual-hose design, which pulls its exhaust air from outside instead of from the room, recovers most of the loss. Consumer Reports states the practical rule plainly: portables use more energy than similarly sized window units, and you should not compare a portable to a window unit on BTU alone. We cover the full labeling trap in our sizing guide.

Because a portable does more work than its rating suggests to cool the same room, it also draws more electricity than an equivalently rated window unit, not less. So the honest statement is not just "we lack a certified number." It is that a portable is very likely the most expensive way, per degree of cooling, to run an air conditioner, precisely the opposite of the impression its low sticker price creates. The category exists for good reasons, and we recommend portables in the right situations, but the reasons are about installation freedom, not cost:

• Renters and restricted windows where a permanent or window-mounted unit is not allowed or not possible.
• Rooms with no suitable window, such as a casement or a room where the window must stay usable.
• Temporary or occasional cooling, where you will run the machine rarely enough that lifetime electricity is not the deciding factor.

If one of those describes you, a portable is a perfectly rational buy, and our best portable air conditioner guide ranks the ones worth owning. But choose it with eyes open. You are paying a running-cost premium for installation flexibility, and the low purchase price is not the bargain it appears to be. The one thing we will not do is manufacture a tidy annual figure to make that trade-off look better than it is. When a certified number does not exist, the honest answer is to say so, explain why, and point you at the real trade-off, which is what this whole section is.

If a portable genuinely is your only option, there are three ways to keep its running cost from getting out of hand, and they follow directly from the physics above. Favor a dual-hose model over a single-hose one, because pulling exhaust air from outside instead of from the room removes most of the negative-pressure penalty and closes much of the efficiency gap with a window unit. Size it honestly to the SACC number, not the ASHRAE headline, so you are not buying a unit that is quietly undersized for your room and therefore running flat out all day, which is the most expensive way to operate any air conditioner. And seal the window kit properly, because a leaky exhaust setup lets the hot air you just pushed out slip straight back in, forcing the compressor to work longer for the same result. None of this makes a portable as cheap to run as an efficient window unit, but it narrows the gap, and it is the honest advice for the shopper whose situation leaves no other choice. The point of naming a category's weakness is never to shame the people who need it; it is to help them buy the least-bad option and run it well.

5. What You Actually Pay Upfront, by Type

Now the visible number, the one you came in knowing to look at. Even here, most buyers work from a vague sense of price rather than the real distribution, so it helps to see what each type of air conditioner actually costs across a live catalog. These are the median prices from real retailer listings in our data, refreshed daily, not manufacturer list prices or a reviewer's guess. The median is the middle of the market for each type, the price at which half the units cost more and half cost less, which is a far better anchor than the cheapest headline unit or the flagship at the top.

The first surprise for many shoppers is how close portable and window prices are. A portable air conditioner is the most expensive per unit of real cooling, yet its median price ($539) is actually lower than a window unit's ($599), which is exactly the trap: the lower sticker hides the higher running cost and the lower effective capacity we covered above. The window unit at a $599 median is the workhorse of the category, spanning a wide range from a $159 budget box to a $2,799 commercial machine, with most real buying happening between about $449 and $849. That range is where the efficiency-versus-price decision from Section 3 actually plays out, and it is where a few minutes of comparison pays for itself many times over.

It helps to know what moving along the window-unit price range actually buys, because the spread from $159 to $2,799 is not arbitrary. At the bottom of the range you get a small, single-speed, manual-dial unit: it cools, but it is louder, less efficient, and has no inverter, no Wi-Fi, and no fine temperature control. In the $449 to $849 band where most buying happens, you reach the modern inverter units that carry the high CEER ratings and the ENERGY STAR badges this guide keeps pointing to, which is exactly why that band is the sweet spot: it is where a modest increase in purchase price buys the efficiency that pays itself back. At the top of the range sit large-capacity, heat-pump, and commercial-grade window units, which cost more because they do more (higher BTU, winter heating, heavier-duty build), not because they are more efficient per BTU. The lesson is that spending more within the window category mostly buys capacity and features, while spending in the middle buys efficiency, and efficiency is the part that lowers your bill.

The step up to a single-zone mini-split, at a $1,615 median for the hardware alone, is where the category changes character. You are no longer buying an appliance you carry home; you are buying a fixed installation with an indoor head and an outdoor condenser, and the price reflects a more capable, far more efficient, and much longer-lived machine. A multi-split system, at a $3,050 median, extends that to several indoor units on one outdoor unit, cooling a whole home rather than a room. Crucially, and this is the number that ambushes people, the mini-split price above is hardware only. It does not include the installation, which for a fixed system is not optional and is often the larger cost. That is the subject of the next section, and it is the reason a mini-split's true entry cost is not $1,615 but something closer to double that.

It is worth naming what the price buys as you move up the ladder, because the higher number is not simply "more expensive," it is a different value proposition. A window or portable unit is cheap, self-installed, and movable, but louder, less efficient, and shorter-lived. A mini-split is expensive to buy and install, but quiet, far more efficient, and long-lived. The efficiency gap between the tiers is real and large: a window unit runs roughly SEER 8 to 15, while a modern ductless mini-split runs SEER 17 to 25 or higher, using 30% to 50% less electricity for the same cooling, and ENERGY STAR certified ductless systems set the efficient benchmark for the category. The lifespan gap is just as important: a mini-split lasts 15 to 20 years against a window unit's 8 to 10, so the higher upfront cost is spread across nearly twice the service life. The right choice depends entirely on your time horizon and your situation, which is why the ten-year total in Section 7 matters more than any single price here. If you want to see the live prices behind these medians, the whole catalog is browsable and sortable at compare.ac, and current markdowns are tracked on our deals page.

6. The Install Cost Nobody Prices In

The most misleading price comparison in the entire air-conditioner market is putting a mini-split's hardware price next to a window unit's total price, because one of them is complete and the other is missing its largest component. A window or portable unit installs itself: you lift it in, secure it, plug it in, and you are done, for zero additional dollars. A mini-split cannot be installed that way. It requires a certified technician to mount the indoor head, place the outdoor condenser, run and evacuate the refrigerant line set, make the electrical connections, and commission the system. That labor is not a formality; in most jurisdictions handling the refrigerant legally requires certification, which is one of the regulations we track on our rules page.

For a single-zone mini-split, professional installation typically runs $1,500 to $4,000 on top of the hardware, depending on the length of the line-set run, the electrical work required, and how accessible the mounting locations are. Independent cost surveys put the fully installed price of a single-zone system in a similar band, roughly $1,200 to $5,500 all in, with a 12,000 BTU system commonly landing around $1,500 to $3,100 for labor and materials. That is not a small adjustment to the sticker; it frequently equals or exceeds the price of the equipment itself. A $1,615 mini-split is really a $3,100 to $5,600 project once it is on the wall and running, which puts it in an entirely different budget conversation than a $599 window unit. This is the single most common way people underestimate what cooling will cost them, and it is invisible on every product page because the product page sells hardware, not projects.

It is fair to ask why you cannot simply install a mini-split yourself and skip the labor cost, the way you would a window unit, and the answer is part regulation and part physics. A split system's refrigerant lines usually arrive either pre-charged or needing to be vacuumed and charged on site, and handling that refrigerant is legally restricted to certified technicians in the United States, for environmental reasons we track in detail on our regulations page. Beyond the law, a botched line-set connection or an improperly evacuated system leaks refrigerant, runs inefficiently, and voids the manufacturer's warranty, so the "savings" from a self-install evaporate the first time the system underperforms or fails. The professional install is not a markup you can route around; it is a real, required part of owning a fixed system, which is precisely why it belongs in the cost comparison from the start rather than as a surprise line item after you have already fallen for the low hardware price.

We treat installation exactly the way we treat running cost: as a labeled category estimate with a disclosed method, kept out of the sort and the sticker price so it never silently distorts a comparison, but always shown so it can never ambush you. The reason we quote a range rather than a single figure is that installation genuinely varies:

• Line-set length drives both material and labor; a condenser far from the indoor head costs more to connect.
• Electrical work matters; a unit needing a new dedicated circuit or a panel upgrade adds cost.
• Access and mounting matter; a second-floor head or an awkward condenser location adds labor.

Because those factors are specific to your home, no honest guide can give you a single installation number sight unseen, and any that does is guessing. What we can tell you is the realistic range and the fact that, for a fixed system, you must budget for it as a real and often dominant part of the cost. This is also why a mini-split only makes financial sense over a long horizon: you are paying a large one-time premium up front in exchange for years of lower running costs and a longer service life, and that trade only pays off if you stay long enough to collect. The Department of Energy's guidance on ductless mini-splits makes the same point about matching the investment to how long you will use it. The next section puts real numbers on that trade.

7. Total Cost of Ownership Over Ten Years

Now we stack the three numbers into one. This is the figure that should drive the decision, and it is the figure no product page will ever show you, because it requires combining a price the retailer knows, an installation cost they do not quote, and a running cost they never mention. When you do combine them, the ranking of "cheap" and "expensive" air conditioners often reverses, and the reversal is the entire reason this exercise is worth doing. Building a total-cost model like this is everyday finance work, and AI labs pay finance professionals by the hour to build and grade the models a professional would; the finance roles listed on AITraining.jobs, a jobs board our own company runs, show where that work is open.

Start with the simplest and most common case, a typical window unit, and look at what a decade of ownership actually costs. A $599 window unit, at $0 installation and $110 a year in electricity, costs about $1,700 over ten years. Of that, roughly $1,100, about two-thirds, is electricity. The number you compared in the store, the purchase price, was the smaller third of the real cost. Here is that breakdown, and it is the single most important chart on this page:

That two-thirds ratio is the reason the efficiency gap from Section 3 is not a rounding error. If electricity is the majority of what a window unit costs you, then a 27% difference in electricity is a large difference in the total, and the machine's CEER rating deserves at least as much of your attention as its price. It also reframes what "expensive" means. A slightly pricier but more efficient unit is frequently the cheaper machine to own, because the purchase-price premium is repaid, with room to spare, out of the running-cost savings. This is the core insight of the whole guide, and it falls straight out of the arithmetic once you stop looking only at the sticker.

Now compare across types over the same ten years. Because running cost is only certified for room units, we are honest in the table below about where a comparable electricity figure exists and where it does not, rather than inventing one to fill the cell. The comparison is still decisive, because the upfront and installation numbers alone tell most of the story:

The table makes two things clear at once. For a single room on a modest budget or a short horizon, the window unit is almost unbeatable on total cost: it is cheap to buy, free to install, and its running cost, while real, is modest for a well-sized efficient model. For a whole home, a long stay, or a priority on quiet and efficiency, the mini-split wins despite its far higher entry cost, because it is rated on the seasonal SEER2 standard that reflects genuinely superior efficiency and it lasts fifteen to twenty years rather than eight to ten, spreading that large upfront cost across nearly twice the lifespan. The portable sits in an honest middle: chosen for freedom, not for cost.

The fairest way to compare machines with different lifespans is to divide the total by the years of service you actually get, because a $1,700 window unit that lasts a decade and a $4,000 mini-split that lasts twenty years are not really competing over the same period. A window unit lasts about 8 to 10 years, which is why we use a ten-year horizon for it; a mini-split lasts 15 to 20. So a $1,700 window unit over 10 years is about $170 a year of ownership, while a $4,500 installed mini-split over 18 years is about $250 a year before you even count its lower electricity draw. On that per-year basis the two are far closer than the raw sticker suggests, and once you fold in the split's 30% to 50% lower running cost and its quiet, whole-comfort operation, the long-horizon owner often comes out ahead. The window unit still wins decisively for a renter or a short stay; the mini-split wins for the owner who will be there to collect the back half of its life. The deciding variable is not the price. It is how many summers you will own the machine.

A decision this multi-dimensional is easier to see as a flow than as a paragraph. The logic that follows the money looks like this:

graph TD
  A["Choosing an air conditioner"] --> B{"Can you install a fixed unit,<br/>and will you stay 10+ years?"}
  B -->|"No: renting, short stay,<br/>or no permission"| C{"Is there a usable window?"}
  C -->|Yes| D["Window unit:<br/>lowest 10-yr total for one room"]
  C -->|No| E["Portable:<br/>freedom, at a running-cost premium"]
  B -->|"Yes: owner, long horizon"| F{"One room or whole home?"}
  F -->|One room| G["Single-zone mini-split:<br/>higher upfront, lower running cost,<br/>15-20 yr life"]
  F -->|Whole home| H["Multi-split:<br/>highest upfront,<br/>whole-home efficiency"]
  D --> I["Then pick the highest CEER<br/>you can afford at the right size"]
  G --> I
  H --> I
  E --> I

The last box in that diagram is the one that pays you back regardless of which type you land on. Once you have chosen the type, the efficiency rating is what determines your bill for the next decade, so within your chosen category and size, buy the highest CEER (for room units) or SEER2 (for splits) you can afford. That single habit, applied at the moment of purchase, is worth more over the life of the machine than almost any other choice, and it costs you nothing but the discipline to read one more number. Our cheapest-to-run and most-efficient guides rank the catalog on exactly that number so you do not have to hunt for it.

Two quick scenarios show how the same framework produces opposite answers. Consider a renter cooling one bedroom for the two years left on a lease. A mini-split is off the table (they cannot install it and would leave it behind), so the real choice is window versus portable. An efficient 8,000 BTU window unit costs about $400 to buy, $0 to install, and $70 a year to run, for a two-year cost near $540. A portable at a similar sticker will run more and cool less, so for this renter the window unit is simply the better machine on every axis that matters, and the decision is easy. Now consider a homeowner cooling a main living space they will keep for fifteen years. Here the window unit's $110-a-year running cost and its need to be replaced twice over that span start to add up, while a mini-split's higher upfront cost is amortized across a much longer life at a much lower running cost. The homeowner who runs the per-year-of-service math from earlier in this section will often find the mini-split is the cheaper machine to own by the time the fifteenth summer arrives, on top of being quieter and more comfortable throughout. Same three numbers, same method, two different right answers, because the inputs (installation freedom and time horizon) are different. That is what a real decision framework looks like: not a single "best AC," but a repeatable way to find the best AC for your situation.

8. Your Electricity Rate and Climate Change the Number

Every running-cost figure in this guide is priced at one rate, the US average of 18.44 cents per kilowatt-hour, and that is a deliberate honesty choice that deserves explaining, because your real bill depends on where you live and how hard you run the machine. We use a single national rate so that every unit is comparable to every other unit: the difference you see between two air conditioners on our site is a difference in the machines, not a difference in geography. If we priced each unit at a local rate, the comparison would be contaminated by which state we happened to assume, and two identical machines could show different running costs for no reason that has anything to do with the hardware.

The temptation to localize the number is strong, and resisting it is one of the least obvious but most important calls we make. Here is the trap. The hottest US states, where an air conditioner runs the most hours, tend to have the cheapest electricity. Florida is around 15 cents a kilowatt-hour and Arizona around 15 as well, both well below the national average, while California is around 33 cents and New York around 30. If we priced a Phoenix shopper's running cost at Arizona's low rate, we would show a reassuringly small annual figure on a machine that is going to run for six months straight, understating the real cost exactly where cooling demand is highest. A per-region price would point the wrong way, because we know regional prices but not regional cooling hours, and price alone tells a backwards story. So we quote one standardized national figure and disclose the real spread as context rather than folding it into a number that would mislead.

To make that concrete, here is what a single window unit, the one that costs $110 a year at the national average, actually costs to run in a handful of states, scaling our standardized figure by each state's residential electricity rate:

The same physical machine, cooling the same room, costs more than twice as much to run in California as in Florida, purely because of the price of electricity. That spread is the entire reason we refuse to bake a regional rate into the headline number: it would tell a shopper in a cheap-power, hot-climate state a reassuring story about a machine that is going to run for half the year, and it would understate exactly the cases where cooling costs matter most. It also shows why the efficiency premium is not a fixed dollar amount but a multiplier on your local rate. Cooling demand is climbing on top of this: warming has pushed cooling degree days up in 97% of major US cities since the early 1970s, so the "average season" the rating assumes is itself getting longer.

What that means for you is simple to apply once you know your own rate:

• If your electricity is cheaper than 18.44 cents, your real running cost is lower than the figure we show, roughly in proportion.
• If your electricity is more expensive, as in California, New York, or much of the Northeast, your real cost is higher, and the efficiency gap between units matters even more.
• If you run the machine far more or far less than an average season, scale the figure accordingly, because it assumes the standard 750 cooling hours.

The practical upshot is that efficiency matters most exactly where it is easy to ignore it: in a high-rate state, or in a long, hot cooling season, the gap between a CEER 13 and a CEER 16 unit is not $30 a year, it is considerably more, compounded over a decade of heavy use. A shopper in an expensive-electricity state who buys on sticker price alone is making the costliest version of the sticker-price mistake. If you want to translate any figure on our site to your own situation, take our number, multiply by your local rate over 18.44 cents, and adjust for how your cooling season compares to a standard one. The EIA's electricity data publishes the current average residential price for every state, which is the one input you need to make our standardized figure personal.

9. How to Read a Spec Sheet for Lifetime Cost

Everything in this guide reduces to a small number of specs you can read on any product page, if you know which ones carry the lifetime cost. The problem is that product pages bury these among a dozen marketing bullets, and the ones that actually predict your bill are rarely the ones in large type. So here is how to read a spec sheet the way this guide reads a catalog, pulling out the three or four numbers that decide total cost of ownership and ignoring the rest. Do this and you can estimate a unit's ten-year cost in the store, before you buy, from information the retailer already prints.

The first number to find is the efficiency rating, and for a US room air conditioner that is the CEER. This is the single strongest predictor of your running cost, because it is literally cooling output divided by electricity input. Higher is better, and for modern inverter units in our catalog the practical range runs from around CEER 13 up to CEER 16 at the top, with the median sitting at 15, while older and legacy units run lower still. The second is the ENERGY STAR badge, which is not marketing fluff: it means the unit met a federal efficiency threshold and, just as importantly for us, that a certified annual-energy figure exists, which is the only reason we can quote a trustworthy running cost at all. If a room unit lacks that badge, treat any running-cost claim about it with suspicion, including your own mental estimate.

The remaining numbers put those two in context and prevent the classic sizing mistakes:

• Cooling capacity in BTU tells you the size, which must match your room; too small never cools, too large short-cycles and wastes energy.
• Annual energy use in kWh, when published, is the direct input to running cost; multiply by your electricity rate and you have the bill.
• The rating standard itself, ASHRAE versus SACC for portables, tells you whether the headline BTU is real cooling or an inflated number.

Once you have those, the arithmetic is the same one we run: take the annual kWh, multiply by your electricity rate in dollars, and you have the yearly running cost; multiply by the years you will own it, and add the purchase and installation, and you have the total cost of ownership. There is no secret to it beyond knowing which numbers to pull and refusing to be distracted by the rest.

Here is the whole exercise on one real unit from our catalog. A GE Profile 12,000 BTU inverter window unit lists an efficiency of CEER 15, carries the ENERGY STAR badge, and publishes an annual energy use of 600 kWh. Multiply 600 kWh by the national rate of $0.1844 and you get about $110 a year to run. Over a ten-year life that is $1,100 in electricity, and at a roughly $394 sticker with no installation, its ten-year total cost of ownership is close to $1,500. Now put a less efficient 12,000 BTU unit beside it, a through-the-wall model whose published ENERGY STAR figure is about 767 kWh: it costs roughly $140 a year and burns $1,400 in electricity over the decade, about $300 more than the efficient unit, for the identical cooling. That single comparison, done in under a minute from two spec sheets, is the entire argument of this guide in one worked example: the numbers that decide the real cost are all published, and they routinely disagree with the sticker. If a spec sheet omits the efficiency rating or the annual energy figure, that omission is itself information; a manufacturer proud of a unit's efficiency prints it in large type, and one that hides it usually has a reason. For the units where the makers do publish honestly, we surface all of this on the per-unit pages so you do not have to reconstruct it, and we explain each figure on our methodology page. For the deeper mechanics of matching BTU to a room, our BTU sizing guide is the companion to this one: sizing decides the size row of the running-cost chart, and efficiency decides where in that row you land.

10. The Running-Cost Mistakes Buyers Make

Most of the money lost on air conditioners is lost through a handful of predictable errors, each of which comes from optimizing a visible number at the expense of an invisible one. Naming them plainly is the fastest way to avoid them, because every one of them is obvious in hindsight and nearly invisible at the moment of purchase. We see these patterns across the catalog constantly, and they are the reason a buying-advisor site like this one exists at all: the market is structured to make the cheap-to-buy, expensive-to-own choice feel like the smart one.

The first and largest mistake is buying on sticker price alone, treating the purchase as a single payment when it is really a down payment on a decade of electricity. The second is assuming a portable is the budget option, when its low price hides both a running-cost premium and a real cooling deficit from the single-hose problem. The third is oversizing "to be safe," which we see constantly; a bigger air conditioner in a small room costs more to buy, more to run, and cools worse because it short-cycles and never dehumidifies properly, a failure mode we cover in depth in our sizing guide.

The remaining two are subtler but just as costly over time:

• Ignoring the CEER rating, and so paying roughly 20% to 40% more for electricity than an equivalently sized, more efficient unit would have cost.
• Forgetting installation on a mini-split, and so comparing a complete window-unit price against a hardware-only split price, understating the split's real cost by thousands.

There is a sixth error that amplifies all the others: buying in a panic during the first heat wave. When the temperature spikes, demand for air conditioners spikes with it, the good efficient units sell out, prices firm up, and a hot, uncomfortable shopper makes a thirty-second sticker decision because they want relief today. That is the worst possible moment to be optimizing for the lowest immediate price, because it is exactly when the running-cost and sizing discipline gets abandoned. The defense is to decide before you are desperate. We track real price history and current markdowns across the catalog on our deals page, precisely so the efficient unit you want can be bought at a good price on a calm day rather than the thirstiest unit still in stock at a premium on the hottest one. The best time to buy an air conditioner is the week before you need it, not the hour after.

The thread running through all of them is the same: the expensive mistakes are the ones that stay hidden until the bills arrive, month after month, long after the checkout felt like a win. A fan-in-disguise or an evaporative cooler mislabeled as an air conditioner is the extreme version of this, where the buyer does not even get real cooling for their money, a problem we wrote about separately in most air conditioners online are not air conditioners. The defense in every case is the same discipline this guide has argued for throughout: look past the sticker, find the efficiency rating and the size, estimate the running cost, add the installation if it applies, and compare machines on the total. It takes a few extra minutes and it routinely saves hundreds of dollars.

11. Your Decision Framework

Pull all of it together and the decision comes down to two questions and one habit. The two questions decide the type of air conditioner. The habit decides how much it costs you within that type. Neither is complicated, and between them they capture almost everything that matters about what an air conditioner will cost you to own, which is a far better basis for a purchase than the price tag that dominates most people's thinking.

The first question is how long you will keep it and whether you can install a fixed unit. If you rent, expect to move within a few years, or cannot install a permanent system, a window unit is the lowest-total-cost choice for a single room, and a portable is the fallback when a window unit is not possible, chosen for its installation freedom rather than its economics. If you own your home, plan to stay a decade or more, and can install a fixed system, a mini-split's higher upfront and installation cost is repaid over its longer life and lower running cost, and it is the better long-horizon machine. The second question is simply one room or the whole home, which decides between a single-zone and a multi-split system on the fixed-installation path.

The habit is the one that pays you back no matter which type you chose:

1. Size the unit to the room first, using real BTU math, because oversizing is a permanent surcharge and undersizing never cools.
2. Then buy the highest efficiency you can afford at that size, CEER for room units and SEER2 for splits, because efficiency is the majority of your lifetime cost.
3. Add installation into the comparison for any fixed system, so you are comparing complete costs, not hardware against projects.

Do those three things and you will land on the air conditioner that costs the least to own for your actual situation, which is rarely the one that costs the least to buy. That is the whole argument of this guide in one sentence: the cheapest air conditioner to own is almost never the cheapest on the shelf, and the difference is hundreds of dollars hiding in a number no product page shows you. We built compare.ac to put that number back on the table, with live prices, honest running-cost estimates, and per-room sizing across the whole catalog, so the total cost of ownership is visible before you spend rather than obvious a year later. Start with the cheapest-to-run guide if running cost is your priority, the best window air conditioner guide if you want the workhorse pick, or the best mini-split guide if you are ready for a fixed system, and let the total, not the sticker, decide.

This guide reflects compare.ac catalog data and US electricity prices as of September 2026. Running costs are standardized estimates computed from ENERGY STAR annual energy use at the US average residential rate of 18.44 cents/kWh (EIA Electric Power Monthly, residential, May 2026) and 750 cooling hours per year (the DOE test standard); your real cost depends on your local electricity rate, how much you run the unit, and your climate. Prices are live retailer figures that change frequently. Verify current details on the unit page before purchasing.