A woman in Savannah, Georgia, called an energy consultant because her heating bills felt wrong. Allison Bailes III of Energy Vanguard ran a Manual J load calculation on her 2,600-square-foot home. Her gas furnace was rated at 89,000 BTU per hour of output. Manual J said she needed 60,000, and when Bailes then measured the furnace's actual runtime during the coldest days of winter, when outdoor temperatures hovered near the 29°F design point, real heating demand turned out to be 46,000 BTU per hour. Her furnace was nearly twice the size her house required.
She is not unusual.
In a study of 40 new-construction homes, Energy Vanguard found the average load was 1,431 square feet per ton of cooling capacity. Most HVAC contractors still size systems using a rule of thumb: 500 square feet per ton. Do the math on those two numbers and the typical contractor is installing an air conditioner two to three times larger than what the house actually needs. In some well-insulated homes, the gap reaches five times.
The Largest Heat Pump Study Ever Says This Is a Continental Problem
In March 2025, a team of researchers published what is, by sample size, the most comprehensive field study of residential heat pump performance ever conducted. The study, published in Nature Communications, analyzed real operational data from 1,023 heat pumps installed across ten countries in Central Europe, monitored over a two-year window from 2021 to 2023. It is not a laboratory simulation. Every data point comes from an internet-connected heat pump reporting actual energy consumption and thermal output in someone's home.
The findings should alarm anyone who has recently installed one of these systems. Efficiency varied by a factor of two to three between the worst-performing and best-performing installations. Seventeen percent of air-source heat pumps fell below existing European efficiency standards. Roughly 10% of all systems were oversized. And up to 24% of systems performed worse than the numbers printed on their product certificate labels.
That last figure is worth pausing on, because it means the efficiency rating you compared when shopping for your heat pump may bear little relationship to what the machine actually delivers once it is bolted to your house and connected to your ductwork.
Why Oversizing Makes Your Home Worse
An oversized air conditioner or heat pump does not just waste electricity. It makes your house less comfortable, and the mechanism is counterintuitive enough that most homeowners never connect the symptoms to the cause.
The system blasts cold air, drops the thermostat temperature fast, and shuts off. Then it sits idle while the temperature creeps back up. Then it kicks on again. This is called short-cycling, and it does three things simultaneously: it prevents the system from running long enough to pull moisture out of the air, leaving the house feeling clammy even when the thermostat reads 72°F; it subjects the compressor to repeated start-stop stress that wears it out years ahead of schedule; and it burns more electricity than a properly sized system running steadily through a longer cycle, because the National Renewable Energy Laboratory found that the real energy penalty of oversizing comes from off-cycle parasitic power consumption by controls and crankcase heaters that run whether the compressor is on or not.
If your house feels cold but sticky in summer, or if your AC seems to turn on and off every few minutes rather than running in long, steady cycles, you probably have an oversized system. And you probably paid more for it than the right-sized one would have cost.
The Slide-Rule Problem
Proper HVAC sizing has had a correct answer since 1986. The Air Conditioning Contractors of America published Manual J, a room-by-room calculation that accounts for insulation R-values, window types, building orientation, air leakage rates, and local climate data to produce a heating and cooling load in BTU per hour for each zone and the whole house. Building codes in most U.S. jurisdictions require it.
Almost nobody does it right.
The full Manual J process requires measuring every room, every window, every door. Cataloging insulation levels. Noting ceiling heights and duct locations. Entering it all into ACCA-approved software. Then pairing the results with Manual S for equipment selection and Manual D for duct sizing. A thorough job takes a trained technician two to four hours of on-site data collection followed by another hour or two of software modeling. And even then, as the ACCA itself acknowledges, the protocol builds in approximately 15% oversizing by default.
Contractors skip it because they cannot afford it. A residential HVAC replacement bid that includes a four-hour Manual J is a bid that loses to the competitor who shows up, eyeballs the existing unit, adds a half-ton "just in case," and quotes $800 less. Competitive dynamics punish accuracy. ENERGY STAR requires no more than 15% cooling oversizing for certified new homes, but enforcement is spotty and existing homes have no such constraint.
What LiDAR and AI Change
Conduit Tech, a startup founded in 2022 by Shelby Breger and Marisa Reddy, built an HVAC design platform that uses LiDAR scanning, the same depth-sensing technology in recent iPhones and iPads, to create a precise 3D model of a home in minutes rather than hours. Its software maps room dimensions, window placements, and wall orientations from a walkthrough scan, then runs an ACCA-certified Manual J calculation automatically. ServiceTitan, the $9 billion field-services software company, acquired Conduit Tech in September 2025 to integrate the tool across its contractor platform.
It is a straightforward value proposition: give a technician a tablet, let them walk through a house scanning with LiDAR for ten minutes, and get a permit-ready Manual J report before they leave the driveway. Proposal creation that took hours collapses to minutes. Contractors can serve more homes per day. Homeowners get a right-sized system with documentation proving it.
Other tools are converging on the same gap from different angles. IESVE offers physics-based HVAC load simulation using the ASHRAE Heat Balance Method with sub-hourly calculations across more than 100 HVAC system configurations. A November 2024 NREL technical report by Grimes-Casey, Thomas, and Masi-Perkins documented the gap between Manual J block-load calculations and full EnergyPlus building simulation for cold-climate heat pump sizing, an attempt to quantify exactly how much accuracy the simplified calculation sacrifices. And researchers at Big Ladder Software published a paper at SimBuild 2026 analyzing voluntary performance data from more than 130,000 variable-capacity heat pump products listed in the NEEP Cold Climate database, developing unified modeling approaches that work when only verified AHRI ratings are available.
None of this is magic. All of it is overdue.
What to Do If You Are Buying or Replacing HVAC
Demand a Manual J calculation. If a contractor quotes you a system size without one, that is a red flag on the level of a surgeon who skips the imaging and goes straight to the scalpel. Any contractor worth hiring has access to ACCA-approved software, whether it is Wrightsoft Right-J, CoolCalc, or one of the newer AI-assisted platforms.
Ask for the room-by-room output. A whole-house number alone is insufficient because it does not tell you whether the sensible and latent loads are matched properly, and getting this wrong can leave you with equipment that meets total cooling capacity on paper while falling 12% short on the temperature-control side that actually matters.
If the calculated load comes in at 500 or 600 square feet per ton, question the inputs. For modern code-built homes with decent insulation and windows, loads below 800 square feet per ton are suspicious. Well-built high-performance homes routinely come in at 1,500 to 2,500 square feet per ton.
Be suspicious of any proposal that adds a half-ton or a full ton "for safety." Manual J already builds in safety margin. Additional padding compounds an oversizing problem that the Nature Communications study found affects one in ten installations even in the European market, where regulatory oversight is tighter than in the United States.
If your contractor has access to LiDAR scanning tools, ask them to use them. ServiceTitan's Conduit platform and similar products are rolling out across contractor networks now. A scan-based calculation is faster, cheaper, and more accurate than a tape-measure-and-clipboard Manual J, and it generates documentation that helps if you ever need to prove your system was properly designed.
The Strongest Argument Against Trusting the New Tools
AI-powered sizing tools are only as good as the assumptions baked into their models. LiDAR captures geometry with precision, but geometry is only part of the problem. Insulation quality behind drywall is invisible to any scanner. Air leakage rates require a blower-door test that no scanning app replicates. Duct condition in attics and crawlspaces, where half of residential HVAC energy loss occurs, is not something a ten-minute walkthrough captures. And occupant behavior, which the Nature Communications study identified as a significant source of performance variability, is not something any algorithm can measure at the point of sale.
A contractor who runs a LiDAR scan, clicks "calculate," and quotes you a system without ever looking at your attic, testing your ductwork, or asking about your family's thermostat habits is doing the same thing as the contractor with the 500-square-foot rule of thumb. A more sophisticated tool does not prevent a more sophisticated mistake. Substituting speed for judgment is the same error regardless of the instrument.
What We Did Not Prove
The Nature Communications study covers European heat pumps manufactured by a single unnamed company. The researchers acknowledge that the findings need validation across other manufacturers and markets. American homes differ from European ones in construction methods, insulation standards, climate diversity, and HVAC system prevalence. The 10% oversizing rate found in Europe may be higher or lower in the United States, where Manual J enforcement is inconsistent and the competitive pressure to oversize is more acute.
Our waste calculation of $250 per household per year applies the ACCA's 30% energy penalty figure, which represents an upper bound for improperly designed systems, to the average American home. Many homes with oversized systems will waste less; some will waste more. The calculation also assumes the system is oversized and running, not that all 84 million owner-occupied households are affected equally. We could not find a peer-reviewed estimate of the percentage of U.S. homes with oversized HVAC systems, which is itself a damning data gap given how much energy this sector consumes.
The ServiceTitan/Conduit Tech acquisition closed in September 2025 with an expected deal close in Q3 2026. Independent evaluations of the platform's sizing accuracy versus traditional Manual J, or versus full EnergyPlus simulation, have not been published. We are relying on ACCA certification and ServiceTitan's corporate claims about speed and accuracy gains.
← Back to AI Home Building