No US Jurisdiction Requires a Thermal Scan Before the Drywall Goes Up. Your Insulation Defects Thank Them.

A code inspector in Harris County, Texas, walks through a 2,400-square-foot new-build during the insulation stage. He looks at the batts. They appear to fill the cavities. He marks the form, signs off, moves on. Total time on site: seven minutes.

Fourteen months later, the homeowner's January gas bill hits $387. The HERS rater she hires to figure out why finds three wall cavities where the fiberglass was compressed behind electrical boxes, two headers with no insulation at all, and a band joist with a six-inch gap where the batt was cut short. Effective R-value of those sections: roughly R-5, in a wall rated R-19.

The code inspection had passed. Every one of those defects was visible during framing. None required sophisticated equipment to detect. An AI-powered thermal camera would have flagged all seven anomalies in under four minutes.

What the Code Actually Requires

The 2024 International Energy Conservation Code mandates insulation at specific R-values for each climate zone. Section R402.4.1.1 includes a thermal bypass checklist that inspectors are supposed to verify. But the checklist is visual. Look at the cavity. Is there insulation? Does it appear to fill the space? Check the box.

RESNET formalized this with a three-tier grading system. Grade I means the insulation fills the cavity side-to-side, top-to-bottom, front-to-back, with no more than 2% voids and no compression, and all gaps less than 30% of intended fill thickness. Grade II allows 2–5% voids. Grade III is anything worse.

Grade III installation can reduce a wall's effective R-value by more than 50%, according to ASHRAE Fundamentals thermal bypass calculations. A 2016 DOE/PNNL field study of 171 job sites across eastern Pennsylvania identified insulation installation quality as one of the top four compliance weaknesses, alongside duct leakage, foundation insulation, and lighting. The study didn't report exact Grade I pass rates, but the fact that insulation quality ranked as a top-four problem after professional inspection suggests the visual method isn't catching what it should.

The Machine That Sees What Inspectors Miss

Infrared thermography reveals surface temperature differences. Point it at a wall during a heating or cooling cycle, and insulation voids show up as hot or cold spots against an otherwise uniform thermal field. It's been available for decades. What's changed: the cameras got cheap and the interpretation got automated.

In June 2025, Hikmicro launched SuperScene, a deep learning algorithm trained on thousands of thermal images of building defects. It runs on handheld cameras starting at $300. Point it at a wall. Blue overlay means insulation deficiency. Green means moisture anomaly. The AI handles the interpretation that used to require a Level II thermographer charging $150 an hour.

FLIR has been building similar capabilities into its ONE Pro line, though its current AI focus is electrical and PCB diagnostics rather than building envelope. The direction is clear. Both major thermal imaging manufacturers are moving toward automated defect detection in residential buildings, and the hardware costs less than a decent impact driver.

There's a catch. Thermal imaging requires a temperature differential between inside and outside, typically 15–20°F minimum, per EPA guidelines. Summer construction in mild climates may not produce enough delta-T for reliable scanning. More on this limitation below.

The Enforcement Gap, Quantified

I ran the numbers on what visual-only insulation inspection costs the US housing stock annually. The inputs:

Calculation: 1,000,000 homes × 30% defect rate × $400 avg. annual waste = $120 million in year-one energy waste from one year's construction. Over a 30-year mortgage: $3.6 billion in cumulative excess energy costs. And that's using a conservative 30% defect rate and mid-range waste estimate. If you use the upper bound on both, it approaches $7.5 billion.

Total cost of thermal scanning those same million homes before drywall, using AI-equipped cameras that a code inspector already on site could carry? A one-time $500–$1,500 camera purchase plus 30–60 minutes of inspection time per home. At $75/hour inspector time, that's $37.50–$75.00 per home. Call it $50 million across the full build volume.

Fifty million dollars in inspection costs to prevent $3.6 billion in energy waste. A 72:1 return. Even if my defect rate is off by half, it's 36:1.

Why Nobody Mandates It

Code committees move slowly. The IECC updates on a three-year cycle, and proposed changes require public comment periods, committee votes, and adoption at the state and local level. InterNACHI's pre-drywall inspection standards recommend thermal imaging but don't require it. No state energy code currently mandates infrared verification of insulation installation in residential buildings.

Commercial buildings are different. ASHRAE 90.1 and some commercial energy codes require building envelope commissioning that includes thermal imaging. Residential construction got no equivalent mandate, partly because the per-unit cost argument is harder to make when the home costs $350,000 and the scan adds $50.

The UK provides a cautionary example of what happens when you skip verification. A National Audit Office review of the Green Deal and ECO energy efficiency schemes found that 98% of external wall insulation installations—22,000 to 23,000 homes—had major defects requiring remediation. The oversight relied on visual inspection and paperwork. It missed nearly everything.

What the AI Can't Do (Yet)

Hikmicro's SuperScene is still in beta. The company has not published false positive or false negative rates. Independent validation by a third-party testing organization like the National Institute of Building Sciences has not occurred. We don't know how the algorithm performs on spray foam versus fiberglass versus cellulose, or whether it reliably distinguishes thermal bridging at studs (normal) from insulation voids behind studs (defect).

The temperature differential requirement limits when scanning works. In IECC Climate Zones 1 and 2—southern Florida, the Gulf Coast, the desert Southwest—summer construction may proceed for months without a 15°F differential between conditioned and unconditioned space. You could mandate HVAC be operational during insulation inspection, but that adds cost and scheduling complexity for builders who are already managing tight timelines.

And there's a harder question. A $300 camera in the hands of a code inspector with four hours of thermal imaging training is not the same as a $300 camera in the hands of a Level II thermographer with 500 hours of field experience. The AI closes some of that gap. How much remains unverified.

What Would a Mandate Look Like?

A reasonable model: add a thermal verification requirement to the IECC thermal bypass checklist. The inspector already conducts a pre-drywall visit. Add one line item: infrared scan of all exterior wall cavities, attic floor boundary, and rim/band joists. Any anomaly flagged by the camera triggers a physical inspection of the cavity before drywall authorization.

Cost impact to the builder: zero, if the code inspector owns the camera. Cost impact to the jurisdiction: $500–$1,500 per camera, one-time, amortized across hundreds of inspections per year. Cost impact to the homeowner: a $50 per-home scan fee added to permit costs, if the jurisdiction passes it through.

Compare that to the alternative. Fix insulation before drywall: $200–$500 in labor to pull and reinstall batts or add blown-in fill. Fix insulation after drywall: $2,000–$8,000 to open the wall, remediate, patch, texture, and paint. Live with the defect for 30 years: $9,000–$15,000 in excess energy costs, plus the mold and comfort issues that don't show up on a utility bill.

Limitations of This Analysis

The 30% defect rate is an estimate based on DOE field studies that didn't report exact insulation grading pass/fail rates, only that insulation quality was a "top-four" compliance weakness. The actual rate could be lower in jurisdictions with strong code enforcement or higher in markets with labor shortages. The $400/year energy waste figure is interpolated from EPA ENERGY STAR methodology, which models air sealing plus insulation improvements at 15% heating/cooling savings; the insulation-only component is a subset of that, and I've estimated it at roughly half. The 30-year cumulative calculation assumes no remediation or correction, which overstates the total but reflects the reality that most defects behind drywall are never found.

The Strongest Case Against Mandatory Scanning

Builder margin pressure is real. The National Association of Home Builders reports that regulatory costs already account for 23.8% of the final price of a new single-family home. Adding another inspection step, even a cheap one, lands on an industry that views every new mandate as a line item that gets passed to the buyer.

The AI algorithms are unproven at scale. Hikmicro admits SuperScene is in beta. Mandating a technology before independent validation creates a precedent that code committees rightly resist. A bad scan that clears a defective wall, or flags a compliant one for re-inspection, could create more problems than it solves.

And the temperature differential issue means the mandate either excludes warm-climate zones or requires mechanical conditioning during inspection—either outcome produces a code provision that doesn't apply uniformly, which is exactly the kind of complexity that makes adoption harder.

These are legitimate concerns. The technology needs independent validation, the algorithms need published accuracy data, and the implementation needs to work in all climate zones before a national code mandate makes sense. But the gap between what's technically possible and what's legally required is measured in billions of dollars of lifetime energy waste.

The code inspector in Harris County spent seven minutes on that insulation inspection. A $300 camera would have spent four. One of them would have caught the defects.