Ryan Smith had a problem nobody could see. As regional safety manager for Garney Construction in Kansas City, he ran crews that worked outside through Missouri summers, the kind where the heat index sits at 105 for six consecutive days and the humidity turns every breath into a transaction. His foremen walked the site hourly, checking on workers the way construction foremen have always checked on workers: with a question. Hey. You good? And the workers answered the way construction workers have always answered: yeah, I'm good. They said this while their core body temperatures climbed past 101 degrees and the capillaries in their skin dilated and their cognitive processing degraded by measurable percentages with every ten minutes of continued exertion, because the distance between feeling fine and collapsing is shorter than most people believe, and because admitting you need to stop on a job site carries a cost that a thermometer cannot measure.
So Smith bought armbands, Kenzen sensors that strap to the forearm and sample heart rate, skin temperature, ambient temperature, and physical activity every two seconds, feeding the data through a proprietary algorithm that predicts core body temperature in real time and pushes tiered alerts to a supervisor dashboard before the worker knows to be worried. The armband vibrates, the dashboard turns red, somebody walks over and pulls the worker into shade before the worker's own body has finished deciding whether to tell him something is wrong.
The Rule That Didn’t Ship
On August 30, 2024, OSHA published a Notice of Proposed Rulemaking for heat injury and illness prevention that would have applied to every construction site in the country. Two trigger temperatures: an initial heat index of 80 degrees Fahrenheit, requiring water, shade, and acclimatization protocols, and a high heat trigger at 90 degrees requiring mandatory rest breaks, buddy systems, and written prevention plans. Public comments closed January 14, 2025, post-hearing comments closed October 30, 2025, and then the rulemaking went quiet in a way that suggested nobody planned to wake it up. No target date for final action appears in the current administration's unified regulatory agenda. What remains is a 437-page proposal in a filing cabinet, politically orphaned, waiting for a government that wants to finish it.
OSHA's National Emphasis Program on heat, the enforcement initiative that allowed inspectors to cite employers under the General Duty Clause, expired on April 8, 2026. A revised NEP was extended in June for five more years, but it remains guidance, not a standard, and its citations rely on the General Duty Clause's notoriously high burden of proof. Five states have their own heat standards: California, Oregon, Washington, Colorado, and Maryland. The other forty-five do not. That leaves roughly 7.5 million construction workers in states where no specific regulation requires their employer to provide shade when the heat index exceeds 90 degrees.
Into this vacuum, the market shipped.
What the Devices Actually Do
Two companies dominate the construction wearable heat monitoring space, and neither is large. Kenzen, founded in 2014, sells a forearm-mounted sensor on a subscription basis, roughly $25 to $39 per user per month. Its algorithm fuses skin temperature, ambient temperature, heart rate, and accelerometer data to estimate core body temperature continuously, without the rectal or esophageal probes that remain the clinical gold standard but that nobody outside a research lab has ever convinced a framer to wear. SlateSafety, deployed at the Department of Energy's Savannah River Site in South Carolina, uses a similar armband form factor and monitors heart rate limit value, a metric comparing current heart rate to a calculated maximum sustained rate. When a worker exceeds 100 percent HRLV for five consecutive minutes, the supervisor intervenes.
Credible science backs both products, but it remains young. An April 2026 study from researchers in Saudi Arabia used Garmin Vivosmart 5 smartwatches to monitor 19 construction workers and trained attention-based LSTM neural networks to predict heat stress events, achieving 95.40 percent testing accuracy with precision, recall, and F1 scores of 0.982. Peer-reviewed and technically impressive, though conducted on just 19 people. A separate paper presented at the 43rd ISARC conference in Singapore in June 2026 demonstrated an entirely offline wearable edge-AI system running on an NVIDIA Jetson Nano that provides bilingual English-Spanish heat safety alerts without cloud connectivity, addressing the language barrier that contributes to the disproportionate death rate among Hispanic construction workers, who now represent 34 percent of the U.S. construction workforce while accounting for a growing share of heat fatalities: 197 deaths in 2011, climbing to 408 by 2022, a 107 percent increase over eleven years.
Beyond heat-specific devices, the broader construction wearable market is substantial and growing. ResearchAndMarkets estimates the global construction wearable technology market at $4.4 billion in 2024, projected to reach $7.3 billion by 2030. That includes smart helmets, exoskeletons, air quality sensors, and AR glasses alongside heat monitors, but the category is real and growing at 8.8 percent annually.
The Residential Blind Spot
Every published deployment of AI heat monitoring wearables involves commercial or industrial construction: DOE nuclear sites, oil refineries, infrastructure mega-projects with dedicated safety departments and budget lines for occupational health technology. Kenzen's testimonials come from companies like Garney Construction, which builds water and sewer systems. SlateSafety's showcase client is a federal contractor managing radioactive waste. Both the ISARC paper and SlateSafety's marketing envision deployment on large job sites with multilingual crews of dozens or hundreds.
Residential construction is none of these things. A custom homebuilder runs a crew of five to fifteen people, mostly subcontracted, on a project that lasts three to twelve months. The decision-maker is a general contractor operating on margins thin enough that a single blown inspection can erase a month of profit. There is no safety department. There is a foreman who doubles as the safety officer, and his training might consist of an OSHA 10 card earned six years ago, and the question he is equipped to ask when the temperature exceeds 100 degrees is the same question Ryan Smith's foremen were asking before the armbands arrived. You good?
Nobody has published data on heat monitoring wearable adoption among residential builders. We contacted Kenzen and SlateSafety and reviewed their public case studies, client lists, and marketing materials: zero residential references. We checked whether any of the four largest production homebuilders in the United States by revenue, D.R. Horton, Lennar, NVR, and PulteGroup, have publicly disclosed deployment of heat monitoring wearables or AI-based physiological safety systems. None have. D.R. Horton, the largest homebuilder in the country, was named to the National COSH 2026 Dirty Dozen list for "repeated safety violations and hazardous construction jobsite conditions," with OSHA citations spanning multiple states and inspection types. D.R. Horton builds roughly 90,000 homes a year and employs approximately 4,000 workers directly, with tens of thousands more working through subcontractors on its sites.
Mathematically, residential adoption costs the same as commercial: $39 per worker per month, $468 per worker per year, $4,680 for a ten-person crew. The decision calculus is not. A safety director at a DOE facility evaluates wearable technology against a regulatory compliance framework, an insurance liability matrix, and a corporate occupational health policy with defined ROI metrics. A residential GC evaluates it against whether the $4,680 comes out of the same budget as the next lumber delivery. Both math problems produce the same dollar figure, and only one of them gets approved.
The Case Against the Armband
The strongest argument against AI heat monitoring wearables deserves its full weight, not a paragraph of hedging before the rebuttal lands.
First, accuracy in controlled conditions does not equal accuracy in the field. That Saudi Arabian LSTM study's 95.40 percent accuracy was achieved on 19 workers wearing consumer-grade smartwatches in a structured research environment. A five percent false negative rate sounds acceptable until you scale it across the approximately 730,000 construction workers in the United States with limited English proficiency, each wearing a sensor in environments where sweat, vibration, dust, and physical contact with tools and materials can degrade signal quality in ways that a university lab does not replicate. At five percent, that is 36,500 missed alerts per year across the LEP population alone.
Second, the armband monitors the symptom, not the cause. A vibrating sensor tells you that a roofer's core temperature has reached 101.5 degrees. It does not change the production schedule that required a crew of four to complete a tear-off and reshingle in two days because the homeowner's financing closes Friday. It does not eliminate the piece-rate compensation structure that penalizes slower work. It does not alter the power dynamic in which a subcontracted laborer working without documentation decides, rationally, that stopping means risking more than his health. A vibrating sensor detects the crisis. The system that produced the crisis remains undisturbed.
Third, the data flows to the employer. Kenzen's architecture routes physiological alerts to supervisors through a dashboard. SlateSafety's deployment at Savannah River gives management real-time access to every worker's heart rate limit value. In an industry where 34 percent of the workforce is Hispanic, where NASP survey data shows that more than 50 percent of injured Hispanic workers reported being "ignored or not taken seriously" by their foreman, and where immigration enforcement actions are explicitly cited in D.R. Horton's Dirty Dozen listing, continuous biometric surveillance of workers raises questions that the marketing materials do not address. Can the data be used to screen out older workers, heavier workers, workers who trigger alerts more frequently? Can it be subpoenaed? Can it be weaponized in a wrongful termination dispute or used to contest a workers' compensation claim? The companies say no. But the policies governing what happens when the data exists will be written by employment lawyers, not engineers.
What We Could Not Verify
No peer-reviewed field trial of any heat monitoring wearable has been conducted in residential construction. That 28,000 annual heat-related injury figure from Alahmad et al. is a modeled estimate derived from workers' compensation claims data and temperature records, not a direct count of heat-specific incident reports, and the authors acknowledge that OSHA's lower estimate of 6,575 injuries per year is "extremely unlikely" while their own figure is probably an undercount. Kenzen and SlateSafety are private companies; independent verification of their algorithmic accuracy outside controlled studies is limited to self-reported deployment outcomes. We could not confirm whether any production homebuilder has deployed or piloted heat monitoring wearables at any scale, including limited trials. OSHA's heat-related fatality data for construction, frequently cited as "more than 40 percent of all heat-related worker deaths," reflects the General Duty Clause citation record rather than a systematic epidemiological survey, meaning deaths attributed to other causes, cardiac arrest in particular, may mask heat as a contributing factor.
The Question for the Person Paying the Mortgage
If you are building a home this summer, the person framing your walls is working in conditions that no federal standard specifically regulates. Technology to monitor that worker's core body temperature in real time, to alert a supervisor before heat exhaustion becomes heat stroke, exists and costs less per month than your streaming subscriptions combined. Your general contractor almost certainly is not using it. Its largest homebuilder was named to a national safety offender list four months ago, and the federal agency responsible for workplace heat safety spent five years drafting a rule that may never be finalized.
An armband is not the answer. It is a $39 monthly reminder that the answer was supposed to be a regulation, and the regulation did not come, and the market filled the space the way markets always do: with a product that works for the buyers who can afford it and ignores the workers who need it most. Every residential framing crew in Phoenix and Houston and Raleigh and Sacramento working through July without shade requirements, without mandatory rest breaks, without any physiological monitoring more sophisticated than a foreman's glance, is operating in the gap between what technology can do and what the industry chooses to pay for. The armband knows the worker is not okay. What matters is whether anyone with the authority to change the schedule is listening when it vibrates.