The 2025 Residential Code Shift: What the New IECC §C405.3.2 Means for Your LED Recessed Can Spec
You’re specifying recessed downlights for a new single-family build in a state adopting the 2025 IECC—and your cut sheet just got rejected at plan review. Not because of color quality or dimming compatibility, but because the fixture’s tested efficacy is 118.3 lm/W. That half-lumen shortfall? It’s now a hard stop.
IECC §C405.3.2 doesn’t nudge—it resets. As of January 1, 2025, all newly installed recessed downlights in residential new construction must deliver ≥120 lm/W at the system level, verified per Title 24 Part 6 Appendix J. No averaging. No “typical” values. No manufacturer-declared nominal efficacy. Just measured lumens out, watts in—under standardized thermal and electrical conditions, with driver, trim, and housing fully assembled.
Why 120 lm/W changes everything—starting with the housing
I’ve reviewed over 40 spec packages this year. The ones that passed first time all shared one trait: they skipped “LED retrofit kits” entirely and specified integrated luminaires—meaning LED arrays, drivers, and thermal management built into the can itself. Retrofit kits—even high-performing ones—almost always fail Appendix J testing. Why? Because Appendix J requires testing the complete installed assembly: housing + trim + light engine + driver. Most retrofit kits assume an existing IC-rated can with unknown thermal mass, airflow, and junction box clearance. You can’t control those variables. So even a 135 lm/W module drops to 112 lm/W when wedged into a cramped 6" non-IC can with no heatsink contact. This falls flat because it treats efficacy as a component metric—not a system behavior.
What works instead: UL-listed, IC/AT-rated integrated downlights (e.g., 6" aperture, 9W input, 1,080 lm output). I’ve seen three pass consistently: shallow-profile torsion-spring housings with extruded aluminum heat sinks; deep-can designs with active thermal shunts to the ceiling plane; and modular systems where the driver mounts directly to the can’s outer flange—not tucked inside the junction box. All hit 120–126 lm/W in third-party Appendix J reports. Their packaging says so—literally. More on that shortly.
Occupancy sensors: not optional, but not universal
§C405.3.2 mandates occupancy sensor integration unless the fixture is installed in a space exempted under Table C405.3.2(1). That table isn’t a loophole—it’s precise. Bedrooms ≥120 sq ft? Sensor required. Hallways < 36 sq ft? Exempt. Closets with door-mounted switches? Exempt—if the switch is within 36" of the doorway and controls only that closet. But here’s what trips people up: “bathroom” isn’t a blanket exemption. Powder rooms (< 50 sq ft) are exempt. Full baths with shower/tub combos? Not exempt—even if they’re 48 sq ft. I think this reflects real-world usage patterns: people linger in master baths, but not in half-baths.
Crucially, the sensor must be *integrated*—not just “compatible with.” A wall-mounted PIR switch feeding line voltage to a standard downlight doesn’t satisfy the requirement. The control must reside within the luminaire’s listed assembly or be factory-wired as part of a listed system. That means either built-in vacancy sensing (with manual-on, auto-off), or a listed kit where the sensor head, wiring harness, and control logic are all submitted and tested as one unit.
Documentation: what your AHJ will actually check
Don’t send generic cut sheets. Your submittal needs:
- A signed, stamped test report showing Appendix J compliance—not LM-79 or ENERGY STAR data;
- A product ID label visible on the housing interior (not just the trim) listing: model number, tested efficacy (lm/W), input wattage, lumen output, CCT, and CRI;
- Manufacturer’s declaration that the listed efficacy applies to all trims offered for that housing (gimbal, baffle, adjustable)—not just the bare reflector;
- For sensor-integrated units: UL 1449 or UL 2750 listing evidence, plus wiring diagram showing sensor power source and load interface.
No AHJ I’ve worked with accepts “pending certification” or “tested to Appendix J draft.” They want the final, filed report—with lab accreditation seal, dated ≤12 months prior to submittal.
Labeling: small print, big consequences
New packaging and cut sheets must display efficacy in bold, 10-pt minimum font—positioned above any marketing claims (“Ultra-slim!” “Dimmable to 0.1%!”). And it must be the Appendix J value, not the LM-79 value. I’ve seen one major supplier get flagged twice because their cut sheet led with “132 lm/W (LM-79)” in large type, then buried “119.7 lm/W (Appendix J)” in footnote 8. That’s a rejection. Period.
Also new: the phrase “Complies with IECC §C405.3.2 (2025)” must appear on the primary product label—affixed to the housing, not the carton. Not “meets code requirements.” Not “designed for compliance.” Exact language. The AHJ told me last month: “If it’s not verbatim, it’s not accepted.”
Retrofit eligibility: the narrow path
Retrofits aren’t banned—but they’re cornered. To qualify, the entire assembly (housing + retrofit kit + trim) must be submitted as a single listed system—and tested per Appendix J in the exact housing it’s rated for. That means no “fits most 6-inch cans.” Instead: “Certified for use only in Brand X Model Y IC-AT housing, tested with Brand X Baffle Trim Z.” I’ve found that only two retrofit systems meet this today—and both require proprietary housings sold as matched pairs. If you’re working with existing drywall and standard stock cans, assume retrofits are off the table. Specify integrated from day one.