Smart Home Devices That Survive Wildfire Evacuation Shutoffs

The useful question is not whether a device looks good in a wildfire-preparedness roundup. It is whether it still performs an evacuation job after the utility has deliberately cut the power and the internet path has collapsed. Southern California Edison says a Public Safety Power Shutoff can be announced up to three days ahead, may last over 24 hours, and in some situations can arrive with only 1 to 4 hours of notice.[1] That is the operating condition. Any recommendation that assumes a powered router, a live cloud account, and a responsive app is already leaning on the part most likely to disappear.

Dated outage map: which devices still matter when PSPS overlaps evacuation

Status labels below are dated to Q3 2026, current as of August 25, 2026. “Confirmed” means the category has a clear evacuation job and a credible path to keep doing it locally or on backup power. “Conditional” means the device can help, but only after model-level verification. “Pre-outage only” means it may be useful before the shutoff, but should not be counted on during the same window an evacuation order may arrive.

The table is deliberately harsher than a normal buyer guide because wildfire evacuation is not a normal smart-home use case. In an ordinary evening automation, a cloud bulb missing a command is annoying. In a PSPS window, that same dependency can remove the only feature that made the device relevant to evacuation.

The shutoff is the compatibility test

A PSPS is not a surprise blackout in the usual sense. It is an intentionally operated safety event, and the warning window is awkward: sometimes there is enough time to charge batteries and run a departure scene; sometimes there is only a few hours’ notice; sometimes the outage lasts into the period when the owner is already gone.[1] That puts smart-home preparedness into a narrow lane. The device has to do useful work before the shutoff, or it has to survive on local control and backup power after it.

The rough resilience order is simple enough to use when buying: local mesh protocols first, then other local-control systems, then Wi-Fi devices with documented offline behavior, and finally cloud-only Wi-Fi devices at the bottom. A commercial installer’s PSPS guide makes the same practical distinction: Zigbee, Z-Wave, and Thread devices can continue responding locally as long as the hub is powered, while cloud-dependent devices lose function when the internet path disappears.[8] Treat the source for what it is—a trade installer blog, not a public-utility notice—but the compatibility principle is the right one.

Backup power matters just as much as protocol. A Zigbee sensor paired to a dead hub is not local resilience; it is a quiet plastic puck. The practical chain is the alarm, hub, router or LAN gear if needed, and the devices that perform the evacuation job. For the broader outage-survival architecture, the companion smart-home power-outage matrix and backup-battery sizing guide are where the watt-hours and runtime decisions belong. Here, the buying rule is narrower: if the hub is part of an evacuation function, it needs backed-up power or the device should not be credited for that function.

Detection comes first, and “smart” is the overlay

Smoke and CO detection is the least glamorous category here and the easiest one to underrate. It should not be. Wirecutter’s 2026 smart-smoke-alarm guide reports NFPA reliability figures showing hardwired alarms operated in 94% of house fires, compared with 85% for battery-only alarms; it also reports that 66% of home-fire fatalities occurred in homes with battery-only alarms.[2] Because those numbers are reported through Wirecutter rather than pulled directly here from the NFPA report, the attribution matters. The conclusion is still blunt: detection belongs in the life-safety tier before it belongs in the gadget tier.

A smart smoke alarm is useful when it adds reach without weakening the base alarm. Room naming helps. A phone alert helps before the internet fails. A hub trigger can help turn on lights, shut down HVAC, or release a verified lock. None of that replaces an interconnected alarm system with backup batteries and a siren that works when everyone is asleep and the router is not.

The 2026 product landscape is also in transition. Nest Protect, long treated as the default smart-smoke-alarm answer, is discontinued. Wirecutter identifies First Alert SC5 as the Google Home successor and says it interconnects with existing Nest Protect units. The same guide names Kidde Smart Smoke + CO and a Ring version with professional monitoring among current picks.[2] That is a dated status, not a permanent truth. Anyone buying for fire season should verify the current SKU, the app platform, and the interconnect rules before installing a mixed system.

The pre-purchase questions are not complicated, but they need to be asked before the device is on the ceiling:

• Does the alarm still sound locally on backup battery when power is cut?
• Do interconnected alarms still trigger one another without Wi-Fi or broadband?
• Does the smart alert require a vendor cloud path, or can the hub see the alarm locally?
• If the alarm is meant to trigger lights, locks, or HVAC, does that automation run when the internet is unplugged?
• What happens when the alarm, hub, router, and phone are each tested in failure, one at a time?

That last test is where many tidy automations fall apart. A phone push that arrives while the house still has broadband is not proof that the departure scene will run during PSPS. The wildfire alert failure-points guide is the right place to debug that chain; the buying decision here is whether the alarm still detects and alerts locally after the smart layers are stripped away.

Exit and visibility devices need local behavior, not just integrations

The exit-and-visibility tier is where smart-home advice gets sloppy fastest. Lights, locks, garage doors, and HVAC all map to real evacuation actions: leave lights on, leave access usable, know how to open the garage manually, and shut off air conditioning when instructed.[3][4][5] The smart version of those actions is allowed to be clever only if the manual version remains obvious.

Lights: automate the departure state before you lose the circuit

Lighting is a good smart-home wildfire category because the job is narrow and physical. People need to move through hallways, stairs, garages, and gates. Responders benefit from a visible address and an illuminated entry. A local scene that turns on the porch, entry, garage, stair, and gate lights is more useful than a dozen color bulbs scattered around the house.

The catch is power. A smart switch does not make a dead lighting circuit glow. Its value is in setting the home into the right state before PSPS, or keeping lighting available on circuits backed by a battery system, generator, or other legal backup arrangement. If there is no backup power for fixed lighting, add non-smart battery lanterns or emergency lights to the evacuation path rather than pretending an app will solve darkness.

Locks and garage doors: verify the release path, then test the automation

Smart locks are not automatically bad in an evacuation plan. A battery keypad can be useful. A physical keyway can be useful. A local Z-Wave or Zigbee lock paired to a backed-up hub can be useful. The problem is the casual assumption that every smoke event can safely and legally trigger every door to unlock. First Alert SC5 is reported to advertise smoke-triggered auto-unlock with Kwikset/Schlage locks and some garage openers, while Home Assistant documents a smoke-detected trigger for automations.[2][9] That still does not prove a specific lock will accept automation-driven unlocking in the exact platform, firmware, and security mode installed at the house.

The only dependable way to treat lock automation is as model-specific. Confirm whether the platform permits unlock actions from smoke events. Confirm what happens when the internet is down. Confirm whether the keypad works on battery. Confirm where the physical key is. Then write down the manual exit plan as if the automation does not exist.

Garage doors deserve the same impatience. If a vehicle is behind the garage door, the household needs to know the manual release procedure before the evacuation order. A smart garage controller that depends on a cloud service is a convenience while the grid is up. During PSPS, the real preparedness item may be a visible pull cord, a practiced release motion, and a plan for how the door gets closed or left after departure.

HVAC: put it in the departure scene, but do not let it own the plan

Thermostats and HVAC controls sit between evacuation and smoke management. For departure, the useful action is simple: turn off air conditioning when instructed. For sheltering or return, the useful action may be keeping smoke out by using recirculate mode and filtration. EPA guidance discusses reducing indoor wildfire-smoke exposure with measures such as portable air cleaners and HVAC settings.[7]

A thermostat is rarely the device that saves the plan. It is a device that should obey the plan. Put HVAC shutoff into a local departure scene if the system supports it, but keep the thermostat lower on the purchase list than alarms, local lighting, and a working exit path.

Cameras are evidence devices unless they record locally

A camera can be useful before evacuation. It can confirm whether smoke is visible near the property, whether a gate is open, whether a vehicle left, or whether lights are on. Afterward, it may preserve evidence for insurance or reconstruction. That is useful work. It is not the same as a life-safety function.

The split is local recording versus cloud-only recording. A camera that records to an SD card or a local NVR can keep evidence if the camera, network segment, and recorder have power. A cloud-only camera that loses broadband becomes a powered lens with nowhere reliable to send video. If the power is also gone, it is not even that.

For wildfire evacuation preparedness, buy cameras only after the detection and exit categories are handled. Then prefer local storage, wired power where practical, and a recorder on backup power. Do not give a cloud camera the same planning weight as an alarm siren or a manual garage release.

Air-quality monitors help with smoke management, not the evacuation order

Air-quality monitors are worth owning in wildfire regions, but they occupy a different lane. They help answer indoor-smoke questions: should the portable air cleaner run higher, is the room improving, should HVAC stay in recirculate, is this a bad time to air out the house. Wirecutter’s 2026 guide covers current home air-quality monitors, including AirGradient One, and EPA guidance supports using portable air cleaners and HVAC strategies to reduce indoor wildfire-smoke exposure.[6][7]

Outdoor AQI and public smoke context are still separate from indoor readings. EPA AirNow draws from a nationwide outdoor monitoring network, while PurpleAir is commonly used for PM2.5-oriented community sensor readings; Home Assistant also has a PurpleAir integration for bringing those readings into a local dashboard.[10] Those readings are useful for situational awareness, especially when deciding how to manage filtration, but they are not a substitute for evacuation orders or official emergency alerts.

The same restraint applies to Watch Duty and Genasys Protect. They are useful situational-awareness tools and alert channels, and both are worth knowing before smoke is visible.[11][12] They should sit beside county alerts, utility notices, and official evacuation instructions, not replace them. The smart home can react to information; it should not invent authority.

What earns a slot before fire season

The dated map at the top leaves a narrower buying answer than most wildfire smart-home lists. The devices worth buying first are the ones that keep an official evacuation action available locally: detect danger, wake people, light the exit path, preserve manual egress, shut down HVAC when leaving, and record locally if documentation matters.

• Earns a pre-season slot: hardwired interconnected smoke/CO alarms with backup batteries; local-protocol lighting for exits, porches, garages, gates, and address visibility; a backed-up hub if it is part of those functions.
• Earns a conditional slot: smart smoke alarms, smart locks, garage controllers, thermostats, local-recording cameras, and air-quality monitors—only after testing offline behavior, battery behavior, and manual fallback.
• Counts as pre-outage convenience: cloud-only cameras, cloud-only lighting, app-only locks, and Wi-Fi devices whose “smart” function disappears when the router or vendor cloud is unreachable.
• Does not deserve evacuation credit: marketing-heavy sprinkler or exterior-wetting systems treated as if they were proven substitutes for evacuation compliance, alarms, defensible space, or responder access.

A good wildfire smart home is not the one with the most categories checked off. It is the one whose important parts still do boring, verifiable work when the house is dark, the broadband is gone, and the owner is standing in the hallway trying to leave.

References

1. Public Safety Power Shutoff (PSPS) — SCE
2. The 3 Best Smart Smoke Alarms of 2026 — Wirecutter
3. Go Evacuation Guide | Ready for Wildfire — Ready for Wildfire
4. Wildfires — Ready.gov
5. How to Prepare Your Home For Wildfire Evacuation — WFCA
6. The 3 Best Home Air Quality Monitors of 2026 — Wirecutter
7. Wildfires and Indoor Air Quality — US EPA
8. What Happens to Your Smart Home During a PG&E PSPS — Full Spectrum Technology Group — August 2026
9. Home Assistant smoke-detected trigger — Home Assistant
10. PurpleAir integration — Home Assistant
11. Watch Duty — Watch Duty
12. Genasys Protect — Genasys Protect