A human body is mostly water, and water absorbs and reflects radio waves. When you walk across a room, the Wi-Fi signals crossing that room arrive at the receiver slightly differently than they did a moment earlier. Wi-Fi sensing is the practice of reading those differences and inferring what caused them — someone entered, someone is standing still and breathing, the room is empty.
No camera, no wearable, no new sensor. Just the network you already have, interpreted differently.
What changed in 2026
- Standardization work matured. A dedicated amendment for Wi-Fi sensing gave vendors a common way to request and report the underlying measurements, replacing a decade of proprietary implementations.
- Carriers and ISPs started shipping it. Motion detection as a feature of a provider-supplied router moved from pilot to broad availability in several markets, usually bundled with a home-monitoring subscription.
- Health-adjacent claims got more careful. Breathing and sleep monitoring via ambient radio drew regulatory attention, and vendors softened medical framing toward wellness language.
- The consent question surfaced publicly. Because sensing can be enabled by a firmware update on equipment already installed, several jurisdictions began asking whether occupants must be told.
What it can and cannot detect
| Capability |
Reliability |
Notes |
| Someone is moving in the space |
Good |
The core, well-solved case |
| Room is occupied but still |
Moderate |
Depends on detecting breathing-scale motion |
| Rough location or zone |
Moderate |
Improves with more access points |
| Breathing rate |
Moderate, in good conditions |
Needs a still subject and low interference |
| Distinguishing person from pet |
Poor |
The main source of false alarms |
| Identifying who it is |
Not reliably |
Research demonstrations do not translate to homes |
| Gestures at a distance |
Poor |
Works in labs with fixed geometry, not living rooms |
The gap between demonstration and deployment is wide here. Laboratory results with controlled geometry and trained models look impressive. A real home with a ceiling fan, a cat, an open window, and furniture that moves is a much harder environment, and false positives are the dominant complaint from people who use it.
Why it matters anyway
Even coarse presence detection is genuinely useful for automation. Lights that respond to a room being occupied rather than to a motion sensor's narrow cone. Heating that does not run in empty rooms. An elder-care check that flags an unusual absence of movement without putting a camera in a bedroom — which, for many families, is the difference between acceptable and unacceptable monitoring.
It also composes well with existing smart home infrastructure, since it produces a simple occupancy signal that any automation platform can consume. If you are building on Matter and Thread, sensing output is just another input, and Thread 1.4 explained covers the mesh side of getting it to your automations reliably.
The privacy problem
A camera is visible. A microphone is usually disclosed. Wi-Fi sensing has no physical tell whatsoever — the hardware looks identical whether the feature is on or off, and it can be turned on remotely by a firmware push to equipment a landlord, employer, or ISP controls.
That combination is new. The data is coarse enough that it is not obviously surveillance and detailed enough to reveal when a home is empty, when someone gets up at night, and how many people are present. Ask whether sensing is enabled on any router you do not own, and treat a provider-supplied device as capable of it by default.
Common mistakes
- Using it as an intruder alarm. Pets and airflow trigger it. You will disable the alerts within a week.
- Assuming it works through everything. Thick walls, metal, and concrete attenuate heavily. Coverage is not the same as your data coverage.
- Expecting identity. It detects that something moved, not who.
- Ignoring it on rented equipment. If the router is not yours, the sensing decision may not be yours either.
- Deploying one access point and expecting zone detection. Location inference needs multiple vantage points.
FAQ
Does Wi-Fi sensing slow down my network?
Marginally. It uses measurement frames that consume a small amount of airtime. On a modern network the impact is not noticeable in normal use.
Can it see through walls?
Partially, and it degrades with material. Drywall is nearly transparent to it; concrete and metal are not. It is not a through-wall imaging system.
Is it safe from a radiation standpoint?
It uses the same transmissions your network already makes. There is no additional emission beyond normal Wi-Fi operation.
Can I turn it off?
On equipment you control, generally yes, in the router settings. On provider-supplied equipment it varies, and it may be tied to a service you are subscribed to.
Where to go next
For the networking generation this rides on, read Wi-Fi 8 explained. For the smart home stack that consumes presence data, see Matter vs Thread and Thread 1.4 explained.