Google's Loss of Pulse Detection Clears the FDA for Pixel Watch 3

What the feature actually does

Loss of Pulse Detection watches for one specific event: the moment a wearer's pulse stops. Google names the causes it targets — primary cardiac arrest, respiratory or circulatory failure, overdose, or poisoning.

When it detects a loss of pulse and the person doesn't respond to a check, the watch automatically prompts a call to emergency services. The key design choice is that it requires no action from the user. Unlike an ECG reading or a manual Safety Check, this is a passive monitor that acts precisely when the wearer cannot.

Google describes it as 'first-of-its-kind.' It joins an existing set of Pixel Watch safety tools — Car Crash Detection, Fall Detection, Irregular Heart Rhythm Notifications, the ECG app, and Safety Check — but it is the only one aimed at detecting the absence of a pulse rather than an abnormality in one.

The regulatory path and who registered it

The feature was announced in 2024 and has been available in 14 countries starting with the EU. U.S. availability was gated on FDA clearance, which Google received the day of this announcement, with rollout beginning at the end of March.

One detail worth noting: the FDA registrant is Fitbit LLC, a wholly-owned Google subsidiary. The health-hardware regulatory identity Google acquired with Fitbit is carrying the compliance work for a feature branded under Pixel.

The boundaries Google draws around it

The announcement is unusually explicit about what the feature is not. It 'may not detect every instance of a loss of pulse,' and it is 'not intended for users with preexisting heart conditions or those who require cardiac monitoring.'

Google also states it 'does not diagnose or treat any medical condition or provide follow-up care.' And the emergency call depends on ordinary conditions — a charged device and adequate cellular connectivity.

Loss of Pulse Detection may not detect every instance of a loss of pulse and is not intended for users with preexisting heart conditions or those who require cardiac monitoring.Montana Labs

These caveats are not incidental. They position the feature for people at low baseline risk who wouldn't otherwise be monitored, and away from patients who need clinical-grade cardiac surveillance — the exact population a consumer wearable is not built to serve.

What clearing a life-critical detector on consumer hardware demands

The interesting engineering problem here is the false-negative and false-positive tradeoff on a device someone wears all day. A model that triggers emergency calls too readily erodes trust and burdens dispatchers; one that misses events fails at its only job. Google's admission that it 'may not detect every instance' is the honest framing of that tradeoff.

For anyone shipping AI features that take consequential automatic actions, this clearance is a concrete template: define the triggering event narrowly, exclude the populations you can't reliably serve, disclose the physical dependencies that can silently break the action, and route the regulatory identity through an entity built for it. The value of Loss of Pulse Detection lives entirely in those constraints, not in the ambition of the promise.