Smartwatch ECG AI Deal: AliveCor Plans Premature-Beat Checks for Apple Watch Only

November 17, 2026:

Smartwatch ECG AI Deal: AliveCor Plans Premature-Beat Checks for Apple Watch Only
Apple Watch Health
Victor Larsen/Unsplash.com

AliveCor and PulseAI said Monday at 09:00 ET that they will add PulseAI’s AI heart-rhythm analysis to the Kardia App for KardiaCare subscribers in the first half of 2027, but what a user gets would depend on the watch. Apple Watch would receive four rhythm determinations plus labels for premature ventricular contractions (PVCs), while Samsung Galaxy Watch, Fitbit Sense and Withings ScanWatch would receive two rhythm determinations.

The FDA clearance behind the software is also narrower than a consumer subscription might imply. The agency’s October 27 clearance letter for PulseAI’s PulseAPI SmartWatch software limits the indication to adults when prescribed or used under a physician’s care, and marks over-the-counter use as not selected. The partnership release says nothing about prescriptions.

If you own one of these four watches, the cleared label decides what the AI could flag and whether a clinician has to be involved. The announcement came 11 days after PulseAI said on November 5 that it had won the clearance, and it gives no price and no firm launch date.

Apple Watch Would Get Four Rhythm Calls and PVC Labels While Three Rivals Get Two

The FDA letter describes PulseAPI SmartWatch as a software analysis library that analyzes Lead I ECG data, which measures voltage between the arms and gives a single view of the heart. On Apple Watch, the release lists four determinations: sinus rhythm, atrial fibrillation (AFib), bradycardia and tachycardia. The release also describes beat-by-beat PVC classification that separates normal sinus rhythm from sinus rhythm with PVCs, and the FDA table labels each beat PVC or non-PVC.

On Galaxy Watch, Fitbit Sense and ScanWatch, the release lists sinus rhythm and AFib determinations, noisy-recording detection and an average-beat calculation. It does not mention PVC labels, bradycardia or tachycardia for those three. AliveCor CEO Priya Abani said the goal is turning daily tracking into “clinically actionable data.”

Clearance Letter Names Apple Watch Series 6 While Release Says Compatible Hardware

The FDA letter, dated October 27, 2026 and filed under 21 CFR 870.1425 as a Class II programmable diagnostic computer, mirrors that split. Its comparison table gives the fullest analysis set, including bradycardia, tachycardia, PVC analysis and beat classification, to wet-electrode ECGs and to Apple Watch Series 6. For Galaxy Watch, Fitbit Sense, ScanWatch and AliveCor’s own KardiaMobile devices, it lists sinus rhythm and AFib analysis only.

The release refers only to compatible devices and does not say which Apple Watch models qualify. Whether newer Apple Watch generations receive the same feature set is not stated in the materials available as of Monday afternoon.

The letter also shows what the engine would not do. The FDA named AliveCor’s own KardiaAI V2 (K201985) as the predicate device, and the subject device does not list the predicate’s sinus-rhythm-with-wide-QRS or sinus-rhythm-with-supraventricular-ectopy detections. Those are two of the six determinations AliveCor sells on KardiaMobile.

Why Premature Beats Trigger False AFib Alerts and How Beat-by-Beat Labels Target Them

A PVC is an extra beat that starts in the ventricles instead of the heart’s natural pacemaker. It arrives early and is followed by a longer pause, and its QRS complex looks abnormal. Up to 80% of apparently healthy people show occasional PVCs on 24-hour monitoring.

That matters because AFib checks on smartwatches look for irregular timing. In a 734-patient hospital study in the Canadian Journal of Cardiology, patients with PVCs were about three times more likely to receive a false-positive AFib result from the Apple Watch ECG app. Lead investigator Marc Strik of Bordeaux University Hospital’s LIRYC institute said automated detection relies on cycle variability, which PVCs inflate. The same study found roughly one in five patients received no automatic diagnosis.

PulseAI’s pitch on Apple Watch is to classify each beat, not just judge overall regularity. Recordings go to the cloud for automated analysis and structured results return to the connected app, and the FDA’s 510(k) summary describes the software as a library meant to be integrated into other companies’ software through an API, the same role the KardiaAI API played for the predicate.

The tradeoffs are concrete. Cloud processing requires secure transmission and compliance with data-protection obligations. The software is not intended for people with pacemakers, ICDs or other implanted electronic devices, and it is not meant for life-sustaining systems, ECG monitors or alarm systems. Whether PVC labels carry clinical weight depends on how often they occur: a burden above 10% of beats is the minimum threshold for PVC-induced cardiomyopathy, while PVCs usually pose no danger. The release does not say how the labels would be presented to subscribers.

Apple’s own ECG app sets a different baseline. Its instructions describe over-the-counter use for adults 22 and older, classification of AFib, sinus rhythm or inconclusive, and no recommendation for users with other known arrhythmias. The document does not mention PVCs.

Inconclusive Readings Fell From 17.6% to 1.4% in Company-Linked Hospital Study

The headline figure is a 92% reduction in inconclusive interpretations versus Apple’s native algorithm. It comes from a prospective study of 296 analyzable participants at an Irish teaching hospital between November 2022 and February 2023, published in Heart Rhythm. First-attempt inconclusive results fell from 17.6% to 1.4%.

Participants were admitted for ablation, cardioversion or electrophysiology testing, and AFib prevalence in the cohort was 21.6%. Wikipedia’s AFib entry puts prevalence at about 2% to 3% of people in Europe and North America as of 2014, so the hospital group is far from everyday wearers.

How inconclusive readings are counted changes the story. Apple Watch sensitivity and specificity were 96.1% and 97.9% when inconclusive outputs were excluded, but 78.1% and 81.0% when they were counted as incorrect. PulseAI’s algorithm scored 98.4% and 96.1% under that stricter count.

Three caveats apply. The authors include PulseAI founder Alan Kennedy, and PulseAI Ltd is listed as a funder, though physicians blinded to PulseAI’s results verified the Apple Watch interpretations and the protocol was preregistered. The study used Apple Watch recordings, so it does not establish performance on Galaxy Watch, Fitbit Sense or ScanWatch. And it does not show fewer strokes, hospital admissions or lower costs. The release says AFib sensitivity improved but gives no figures.

Does Prescription-Only FDA Clearance Cover Consumer KardiaCare Subscribers?

The release says users with a KardiaCare subscription would receive the determinations through the Kardia App, and it describes the purpose as turning wearable data into clinical consultations and specialist overreads. It does not say a prescription or clinician is required, and it gives no clearance number.

The FDA letter points the other way on the label. A 510(k) clearance reflects substantial equivalence to a predicate for a stated intended use, and one analysis of the PulseAI clearance cautions that it is not unrestricted authorization for every wearable or application. The FDA documents also describe the subject as narrower than the predicate, which covers any ambulatory ECG device, while PulseAI’s software is limited to Lead I or II wet-electrode and compatible dry-electrode inputs.

The clearance does not, by itself, show whether a person could simply subscribe or would need a clinician in the loop. For context, AliveCor’s Apple Health page lists KardiaCare at $99.99 per year or $13.99 per month, but the release gives no price for the new determinations.

KardiaCare Already Excludes Apple Watch Users From Six Arrhythmia Determinations

A note on AliveCor’s KardiaCare page says not every KardiaCare feature is provided to Apple Watch users, and gives KardiaMobile’s six FDA-cleared determinations as an example: AFib, bradycardia, tachycardia, PVCs, sinus rhythm with supraventricular ectopy and sinus rhythm with wide QRS. PulseAI’s analysis would add some of those labels for watch recordings inside the same app, but the FDA letter omits the last two.

The gap has a history. AliveCor’s KardiaBand became the first FDA-cleared medical device add-on for Apple Watch in 2017, and AliveCor ended sales in 2019. After years of patent and antitrust fights, the USPTO invalidated AliveCor’s three patents in December 2022, a federal judge ruled for Apple on the antitrust claims in February 2024, and the Ninth Circuit affirmed that outcome in January 2026.

PulseAI Names AliveCor First Partner After FDA Used AliveCor’s Own Algorithm as Predicate

PulseAI CEO Alan Kennedy called this its first partnership for the company’s AI ECG technology. Belfast-based PulseAI announced a seed round on January 12, 2026 led by Innovation Ulster, with The Mortara Group and The Francis Crick Institute participating; the amount was not disclosed.

AliveCor says it has recorded more than 400 million ECGs and that Kardia devices detect six common arrhythmias in about 30 seconds. It also says its Kardia 12L system is cleared for 39 cardiac conditions, though that system is for healthcare providers only.

Five Open Questions: Price, Exact Date, Models, Regions and Sensitivity Figures

  • Price: The release says only that a KardiaCare subscription is required.
  • Exact launch date: First half of 2027 is the only timing given.
  • Models and regions: The release does not specify models or regions, including whether all four platforms launch together.
  • Regulatory path: The release does not explain the regulatory path for delivering a prescription-indicated clearance through a consumer subscription.
  • Sensitivity data: The AFib sensitivity gain comes without numbers.

Until those answers arrive, the practical read is narrow: an Apple Watch owner could gain PVC labels, and Galaxy Watch, Fitbit Sense and ScanWatch owners would get AFib and sinus-rhythm calls, though the evidence for fewer inconclusive readings comes from Apple Watch recordings. The open question for every user is whether the prescription indication applies before counting on the service. The best evidence for the 92% claim comes from hospital patients with a high AFib rate, tested on Apple Watch recordings only.


Frequently Asked Questions

Will Galaxy Watch, Fitbit Sense or ScanWatch users get PVC detection from this deal?

Not according to the release or the FDA letter. The release lists PVC labels, bradycardia and tachycardia for Apple Watch only, while the FDA table gives the other three watches sinus rhythm and AFib analysis, noisy-recording detection and an average-beat calculation.

Do you need a prescription to use PulseAI’s analysis through the Kardia App?

That is unresolved. The FDA clearance is for prescription use by adults under a physician’s care and does not select over-the-counter use, but the partnership release does not mention prescriptions.

How does AI reduce inconclusive smartwatch ECG readings?

Apple’s app returns an inconclusive result when it cannot classify a recording, for reasons such as a noisy signal, an arrhythmia it does not recognize or certain physiological conditions. PulseAI’s algorithm analyzed the same recordings in the hospital study and left 1.4% inconclusive on the first attempt, versus 17.6% for Apple’s native algorithm. PulseAI says its deep learning algorithms were trained on millions of ECGs.

When will the AliveCor and PulseAI features launch, and what will they cost?

The companies say the features will be available in the first half of 2027 to KardiaCare subscribers, with no price or firm date given.

Source link