Smartwatches are getting higher at recognizing well being patterns. Blood sugar could possibly be one of many subsequent massive assessments.
For now, the caveat issues: smartwatches can not measure blood glucose on their very own in an FDA-authorized, noninvasive method. The FDA has warned customers to not depend on smartwatches or sensible rings that declare to measure blood sugar with out pores and skin penetration, as a result of inaccurate readings might result in harmful therapy selections.
That warning doesn’t apply to smartwatch apps linked to accepted steady glucose monitoring techniques, or CGMs, that measure glucose immediately. According to the FDA, these apps are totally different from wearables that declare to estimate blood glucose on their very own.
That distinction is the place the story will get attention-grabbing. The smartwatch could not change a glucose monitor anytime quickly, however AI might assist wearables turn into higher companions to accepted CGMs, meals logs, sleep information, exercise information, and heart-rate traits.
Smartwatches could get higher at explaining glucose patterns
Some CGM customers can already view glucose readings on suitable wearables. Dexcom, for instance, introduced direct Apple Watch connectivity to its G7 CGM system in 2024, permitting customers to see readings from the accepted sensor with out counting on an iPhone because the intermediary.
That makes Apple Watch an necessary bridge between accepted medical sensors and on a regular basis well being monitoring. Apple already positions the watch round coronary heart well being, exercise, sleep, and security options, whereas CGM makers provide the glucose information. AI might finally assist join these streams into extra helpful explanations, reminiscent of whether or not sleep, train, or meal timing seems linked to a consumer’s glucose patterns.
The key level for customers is easy: Apple Watch might help floor glucose information from suitable CGMs, however it’s not an FDA-authorized noninvasive blood glucose monitor by itself.
The subsequent step is not only displaying the quantity. It is explaining what the quantity means.
AI techniques might assist hyperlink glucose spikes and dips to sleep, train, stress, meal timing, and prior patterns. Instead of merely displaying that blood sugar rose after lunch, a future well being assistant would possibly level out that the spike was greater than typical after poor sleep or a late exercise.
That form of sample recognition is already a part of the broader CGM worth proposition. Dexcom says its Clarity software program helps customers and clinicians assessment glucose traits, patterns, and statistics over time. Abbott’s FreeStyle Libre techniques equally emphasize glucose traits and real-time information by way of linked apps.
AI might assist predict highs and lows earlier
Researchers are already testing AI fashions that use CGM information to forecast near-future glucose adjustments. A 2024 arXiv preprint proposed a big sensor mannequin educated on 1.6 million glucose information from sufferers with diabetes to enhance short-term glucose prediction.
That doesn’t imply the expertise is able to information therapy selections from a smartwatch. But it does present the place the sector could also be heading.
Other analysis is shifting in the identical route. A 2024 paper on glucose forecasting described an AI-powered system that mixes glucose information with behavioral and physiological alerts, together with weight-reduction plan, treatment, bodily exercise, sleep, and stress. A more recent 2026 preprint proposed a CGM basis mannequin educated on greater than 109,000 hours of glucose recordings.
For smartwatch customers, the long-term promise is extra helpful alerts. Instead of warning solely when glucose is already excessive or low, an AI-powered system might finally detect {that a} consumer is trending towards an issue and counsel motion earlier.
That distinction issues. A prediction provides customers time to reply. A late alert could solely verify what’s already occurring.
Wearables might personalize well being recommendation
AI may assist flip…
