Ultrahuman to Test Wearable-Based Biological-Age Model in Abu Dhabi
Ultrahuman has signed a two-year memorandum of understanding with the Abu Dhabi Public Health Center (ADPHC) to develop and validate a biological-age model intended for use across the United Arab Emirates.
The Bengaluru-based health tech startup, known for its smart rings, announced the agreement on Oct. 12, 2026. The model combines clinical biomarkers, measurable signals from blood and other lab tests, with continuous data from wearable devices, according to TechCrunch. The two-year period is set aside for research and clinical validation before any population-scale rollout.
The model is described as wearable-agnostic. That means it is designed to work across brands and form factors, rather than being locked to one device. It is built to ingest data from rings, wrist-worn devices, straps and patches alongside clinical biomarkers. Individuals would eventually access their health age through government digital-health platforms including TAMM and Sahatna. Validation comes first. Availability across the UAE is framed as an eventual outcome, not an immediate launch.
Privacy architecture was addressed directly in the announcement. Ultrahuman CEO Mohit Kumar said the model will not be trained on UAE data and that Ultrahuman will not have access to personally identifiable information through the program. In plain terms, validating a model on a population does not always require training it on that population, and controls around names and IDs are separate from handling de-identified physiological signals with personal details removed.
Ultrahuman has sold around 800,000 smart rings globally so far. The UAE is among its 10 largest markets and accounts for about 4% to 5% of overall revenue. The company reports active ring users in more than 180 countries and 130,723,925 nights of sleep tracked to date.
The current lineup shows what kind of continuous signal is available. Ultrahuman states its Ring PRO provides up to 15 days of battery life on a single charge and up to 250 days of on-ring data storage, is made with fighter jet-grade titanium, and uses redesigned PPG, the light-based sensor that measures heart rate, with improved signal quality during sleep and recovery, according to Ultrahuman. Its Ring AIR weighs 2.4g and is 2.4mm thick, offers 4-6 days of battery life, and offers precision health tracking and core features without a mandatory subscription. Long battery life and on-device buffering reduce gaps in longitudinal data, the months-long record used by any aging model that relies on resting heart rate, heart rate variability, sleep architecture and recovery trends rather than spot checks. Heart rate variability is the small change in time between beats. Sleep architecture is how the night divides into light, deep and REM sleep.
Ultrahuman already operates a consumer-facing aging metric called Ultra Age. The company describes it as a science-backed aging marker powered by physiology, built as a single insight combined from Brain Age, Pulse Age, and a Blood metric, according to the Ultrahuman Blog. The ADPHC collaboration is distinct from that product. It pairs wearable streams with clinical biomarkers and subjects the combination to formal validation.
There is parallel activity in the same geography. Croatian-founded GlycanAge expanded access to its biological-age test in Abu Dhabi on Oct. 6, 2026, according to SeeNews. That approach is glycan-based and lab-driven, while the Ultrahuman-ADPHC effort centers on continuous wearable data fused with clinical markers.
The broader context for technologists working on clinical-grade wearables is that the hard problems sit outside the ring itself. They sit in normalization, drift and generalization. A wearable-agnostic model must reconcile different sampling rates, sensor placements, motion artifacts and firmware filtering across rings, watches, straps and patches. It must then align those noisy, high-frequency streams with low-frequency, high-precision lab values without overfitting to one device or one cohort, which is when a model learns quirks of its training set too well and fails elsewhere.
In my view, the two-year validation window is the most encouraging part of the announcement. Consumer biological-age scores have proliferated faster than their validation. A public-health partner with access to clinical biomarkers and a path to TAMM and Sahatna integration offers a route to test calibration, stratification and longitudinal stability at a scale startups cannot achieve alone. Calibration is whether a predicted age matches measured risk. Stratification is whether the score sorts people into the right risk groups. The decision not to train on UAE data places extra weight on external validity testing. The model will need to hold across different demographics, comorbidities, medications and environmental conditions.
Looking ahead to what success would enable, a trusted health-age number inside existing government platforms could give clinicians a compact longitudinal summary and give individuals a feedback signal tied to sleep, activity and cardiometabolic inputs. I have watched my own children ignore detailed charts but respond to a single number that moves. Technology often lands that way. The abstraction succeeds when the metrology, the science of accurate measurement, underneath is sound, and that is what the next two years are meant to establish.
