Hyper growth of AI-enabled wearables
The global wearable AI market size was valued at USD 62.7 billion in 2024 and is estimated to reach USD 138.5 billion by 2029. (M&M)
There are a handful of companies making ultrasound wearables, the clip I saw was from a startup called Sonologi out of Palo Alto.
If you are an athlete, coaches and medical staff can assess real-time cardiovascular health (like arrhythmias), muscle strain, and fatigue levels.
“Imagine a future where real-time, continuous imaging is as ubiquitous as EKG or oxygen monitoring—inside and outside the hospital.”
How does it actually use AI? This device only uses Predictive Machine Learning (ML) for two functions:
Compensating on-the-fly movement to maintain image stability
Recording, monitoring, and alerting potential health issues over time
GenerativeAI: (think ChatGPT)
PredictiveML: (think forecasting models)
Predictive ML focuses on analyzing existing data to forecast outcomes or extract insights. Generative AI creates new data.
Future devices could apply generative AI models to enhance/reconstruct low-quality ultrasound images, filling in missing details or improving resolution.
In the weeds
If you're interested in the science behind this type of AI-enabled wearable, check out this paper.
Hu, H., Huang, H., Li, M. et al. A wearable cardiac ultrasound imager. Nature 613, 667–675 (2023).
Traditional non-invasive imaging methods are limited by bulkiness, and only capture surface signals.
Innovations in device design and materials improve skin-device coupling, allowing for multi-view left ventricular imaging during motion.
By applying a deep learning neural network to extract key information (for example, the left ventricular volume in apical four-chamber view) from the continuous stream of images, you can dramatically reduce variability.
This model generates waveforms for key cardiac metrics, including stroke volume, cardiac output, and ejection fraction.
The technology enables accurate, dynamic cardiac monitoring while biking, running, sleeping, etc.