Every major AI company now has a wearable strategy, and the devices are starting to look less like gadgets and more like a coherent vision of what comes after the smartphone. Apple is developing three AI-native wearables simultaneously: smart glasses, a pendant that can be pinned to clothing or worn as a necklace, and a new generation of AirPods with cameras. OpenAI is targeting 40 to 50 million units of AI earbuds in its first year on market. Amazon has already shipped Bee, a $50 always-on wearable that transcribes your day. Oura, which has sold 5.5 million smart rings, just acquired Doublepoint — a startup specialising in gesture recognition — to add a control layer to what was previously a passive health monitor. The category is accelerating faster than most observers expected.
The common thread across these products is context. The premise of a phone-based AI assistant is that you pull it out when you need it, type or speak a question, and wait for a response. The wearable premise is different: the device is always present, accumulating context about what you're doing, who you're talking to, and what you've said, and offering assistance that is continuous rather than episodic. Amazon's Bee records and transcribes its owner's activities throughout the day, building an ongoing summary that can be queried later. Apple's camera-equipped AirPods are designed around visual context — understanding what you're looking at, not just what you're saying.
Scientific American's analysis of the Apple and OpenAI hardware bets frames them as a recognition that the AI assistant wars will ultimately be won at the hardware layer, not the software layer. A model that lives inside Apple's ecosystem, with access to what the camera sees and what the microphones hear, has a structural advantage over one that lives inside an app on your phone. This is why OpenAI is building hardware at all — it's insurance against Apple closing the ecosystem around its own models, which the company has been steadily doing since its deeper integration with Siri.
Oura's Doublepoint acquisition points to a different design question: how do you interact with a wearable that has no screen? The Oura ring is essentially a sensor array — it tracks heart rate, sleep stages, temperature, and activity — but until now it has been passive. Doublepoint's gesture recognition technology, which lets users control devices by tapping their fingers together or making small hand movements, turns the ring into an input device. The implicit vision is a form of ambient computing where the interface lives on your hand rather than in your pocket, and you interact through small, unobtrusive gestures rather than pulling out a screen.
Privacy is the looming question across all of these products, and it is not a small one. A device that is always recording audio and sometimes video, worn close to your face or on your wrist throughout the day, is capturing an extraordinary amount of data about your life — conversations you have, places you go, faces of people around you. Early wearables from companies like Humane and the Rabbit R1 generated significant backlash partly on privacy grounds, partly because the utility didn't justify the intrusion. The companies building this generation of devices have clearly studied those failures, emphasising local processing, user control, and selective recording rather than always-on surveillance.
Whether the privacy architecture holds up in practice is another matter. The history of consumer technology is full of products where the privacy protections that were prominent in the launch materials eroded quietly over time as the business model demanded more data. The wearable AI companies are asking users to trust them with more intimate data than almost any previous consumer device. The market will tell us whether that trust is warranted — but by the time it delivers its verdict, the data will already have been collected.