From smart footwear to haptic gloves: how Human-Machine Augmentation technology is reshaping ground-up innovation Let’s imagine three everyday situations. A warehouse worker finishes a full shift of lifting without hurting their back. A stroke patient walks through a rehab clinic with a powered frame that moves with him. An engineer working on a car design grabs a virtual door handle and feels it push back. In all three cases, a device worn on the body helps a person do something that would be difficult to do alone. The person is still doing the work. The machine simply helps them do more. This is the idea behind human augmentation. It is different from traditional automation. For decades, automation meant taking work away from people. Assembly lines, factory robots, and conveyor belts did more of the work so people had to do less. Augmentation works in the opposite way. Wearable machines aren’t new. What’s new is that they can now sense and compute enough to follow what you’re doing and react in the moment. The capability picked up a name this year that made it from research papers to the CES stage: physical AI. The last wave of AI generated text, images and video. This one acts — reading a real environment, working out what’s happening, and doing something about it through motors and sensors, close to a person and without hurting them. The market numbers show how physical AI is becoming more common across different industries. Wearable robots and exoskeletons are worth about $6.8 billion in 2026 and could grow to more than $24 billion by 2031. Physical AI is smaller today, but it is growing faster: from less than $1 billion in 2025 to more than $15 billion by 2032. Two key changes made this possible: batteries became cheap enough to power devices for a full shift, and AI chips became small enough to be worn on the body. So, where can this be used now? In hospitals and rehab clinics, on factory and warehouse floors, and in the tools designers and trainees work with. Why it’s happening now: key forces driving human augmentation Here are four reasons companies are starting to put augmentation in their budgets: Ageing populations, and rising demand for rehab and mobility tech. By 2030, one in six people worldwide will be 60 or older. By 2050 that group doubles to 2.1 billion, and the over-80 population triples to 426 million, according to the WHO. Older adults already outnumbered children under five as of 2020, and in Japan over 30% of people are past 60. That drives up demand for assistive, rehabilitative and mobility devices: more strokes and neurological conditions to treat, more people who want to stay independent while there aren’t enough carers and therapists to do all of it by hand. Increasing focus on workforce safety and productivity. In industry, manufacturing and logistics, augmentation solves two problems at once. The first is injury: musculoskeletal disorders (strains, back damage, the wear from lifting and repetitive motion) cause over a million US workplace injuries a year, cost employers around $20 billion in workers’ compensation, and run from $13 billion to more than $50 billion once lost productivity is counted. About half the cases are back injuries, and a single serious one costs $15,000 to $85,000. Moreover, XR helps practicing a dangerous job before doing it for real, which cuts mistakes on the floor and speeds up training. The second is output: with labor shortage, companies want the people they have to get through a full shift without fading or getting hurt. A wearable that takes load off the spine does both: fewer claims and steadier performance. The new generation of immersive technologies. XR and haptic interfaces have moved past gaming demos into training, simulation and product design. Force-feedback gloves now reproduce weight, resistance and texture well enough that someone can handle a virtual part like a real one: rehearse a procedure, test a prototype, walk a trainee through a task with the physical constraints intact. It’s being funded like a serious category, too. AI small enough to run on the device itself. This is the quiet one under the other three. Sensors, sensor fusion and on-device processing let a wearable read what’s happening and adapt in milliseconds, without sending anything to the cloud, which is what makes it responsive, and increasingly connected to the systems around it. Ten years ago that kind of processing needed a server room. Let’s now take a closer look at where physical AI is currently being used. Healthcare: giving people their movement back In medicine, augmentation is about getting back something ordinary that illness or injury took away: standing up, walking across a room, holding a cup, tapping out a message. It usually starts in the rehab clinic. A powered exoskeleton holds a patient upright and moves their legs through the walking motion, so someone recovering from a stroke or a spinal injury can practise the real thing far sooner than they otherwise could. Ekso Bionics’ EksoNR is used this way for stroke, spinal cord injury and brain injury. The therapist dials how much help the device gives from a touchscreen, and it records every step and how the person is progressing. Ekso says its machines have supported hundreds of millions of assisted steps. The evidence is promising: in one review of exoskeleton training after spinal cord injury, 76% of patients could walk with no physical help by the end of a program, and no serious adverse events turned up. Virtual reality solves another problem in the same clinics: boredom. Getting arm and hand movement back after a stroke takes hundreds of repetitions, and the standard exercises are dull enough that people stop pushing, which slows them down. VR turns this into a game, reaching, grabbing and sorting things in a scene that responds, so patients do more of them and stay interested. The headset also tracks every rep, so the difficulty can shift to match how the patient is…