The future of exoskeletons may look less like Iron Man—and more like the world’s smartest brace.
A terrific new Economist story, “Humble exoskeletons work better than combat supersuits”, captures a shift that has been hiding in plain sight. The grand, full-body robotic suit—the cinematic machine that turns a human into a superhero—has repeatedly proved difficult to make comfortable, efficient and useful in everyday life. Meanwhile, much more modest systems are quietly succeeding: powered gloves, lumbar supports, knee and ankle devices, arm orthoses and other machines that help one part of the body do one job better.
Less armor. More augmentation.
The Economist describes a number of these targeted systems: a powered glove helping restore grip after major hand injury; powered and passive back-support systems reducing the burden of lifting; arm and hand exoskeletons responding to faint muscle signals; and lightweight knee, hip and ankle systems beginning to move from rehabilitation into work, sport and even “powered clothing.”
The lesson is wonderfully simple: a Swiss Army knife strapped to the whole body may lose to a screwdriver placed exactly where the torque is needed.
This is not a new idea to those of us who have lived around wearable robotics for a while. The Wearable Robotics Association (WearRA) has built a community around exactly this transition—from spectacle to practical human augmentation. I have had the pleasure of engaging with this group for years, including at WearRAcon, where we discussed the future of human-machine interfaces.
Tom Sugar and the ASU “tinkerer’s paradise”
My longtime colleague Tom Sugar at Arizona State University and his coworkers have been building this future for decades. Tom founded ASU’s Human Machine Integration Laboratory in 1999 and has focused on compliant wearable robots, prosthetics and orthotics designed to improve human mobility rather than simply overwhelm it with motors. His ASU group has produced systems ranging from ankle technologies to wearable devices that assist walking, running and load carriage. Tom is also a leader of WearRA, helping connect the engineering, clinical and industrial sides of this field.
Tom, Sangram Redkar and colleagues also pulled me and Bijan Najafi into work around the second edition of the Wearable Exoskeleton Systems book. That ecosystem—engineers, clinicians, rehabilitation scientists and end users—is exactly where the useful future tends to emerge.
Charles Liu: closing the loop between brain and machine
At USC, my longtime friend and partner Charles Liu, director of the USC Neurorestoration Center, is pushing the same idea in a different and even more ambitious direction: not merely putting a machine around the body, but closing the loop between intention, movement and sensation.
In 2026, the USC/Caltech/UCI team reported a proof-of-concept two-way brain interface with wearable robotic legs: motor-cortex signals are decoded to drive the exoskeleton, while sensors in the robotic system trigger stimulation of sensory cortex so that the wearer can receive step-related sensation. We have followed this arc on the blog for years, including Charles’s work on human augmentation and, more recently, bidirectional robotic rehabilitation.
Repair, regeneration and replacement—revisited, redux… again
This also feels like another chapter in an idea we first framed in 2012 and then formally revisited in 2025: Repair, Regeneration, and Replacement—Redux. Charles Liu, Bijan Najafi, Wei Gao, David Klonoff and I argued that the boundaries between biology and engineering were becoming increasingly porous.
Exoskeletons sit beautifully across all three R’s.
- Repair: assist a weak or injured system while preserving function.
- Regeneration: keep someone safely active while tissue, nerve or muscle recovers—and perhaps provide the repeated mechanical and neural input that recovery needs.
- Replacement: when biology cannot fully recover, provide an engineered layer that restores lost capability.
And in limb preservation, this principle feels particularly familiar. We already know that the goal is not to make the entire body invulnerable. It is to redistribute load, protect vulnerable tissue, preserve mobility and keep the person moving through the world. In that sense, an intelligent exosuit is a cousin of offloading—only dynamic, sensor-rich and potentially adaptive from step to step.
We explored a related idea years ago in our work on exotendon/exoskeleton footwear and amputation prevention, and even earlier in discussions of unpowered exoskeletons that reduce the energetic cost of walking.
The important caution: moving force is not the same as eliminating it
There is an important biomechanical warning here. An exoskeleton does not make load disappear; it moves load somewhere else. The assisted joint may be happier while another joint—or skin interface—quietly pays the bill. That is why sensors, human factors, tissue tolerance and longitudinal measurement matter as much as motors and batteries.
For people with diabetes, neuropathy, frailty or healing wounds, that point is not academic. The most successful wearable robots may eventually be the ones that know not only how much assistance to give, but when, where and for how long—a closed-loop combination of robotics, physiology and real-time risk sensing.
So perhaps the future is not a supersuit at all. It is a quiet layer of intelligence between the person and the world: a glove that grips when needed, an ankle that gives back a little energy, a back support that shares a lift, a robotic leg that listens to the brain—and eventually, systems that protect tissue before injury announces itself.
That may be less cinematic than Iron Man. But it is probably much more useful.
Leave a Reply