Do Exoskeletons Actually Work? Here’s What the Clinical Trials Show.

Somewhere in a rehabilitation hospital in South Korea, a 62-year-old stroke patient stands up from a wheelchair, straps into a powered exoskeleton, and takes twelve steps across the room. It is the first time she has walked in four months. The physical therapist does not touch her. The exoskeleton reads the force in her legs, predicts her intended motion, and delivers assistive torque to her hips in real time. After twenty minutes, she sits back down. Her heart rate is elevated but controlled. She is crying, but not from pain.

Meanwhile, in a clinical research lab in California, investigators are analyzing data from 93 stroke patients who completed a four-week trial comparing robotic gait training to conventional therapy. The results show that both groups improved. But the exoskeleton group reported something unexpected: their quality-of-life scores rose 25%, compared to 14% in the control group. The difference was not just statistically significant. Patients said they felt more motivated, less fatigued, and more confident performing daily activities.

These are not speculative results from a press release. They are peer-reviewed clinical trials published in 2025, and they answer a question that has followed wearable robotics for two decades: do exoskeletons actually work?

The short answer is yes. The longer answer is that the evidence exists, the regulatory pathways are clear, and the technology is ready—but most people still cannot access it. Not because it does not work. Because it costs as much as a car.

The Evidence Is Real

For years, the gap between exoskeleton prototypes and clinical deployment was wide enough to justify skepticism. Early devices were heavy, uncomfortable, and showed modest improvements over standard therapy. Researchers published papers on algorithms. Clinicians remained unconvinced.

That has changed. Between 2024 and 2026, multiple randomized controlled trials demonstrated measurable, clinically meaningful improvements in patients using powered exoskeletons for rehabilitation.

Stroke: Quality of Life Gains

A multicenter randomized trial published in Stroke enrolled 93 subacute stroke patients with severe walking impairment. Half received torque-assisted exoskeleton training for 20 sessions over four weeks. The other half received conventional rehabilitation.

Both groups improved their walking function by similar amounts. The difference showed up elsewhere: composite quality-of-life scores (measured by the Stroke Impact Scale) increased by approximately 25% in the exoskeleton group, compared to 14% in the control group. The between-group difference of 10 points was both statistically significant and clinically meaningful—meaning patients could feel the difference in their daily lives.

Patients in the robotic group also reported higher motivation, reduced fatigue during sessions, and greater confidence in performing everyday tasks. These are not trivial outcomes. Rehabilitation is physically and emotionally exhausting. If a device makes the process less grueling while delivering comparable or better functional gains, that matters.

Spinal Cord Injury: Fewer Adverse Events

For people with complete paraplegia due to spinal cord injury, standing and walking require external mechanical support. Traditionally, that meant knee-ankle-foot orthoses (KAFOs)—rigid braces that hold the legs straight while the user swings forward on crutches.

A randomized crossover trial published in PLOS One compared the ABLE Exoskeleton to conventional KAFOs across 100 gait training sessions for each device. The ABLE exoskeleton proved feasible, safe, and as practical to use as standard orthoses.

The unexpected finding: the exoskeleton group experienced 17 adverse events, while the KAFO group experienced 31—a 46% reduction. The exoskeleton’s active control systems appeared to reduce risk rather than increase it, contrary to the intuition that adding motors and electronics would introduce new failure modes.

Upper Limbs: The Other Half of the Problem

Most exoskeleton research focuses on walking. But roughly 80% of stroke survivors experience upper limb impairment, and recovering hand function is often more critical for independence than recovering gait. You can use a wheelchair. You cannot easily replace your hands.

A trial combining the ReHand robotic system with brain-computer interface, published in Frontiers in Human Neuroscience in 2025, showed enhanced hand motor recovery compared to conventional therapy. The system uses a soft robotic glove that replicates the movements of a healthy hand onto the impaired hand in real time, guided by signals from embedded sensors.

Another study on powered hand exoskeletons tested an 8-week protocol: 30-minute sessions twice per week, split between 15 minutes of conventional therapy and 15 minutes with the exoskeleton. The experimental group showed significant improvements in upper extremity function and quality of life.

Systematic reviews published in 2024 and 2025 confirm that upper limb robotic rehabilitation, including exoskeletons, produces measurable functional gains—though the effect sizes vary depending on device type, training intensity, and patient selection.

The gap in public awareness is striking. Most people have heard of exoskeletons for walking. Far fewer know that robotic gloves exist, let alone that they have clinical evidence supporting their use.

The Regulatory Path Exists (But It Is Not Simple)

Clinical efficacy is necessary but not sufficient. For a medical device to reach patients, it must navigate regulatory approval, meet safety standards, and secure reimbursement. Exoskeletons have done this. The process is slow, expensive, and far from universal—but the pathway exists.

FDA Clearances in 2025

Three exoskeletons received or expanded FDA clearance in 2025:

ReWalk 7 (Lifeward, March 2025): The seventh-generation wearable exoskeleton received 510(k) clearance for individuals with spinal cord injury. It allows users to stand, walk, and navigate stairs with the device providing lower-body support.

EksoNR for Multiple Sclerosis (Ekso Bionics, November 2025): The FDA expanded indications to include MS rehabilitation patients, adding to existing approvals for stroke and SCI. The device is designed for use in clinical rehabilitation settings, not home use.

Atalante X (Wandercraft, November 2025): FDA expanded indications to include MS patients and a broader range of spinal cord injuries. Atalante X is self-balancing, meaning users do not need crutches—a significant usability improvement.

All three devices are classified as Class II medical devices under the 510(k) pathway, which requires demonstrating substantial equivalence to an existing approved device plus evidence of safety and effectiveness. The first exoskeleton clearance came in 2014. A decade later, the regulatory path is well-established, and manufacturers know what data the FDA expects.

What FDA Approval Actually Means

An FDA clearance is not a guarantee of efficacy. It is a statement that the device is reasonably safe and substantially equivalent to existing devices for the approved indications. The agency does not compare the device to the best available treatment—only to predicate devices or to a reasonable safety standard.

That distinction matters. An exoskeleton approved for stroke rehabilitation does not have to outperform conventional physical therapy. It only has to be safe and show some benefit. The clinical trials described above go beyond that standard by demonstrating comparable or superior outcomes, but the regulatory bar is lower than many people assume.

Indications Are Narrow

FDA-approved indications for exoskeletons include hemiplegia due to stroke, paraplegia due to spinal cord injury, and (as of 2025) multiple sclerosis. Devices approved for one indication cannot legally be marketed for another without additional clearance.

If you have Parkinson’s disease, cerebral palsy, or muscular dystrophy, there may not be an FDA-cleared exoskeleton for your condition—even if the device would physically work. The regulatory system moves slower than the technology.

The $70,000 Question

FDA clearance opens the door. Insurance coverage determines who walks through it.

National Reimbursement Policy

In 2024, the Centers for Medicare & Medicaid Services (CMS) issued a national reimbursement policy for exoskeleton use by qualifying beneficiaries. Before this policy, patients or hospitals had to seek individual coverage determinations—a slow, uncertain process that often resulted in denial.

The national policy standardizes eligibility criteria and ensures that Medicare will pay for exoskeleton therapy when medically necessary. For patients over 65 or those with qualifying disabilities, this removes a major barrier. For younger patients with private insurance, coverage remains inconsistent. Some insurers follow Medicare’s lead. Others do not.

Out-of-Pocket Reality

For personal ownership, the numbers are stark. The ReWalk Personal 6.0 costs around $70,000. The SuitX Phoenix runs approximately $40,000. High-end rehabilitation systems for hospitals can exceed $150,000. Batteries need replacement every 2-3 years at $2,000-5,000 per pack.

Compare that to the median U.S. household income of roughly $75,000. Even with insurance covering part of the cost, out-of-pocket expenses can run into the tens of thousands. For most families, that is not feasible.

The price gap to consumer products underscores the regulatory premium. The Hypershell X Pro, a consumer hiking exoskeleton, retails for £1,199 (roughly $1,500). It provides gentle walking assistance using similar motor and battery technology. The 50x price difference to medical-grade devices reflects the cost of FDA clearance, clinical trials, safety certifications, durability testing, and ongoing regulatory compliance—not just better hardware.

A New Model: Rent, Don’t Buy

ExoAtlet, a European medical robotics company, launched a Hardware-as-a-Service (HaaS) program in 2025. In partnership with WeCare MedLease, the company offers exoskeleton rentals to clinics, researchers, and individuals across Europe, Africa, and the Middle East.

The rental rate: €150 per day for professional facilities. That is €54,750 per year if used daily—still expensive, but structured as an operating expense rather than a capital purchase. Hospitals can try the technology without committing six figures upfront. Insurance reimbursement can potentially cover per-session rental fees more easily than lump-sum device purchases.

This model shifts risk from buyer to vendor. The manufacturer handles maintenance, upgrades, and repairs. If the device becomes obsolete, the clinic is not stuck with a $100,000 paperweight. For complex systems that require integration with clinical workflows and trained staff, this makes adoption faster and less risky.

Other companies are exploring similar approaches. Cyberdyne’s Robot Suit HAL is available for lease to medical and welfare facilities in Japan. The exoskeleton market, valued at $0.56 billion in 2025, is projected to reach $2.03 billion by 2030—a 19.2% compound annual growth rate. Much of that growth will depend on whether HaaS models prove sustainable.

What Patients Actually Experience

The clinical trial data is impersonal by design. Outcomes are measured in scales, percentages, and p-values. But behind every data point is a person relearning how to move.

A patient using the Vilpower exoskeleton, which completed trials with 40 participants in 2025, told reporters: “It feels like me again.” The device, designed for full-arm assistance, is expected to launch in Norway in the first half of 2026.

A hiker using the consumer-grade Hypershell X Pro wrote in early 2025: “Much more energy and feel stronger on the steepest mountains,” with plans for 100-mile and 600-mile treks. The exoskeleton is not treating a disability—it is augmenting normal function. But the user’s description echoes what rehabilitation patients report: the feeling of capability returning.

These testimonials matter, but they also obscure the less visible struggles. Research on home-based exoskeleton use identifies persistent challenges: putting the device on and taking it off (donning and doffing) can take 10-20 minutes even with practice. At-home safety requires fall detection, remote monitoring, and emergency protocols that most devices do not yet provide. Patients want independence, but independence requires infrastructure that is still being built.

A participatory study with older adults designing soft exoskeletons emphasized the need to involve users directly in the design process. What engineers think is comfortable and what users actually find comfortable often differ. What looks simple in a controlled lab becomes complicated in a messy living room with furniture, pets, and carpets.

The technology works in clinical settings with trained staff. Making it work at home, unsupervised, for months or years—that is a different engineering problem, and one the field has not fully solved.

The Inequality Nobody Wants to Talk About

Even if exoskeletons become medically effective, regulatory-approved, and insurance-covered in wealthy countries, most of the world will not have access.

A device that costs as much as a luxury car is inaccessible to the majority of the global population. Multiple studies have raised concerns that exoskeleton technology may benefit only wealthy patients in developed countries, thereby exacerbating existing health inequalities rather than reducing them.

A 2025 survey of construction industry stakeholders rated the inaccessibility and unaffordability of exoskeletons as “Very Critical,” with 95.24% agreement that unequal access represents a significant ethical concern. That consensus spans employers, workers, safety officers, and researchers.

The rental model helps, but €150 per day is still expensive in absolute terms. Over a year of daily use, that is more than most households in lower-income countries earn annually. Even in Europe, sustained personal use remains out of reach for most individuals without institutional or insurance support.

There is also a dependency risk. Ethicists have warned that users with mobility impairments could become dependent on devices that they cannot afford to maintain long-term. If a research study ends, if insurance stops covering the device, if the company goes out of business—the user loses not just the technology but the independence it enabled. That creates a vulnerability that did not exist before.

A comprehensive ethical assessment published in Artificial Organs in 2024 argued that exoskeleton deployment in clinical, industrial, and military contexts raises distinct ethical issues around autonomy, privacy, mandatory use, and long-term dependency. The technology is not ethically neutral. How it is deployed determines whether it reduces inequality or entrenches it.

Industrial Reality Check

Medical exoskeletons get the most attention, but industrial exoskeletons represent a larger and faster-growing market segment. These are not devices for people with disabilities. They are tools to reduce injury and fatigue in physically demanding jobs.

Airbus reported in June 2025 that paint shop operators experienced a 10-40% reduction in shoulder and upper back muscle strain when using exoskeletons for sanding tasks. The company has deployed the technology across multiple production facilities.

A scoping review of occupational exoskeleton use found that 42% of studies focused on automotive industry workers, 17% on logistics, and 17% on healthcare workers. The evidence shows measurable reductions in muscle activity, fatigue, and injury risk in static or repetitive tasks.

But the same review noted significant limitations. Most exoskeletons are designed for controlled environments—factories, warehouses—where tasks are predictable and infrastructure supports the technology. Construction sites, with uneven terrain, rapid task switching, and unpredictable loads, remain challenging. Exoskeletons can interfere with other personal protective equipment (PPE), such as fall arrest systems and tool belts. And in dynamic tasks requiring frequent posture changes, exoskeletons can hinder performance rather than help.

The industrial use case is simpler than medical rehabilitation in some ways: workers are healthy, the tasks are defined, and the ROI can be calculated in dollars (reduced injury claims, less worker’s compensation, higher productivity). But even there, adoption is uneven. A human-centered assessment of construction exoskeletons concluded that widespread deployment will require better integration with existing safety systems, clearer evidence of long-term benefits, and designs that account for the variability of real-world work.

Soft vs Rigid: It Depends

One of the most common questions from potential users: should I use a soft exoskeleton or a rigid one?

The answer is not about which technology is better. It is about what you are trying to do.

A 2024 benchmark study tested four commercial back exoskeletons in an industrial workplace: two soft (Darwing, Auxivo) and two rigid (Laevo, Paexo). Workers rated Darwing (soft) as the most comfortable overall. But they also rated Auxivo (soft) as the least comfortable, demonstrating that the soft-versus-rigid distinction matters less than the specific design.

Laevo, a rigid device, achieved the best trade-off between muscle activity reduction and comfort across both static and dynamic tasks. The takeaway: design quality matters more than material category.

A 2025 study comparing soft active and rigid passive exoskeletons during lifting tasks found that the active soft exosuit significantly reduced both mean and peak muscle activity (measured by EMG), while the passive rigid device showed only modest reductions, particularly during unloaded tasks.

The general principle, summarized in a 2025 review: rigid devices excel when high forces, precise positioning, or high-speed dynamics are required. Soft devices excel when portability, long-duration wear, and comfort are priorities.

For stroke rehabilitation in a clinic, where a therapist helps the patient don the device and sessions last 30-60 minutes, a rigid exoskeleton’s superior force transmission may outweigh its bulk. For an industrial worker wearing a back-support device for an eight-hour shift, a soft exosuit’s comfort and breathability may matter more than maximum load capacity.

There is no universal answer. There are only trade-offs.

What This Actually Means

Here is what I believe the data supports:

Clinical exoskeletons work, and the evidence is stronger than it has ever been. Multiple randomized controlled trials from 2024-2026 show measurable improvements in quality of life, functional outcomes, and safety. Stroke patients walk better and feel more confident. Spinal cord injury patients train with fewer adverse events. Upper limb devices help people regain hand function. The skepticism that was reasonable five years ago is less defensible now.

The regulatory pathway is clear, though slow. FDA 510(k) clearance is achievable for companies with the resources to conduct trials and compile submissions. Three new approvals in a single year (2025) signal that the process is maturing. The existence of a national CMS reimbursement policy removes a barrier that previously blocked many patients from accessing the technology. This is progress, even if it is incremental.

Access remains the largest barrier. Devices that cost $40,000-150,000 will never reach most people who need them, regardless of clinical efficacy or regulatory approval. The rental model helps but does not solve the problem—€150 per day is affordable for institutions, prohibitive for individuals. Until manufacturing costs drop by an order of magnitude, exoskeletons will remain assistive technology for the few, not the many.

Inequality is a design problem, not just an economic one. If exoskeleton development continues to focus on high-end medical devices for wealthy markets, the technology will entrench disparities. Alternative approaches—passive devices, lower-cost powered assists, open-source designs—could broaden access, but they receive far less R&D funding than flagship products. The field is optimizing for performance and regulatory approval, not affordability and global scalability.

Industrial adoption will likely outpace medical adoption. Workers in factories and warehouses do not need FDA clearance to use exoskeletons. Employers can calculate ROI directly: fewer injuries, lower insurance costs, higher productivity. The regulatory complexity and reimbursement uncertainty that slow medical deployment do not apply. The market projections reflect this: industrial and commercial exoskeletons are expected to drive most of the 19% annual growth through 2030.

The technology is no longer the bottleneck. The system is. Exoskeletons can restore walking, reduce muscle strain, and improve quality of life. The evidence exists. What does not exist is a healthcare and economic system structured to deliver that technology equitably. Solving that requires policy, not engineering—and policy moves even slower than hardware.

There will be no single moment when exoskeletons “arrive.” There will be gradual expansion: more FDA approvals, broader insurance coverage, lower costs, better rental models, and (eventually) devices simple enough for unsupervised home use. The stroke patient in South Korea will keep attending supervised sessions at the hospital because the device is too complex and expensive to take home. The construction worker in Germany will keep using a passive back-support exoskeleton because the powered version does not yet integrate well with fall arrest harnesses. The clinical trials will keep showing positive results, and most people will keep not having access.

The evidence says exoskeletons work. The system says most people cannot have one. Both statements are true, and the gap between them is not closing quickly.