Exoskeletons Promise to Reduce Injuries, But May Increase Them

The Signal: A Startup's Demand Announcement

At Y Combinator's S26 Demo Day on 10 September 2026, Aaron Edell put $100 million in Ukrainian letters of intent on the screen for Edgerun, a company building a powered device worn around the hips to help soldiers walk while carrying heavy equipment with less physical effort. Edell described these as contracts; Dealroom notes LOIs are prospective demand, not confirmed binding orders or revenue. The supply chain required to turn field-tested prototypes into thousands of units that survive mud, cold, and transshipment hubs does not yet exist at defense scale.

Why the Money Is Moving Now

The Pentagon's budget tells the story.

The fiscal 2025 request sits at $480 million — a 68-fold increase in a decade. Market forecasts show annual growth rates from 16.8% (Straits Research, 2026–2034) to 32.3% (MarketsandMarkets, US 2025–2030) to 17.9% (Fact.MR, global 2026–2036). North America holds roughly 40 percent of the military market; Asia Pacific grows fastest, with China projected at 19.8% CAGR through 2036.

Three operational shifts drive the spend. Urban warfare has gotten denser in places like Mosul and Eastern European cities, forcing troops through rubble under full combat load. The load itself has grown: heavier equipment, longer distances, higher injury rates. Naval operations create a distinct use case: ship boarding and deck movement on unstable surfaces demand balance and lifting support existing gear doesn't provide.

The Pentagon learned from its last big bet. SOCOM's Tactical Assault Light Operator Suit (TALOS), announced in 2013 as a 600–700-pound "Iron Man" concept, was terminated after years of delays. The pivot produced simplified attachments supporting hips, lower back, and legs. The Army's 2024 rotation at Fort Johnson tested the Dismounted Unit Soldier Transport (DUST) and the Silent Tactical Energy Enhanced Dismount (STEED), a motor-assisted wheelbarrow built by Hendrick Motorsports that carries 500 pounds for 15–30 miles. Exoskeleton trials now integrate with the Integrated Visual Augmentation System (IVAS), linking wearable robotics to heads-up displays.

Powered exoskeletons already hold 60 percent share. The funding is real. The operational need is documented. But the gap between prototype and production, including actuators, batteries, sensors, and certifications, remains the bottleneck.

The Manufacturing Gap: Why Prototypes Don't Become Programs

The pattern repeats across every program that has reached advanced development. Biomechanical concepts prove out in labs. Actuators and harnesses get ruggedized for mud, sand, and temperature extremes. Then the program stalls at engineering for maintainability: replaceable modules, onboard diagnostics, standardized interfaces a maintainer can swap at a forward operating base without a PhD. U.S. Army Natick Soldier Center and Army Research Laboratory have documented this progression for decades.

Actuators sit at the center. High-torque, lightweight electric actuation depends on permanent magnets built from neodymium, praseodymium, dysprosium, and terbium. Neodymium and praseodymium form the backbone of NdFeB magnets, the strongest commercially available. Dysprosium and terbium raise the temperature at which those magnets lose coercivity — non-negotiable when a hip actuator runs hot under load. Samarium-cobalt tolerates higher temperatures and resists corrosion better, at higher cost and lower peak energy product. A multi-joint exoskeleton may contain tens to hundreds of grams of NdFeB or SmCo per unit. That sounds modest until you trace the supply chain: a large share of global rare earth separation and magnet manufacturing is concentrated in China, per DOE, USGS, and IEA assessments. Export controls and ITAR restrictions add another layer.

Grain-boundary diffusion now concentrates dysprosium and terbium at grain surfaces, maintaining high-temperature coercivity with less heavy rare earth material. Substitution strategies evolve: ferrite magnets where torque density allows, or passive mechanisms (springs, clutches, elastic elements) requiring no rare earths at all. Recycling from machining waste and end-of-life motors is technically feasible but collection logistics and quality assurance remain hurdles. Diversified sourcing across allies, modular designs accepting multiple motor suppliers, and qualification of alternative magnet grades reduce single-point-of-failure risk — though none eliminates it.

Battery logistics present a second bottleneck. Manufacturer data and peer-reviewed evaluations, drawing on Sarcos systems and Lockheed Martin legacy documentation, cite mission endurance of two to eight hours depending on assist level, terrain, load, and temperature. Those numbers shift significantly under real conditions. The pandemic disrupted electronic component manufacturing across the Asia Pacific semiconductor sector. For a logistics unit, battery swap must be as routine as changing a radio battery. That requires standardized form factors, ruggedized enclosures, and a supply chain delivering replacements without a 16-week lead time.

Sensors and control systems introduce a third tier. Advances in IMUs, force sensors, and model-based control algorithms reduced the "laggy" feel of early prototypes. But when assistance feels unpredictable, users fight the machine, increasing effort — a finding confirmed in Army and NATO STO human performance evaluations. REE-enabled miniaturized components improve signal quality and packaging density, enabling stable control loops. That stability determines whether the exoskeleton reduces injury risk or creates new ones. Poorly fitted, heavy, or unreliable systems can increase perceived exertion and introduce injuries the device was supposed to prevent, per DoD human factors evaluations.

Ruggedization and certification close the loop. Defense primes focus on integrating exoskeletons with comms, power distribution, and the broader soldier-worn ecosystem. That demands environmental qualification (vibration, thermal cycling, ingress protection) that commercial industrial vendors haven't had to meet at scale. Testing and certification verifies compliance with defense and airworthiness requirements, reliability, and operational safety. Advanced robotic systems need significant R&D investment; testing and training for robotic warfare adds cost.

Bottlenecks for the next five to ten years will center on three areas: high-quality magnet supply in sufficient volume, production capacity for ruggedized actuators meeting defense environmental standards, and maintainable battery logistics at the unit level. The company that solves these at production scale, not just in a lab, will field exoskeletons instead of demonstrating them.

The Competitive Landscape: Survivors and the Graveyard

The field is littered with programs that failed to cross the prototype-to-production chasm. The Pentagon has effectively split the problem into distinct mission sets, each demanding a different architecture.

At the heavy end sits Sarcos Defense. The Guardian XO, commercially available since December 2019, earned a 2020 RBR50 innovation award and a Marine Corps Logistics Innovation Office contract for an Alpha unit that year. The spec: a wearable robot for full-shift wear letting one operator lift 200 pounds repeatedly. Sarcos frames it as a logistics multiplier: "one person lifting 200 pounds instead of four or five," targeting Expeditionary Advanced Basing Operations. Leadership stacked with retired U.S. military officers positions the Guardian XO as the "big-rig" in a vehicle-class analogy: not a daily commuter, but the heavy hauler. Delta Air Lines began testing it in January 2020 for baggage and cargo handling, a commercial beachhead funding the military roadmap. Sarcos went public via SPAC in 2022, giving it a capital structure most startups do not have.

Lockheed Martin occupies a different tier: the prime cycling through architectures for two decades. The FORTIS, a passive, unpowered upper-body frame holding heavy tools at chest level, has been tested by the Navy for shipboard maintenance. Separately, Lockheed's ONYX powered lower-body exosuit won a $6.9 million Army enhancement contract in 2018; DEVCOM spokesman Accetta confirmed the initiative ended due to "a number of technical issues" and lack of funding. The pattern: Lockheed sustains passive systems solving specific maintenance ergonomics, but its powered programs haven't survived transition to fielded capability.

The Army's own programs illustrate the powered-passive split. The Soldier Assistive Bionic Exosuit for Resupply (SABER), unveiled August 2022, is an unpowered soft exosuit targeting lower-back strain during artillery resupply. A 2023 study found 90 percent of soldiers using it during field artillery exercises reported increased task performance. The Dephy ExoBoot, tested for years, adds powered ankle assistance for dismounted mobility. Neither is a full-body combat system; both are targeted injury-reduction tools for specific MOS workloads.

Above them looms the TALOS ghost. SOCOM's 600–700-pound powered armor concept was terminated after years of delays — undone by the same power-density and integration challenges haunting every full-body program.

The market has room for multiple architectures: "there won't be one exoskeleton that fits all needs, just like with vehicles," as the Exoskeleton Report noted of Sarcos's positioning. But the graveyard of full-body powered programs is deep, and new entrants face a supply chain problem that has defeated better-funded primes.

The Talent Crisis: Hiring in a Niche Field

The aerospace and defense workforce crisis isn't new, but retiring boomers, pandemic-era layoffs, and tech-sector poaching have sharpened it into a structural constraint. Aviation Week reported in March 2023 that sector attrition climbed to 7.1 percent in 2022, up from 5 percent in 2020 and 4.3 percent in 2017, per an AIA-AIAA-EY survey. Sixty-nine percent of respondents said turnover increased. Roughly a third of industry employees are 55 or older — a "wave of retirements" McKinsey flagged as a dominant factor.

The industry's default for decades has been poaching from large rivals rather than training new cohorts — a habit reinforced by Cold War consolidation and pandemic downsizing. Precision Castparts cut roughly 40 percent of its workforce at the pandemic's height; across the supply base, many suppliers slashed headcount by half. In prior downturns, laid-off workers returned. This time, too many left for good.

The compensation gap compounds it. McKinsey found mechanical and electrical engineering roles outnumbered by software opportunities by as much as 13 to 1. Entry-level software engineers at defense contractors earn about half what Meta, Amazon, and Google pay. Tech talent outflows run potentially twice the inflow rate.

Defense-specific credentials narrow the funnel further. ITAR restrictions, security clearances, and DFARS compliance mean a qualified supply chain professional from automotive or consumer electronics cannot simply step in. The semiconductor industry faces a parallel crunch: Deloitte projects more than one million additional skilled workers needed by 2030, while fewer than 100,000 U.S. graduate students enroll in electrical engineering and computer science annually. Governments are spending over $100 billion to reshore chip manufacturing, but labor efficiency drops as production spreads.

Veteran pipelines offer one countermeasure. VetsinTech, backed by Craig Newmark Philanthropies, has trained over 2,000 veterans in cybersecurity and recently launched a "Vets in AI" track. The model, mission-oriented talent with clearance eligibility and high-stakes operational experience, maps to defense robotics. But it's a niche feed for a niche need.

What Happens Next

The $100 million in Ukrainian LOIs Edgerun announced at S26 Demo Day is a significant demand signal. It is also, as Edell acknowledged, prospective demand — not binding orders, not revenue. The gap between an LOI and a delivered system is where major U.S. programs have stalled.

TALOS was terminated struggling to power a 600-to-700-pound suit. ONYX, backed by a $6.9 million Army contract, ended over technical issues and funding. FORTIS found a niche holding tools but does not walk. SABER showed promise, with 90 percent of artillery soldiers reporting improved performance, yet remains a prototype. Dephy's ExoBoot has been tested for years without a production contract.

Edgerun's hip-worn device targets a narrower problem: helping soldiers walk under load. That focus avoids the full-body integration nightmare that sank TALOS. But supply chain realities are indifferent to scope. The actuators, batteries, sensors, and ruggedized electronics must still meet military environmental and safety certifications.

The sector's projected market size reflects the current reality: pilots, studies, and LOIs. It does not reflect production lines.

Edgerun now competes for the same supply chain talent that primes need. The hire they make will signal whether they are building for a pilot batch or for the thousands of units the Ukrainian LOIs imply. If they cannot secure allocated capacity for custom actuators and military-grade cells, the LOIs expire. If they can, they become the first U.S. exoskeleton company to cross the valley of death since the HULC program began two decades ago.

Working in frontier tech? Zero G Talent tracks the openings: see every open ASML role, browse frontier tech jobs, openings at Stripe, and the people building the field.