Persona AI and Under Armour Are Designing Workwear for Humanoid Robots

Persona AI and Under Armour announced a research collaboration on May 11, 2026, to test performance materials for humanoid robots that will work around welding, heavy manufacturing, extreme heat exposure, and hazardous material handling.

The project is not about making robots look human. It is about whether external material layers can improve durability, thermal regulation, abrasion resistance, flexibility, and long-term reliability when humanoids are expected to use tools in the same places where human crews wear protective gear.

AI-generated image

An abstract materials-testing scene for industrial humanoid workwear. Source: AI-generated editorial illustration.

How Robot Workwear Works

Humanoid robots are usually discussed as software systems with legs and hands. That framing misses a basic industrial problem. A robot working near weld spatter, grinding dust, sheet metal edges, heat, oil, solvents, and repetitive contact needs a protective outer layer in the same way human workers need gloves, sleeves, boots, visors, and flame-resistant clothing.

Persona AI said the collaboration will examine how different textiles behave under heat, friction, repetitive motion, and other real-world stresses. Under Armour brings experience in athlete performance materials, including thermal management, abrasion resistance, and flexible designs that can move repeatedly without breaking down. Persona brings the robot use case: industrial humanoids intended for shipyards, construction sites, energy infrastructure, and manufacturing facilities.

The important technical detail is that humanoid protection cannot simply copy human workwear. Robots have exposed joints, cable routes, sensor windows, cooling paths, access panels, and tool interfaces. A material layer that traps heat, blocks a camera, snags during arm motion, or changes the friction profile of a gripper could reduce performance instead of improving it.

That makes the materials question closer to industrial design than apparel design. The outer layer has to fit the robot's kinematics, meaning the range and path of each joint. It also has to support service workflows. A damaged forearm cover should be quick to inspect and replace. A torso layer should not force a technician to remove major hardware just to reach a battery connector, cooling duct, or compute module.

Under the Hood: Materials as a Robot Subsystem

The collaboration points to a category that robotics companies will likely treat as a real subsystem: robot-specific protective gear. It sits between the mechanical body and the work environment. The design problem includes heat transfer, cut resistance, electrostatic behavior, fluid exposure, washability, replaceable panels, and how field technicians inspect damage.

For humanoids, the material layer also has to preserve motion. A stiff sleeve around an elbow or shoulder can raise actuator loads. A loose cover can catch on fixtures. A reflective patch can interfere with machine vision in certain lighting. A thick glove-like layer can change tactile sensing and grip repeatability. Those are not fashion choices. They are uptime, safety, and maintenance variables.

AI-generated image

Heavy manufacturing environments create heat, abrasion, and contamination problems that lab humanoids rarely face. Source: AI-generated editorial illustration.

Who's Building the Robot Workwear Stack

Persona AI is headquartered in Houston and describes its platform as industrial humanoid robotics for welding, construction, energy, fabrication, inspection, maintenance, grinding, painting, and material handling. The company emphasizes high-risk work settings where labor shortages and safety concerns already affect operations.

Under Armour is entering from the materials side rather than the robot manufacturing side. Kyle Blakely, the company’s senior vice president of innovation, design studio, development, and testing, said the work lets Under Armour apply concepts such as thermal management, abrasion resistance, and flexibility beyond sport.

What This Means for Industrial Humanoids

The news is small in one sense: no commercial robot suit was announced, no deployment count was disclosed, and the project remains early R&D. It matters because it is one of the clearer signs that humanoid companies are moving from demo constraints toward jobsite constraints. A factory or shipyard does not care whether a robot can wave on stage. It cares whether the machine can survive a shift, avoid damaging itself, and return to service quickly after routine wear.

For industrial buyers, protective materials could become part of the total cost of ownership calculation. If a textile layer prevents scratches, keeps sparks away from cables, reduces contamination inside joints, or simplifies field maintenance, it can pay for itself through fewer service calls. If it traps heat or causes snags, it becomes another failure point.

There is also a branding signal. A mainstream athletic apparel company is treating humanoid robots as a plausible future market for performance products. That does not make the market mature. It does show that adjacent industries are starting to plan around robots as working assets rather than lab curiosities.

The bigger industry lesson is that humanoid commercialization will be won by the unglamorous parts as much as the neural networks. Buyers will ask about spares, cleaning, protective ratings, safety procedures, repair time, and how the robot behaves after 500 hours of contact with real tools. Materials are one piece of that checklist, but they touch almost every other piece.

What's Coming Next

The near-term milestones are practical. Watch for material test results, field trials in welding or fabrication settings, photos of prototype covers, and whether Persona identifies replaceable panels as part of normal maintenance. The most useful metric will not be a dramatic demo. It will be mean time between failures in dirty, hot, repetitive work. Customer safety audits will also matter because exterior materials can affect pinch points, emergency stops, and cleaning procedures.

If the collaboration advances, expect other humanoid manufacturers to discuss protective exteriors more openly. Factory robots already use guards, cable tracks, end-effector covers, and environmental ratings. Humanoids will need their own version of that discipline, especially if they are sold into work that humans currently perform with personal protective equipment.

Frequently Asked Questions

The 12-Month Outlook

Persona AI and Under Armour are testing a narrow but important question: how does a humanoid survive industrial work after the locomotion demo is over? If the answer becomes a set of validated materials, covers, and replaceable protective layers, robot exteriors will move from styling to field engineering.

The next year should show whether this remains a research note or becomes a deployable product category. The clearest signals will be field pilots, material specifications, replacement schedules, and whether customers ask for robot protection the way they already ask for safety ratings, tool compatibility, and service contracts.