Key takeaways

• Physical AI is the shift from software that answers to machines that act. A humanoid robot is the most demanding form of it, and its performance is set as much by what it is made of as by the model that drives it.

• Actuators, the motor-and-gearbox units that move the joints, account for 40–60% of the cost of building a humanoid; structure and batteries take 5–10% each. Every one of those subsystems is a materials decision before it is a design decision.

• No single material builds a humanoid. A working body map spans at least five families: engineering plastics for frames and housings, super engineering plastics for actuator internals, elastomers and synthetic rubber for skin, grips and soles, flame-retardant and shielding compounds for electronics, and battery materials for the power pack.

• LG Chem treats robotics materials as one of its three growth businesses and already supplies material families for nearly every region of the body: transparent and shielding materials for the head, reinforced engineering plastics for the frame and actuators, flame-retardant and battery materials for the torso, elastomers and latex for the hands, ABS and ASA for the shell, and synthetic rubber for the feet.

• The contribution of materials is measurable: weight that extends runtime, part consolidation that cuts cost, dry-running surfaces that remove maintenance, and surface properties that let a robot work in a hospital or a kitchen at all.

Ask a search engine what robots are made of and the answers tend to stop at “metal and plastic.” That was an adequate answer for an industrial arm bolted to a factory floor. It is not an adequate answer for a 1.7-meter biped that has to carry its own battery, work beside people, be wiped down with disinfectant and survive a fall. The humanoid robot industry is growing quickly: Goldman Sachs now expects it to ship 6.5 million units a year by 2035, up from a forecast of 1.4 million only two years earlier. As it grows, it is discovering something that is easy to miss from the outside: a surprising amount of the hard engineering is about materials.

This article is a map, and it makes one simple argument: in a humanoid robot, the list of materials is not a detail for the purchasing department. It is a large part of the design itself. That is why materials companies, which used to supply the robot industry from the sidelines, are moving closer to its center, and why it is worth asking which of them can cover the whole body. The sections below look at what physical AI asks of materials, how LG Chem, which recently named robotics materials one of its three growth businesses, reads that demand, and which robot materials go where in a humanoid robot’s body, from head to foot, with LG Chem’s own solutions noted region by region. It closes with what that breadth means for a robot developer. It is written for anyone who needs the whole picture before going deeper on any one part of it, whether engineer, sourcing manager or analyst.

Why physical AI is a materials problem as much as a software problem

Physical AI is the term the industry has settled on for artificial intelligence that perceives and acts in the physical world through a body: a robot, a vehicle, a machine. Where generative AI produces text and images, physical AI produces motion, and motion has a cost that software does not. A chatbot that makes a mistake wastes a few seconds; a robot that makes a mistake wastes battery, wears out a gear or falls over. A humanoid robot must hold its own weight upright, swing its limbs thousands of times an hour and keep a 2 kWh battery, roughly what a home uses in an hour or two, alive for a working shift. TrendForce reported in January 2026 that most current humanoids operate for two to four hours between charges, and lithium-ion cells are already close to how much energy they can hold per kilogram. The obvious fix, a bigger battery, mostly adds weight that the robot then has to carry.

The consequence shows up in the bill of materials, the itemized list of everything that goes into one robot and what each item costs. McKinsey’s April 2026 analysis of the humanoid supply chain puts actuators at 40–60% of a humanoid robot’s unit cost, sensing and perception at 10–20%, compute at 10–15%, structural components at 5–10% and battery modules at 5–10%, with cooling and wiring taking the rest. Read it as a list of materials rather than a list of costs and it becomes gears, housings, bearings, frames, enclosures, skins, cells and cables. Every one of them has a weight, a level of friction, a fire rating and a surface finish that somebody has to choose. The same analysis names two of the industry’s tightest bottlenecks: tactile sensors, which give a robot a sense of touch and have no standard design yet, and precision reducers, the gearboxes that turn a motor’s fast spin into slow, strong joint movement. Both are, at root, materials problems.

Weight is the thread that runs through all of it. Think of a humanoid robot’s leg as a lever. The further a kilogram sits from the hip, the more twisting force, or torque, the hip motor needs to swing it, and the effect grows with the square of the distance. So a kilogram removed from a shin saves more energy than a kilogram removed from the torso. A lighter limb also needs a smaller motor and gearbox, which saves still more weight. Noise, hygiene and safety follow close behind. A robot that squeaks at the joints, stains under disinfectant or struggles with a flammability test is unlikely to find its way into a hospital or a restaurant, however capable its model. None of these constraints can be fixed with better software. Together, they are what has turned the humanoid from an occasional customer for materials into a market in its own right.

How LG Chem reads the humanoid robot opportunity

LG Chem has set out its view of robotics in public, and in some detail, which makes it a reasonable place to start. In June 2026 the company announced a KRW 15 trillion research and development program running to 2035, with around 70% directed at three growth businesses: semiconductor materials, mobility materials and robotics materials. Within robotics it named three categories it intends to build: structural materials for the robot’s frame and body, precision actuation materials for the parts inside joints and drives that have to move exactly, and bonding materials, the adhesives and joining materials that hold the assembly together. The stated goal is to move from selling resin, the raw plastic that parts are molded from, to helping customers design the parts and the manufacturing processes themselves. The company has also begun to show what this looks like in practice. At Chinaplas 2026 in Shanghai it presented robot materials ranging from paint-free metallic ABS for exterior panels and high-refractive lens materials for robot eyes and sensors to heat-resistant PVC for internal wiring and battery-safety materials that delay thermal runaway.

The company calls this becoming a “tech-driven converting company”. In plain terms, it means using everything the company knows how to make in order to change its business as fast as its markets change, with three directions in mind: semiconductor materials, mobility and robotics, and cancer drugs. The reason this matters for the rest of the article is the range of materials behind it. LG Chem makes commodity and specialty petrochemicals, including ABS, polycarbonate, synthetic rubber and specialty resins; it has an engineering-materials division covering polyamide, PBT, POM and high-performance compounds; and its advanced-materials company positions itself as a battery and electronic materials supplier for e-mobility, with divisions for cathode materials and separators, two of the main ingredients of a lithium-ion cell. It also sits inside a group that is building the robots themselves: LG Electronics has announced a next-generation humanoid built on NVIDIA’s Isaac GR00T platform, and LG Energy Solution has signed battery supply agreements with Tesla, Boston Dynamics and Figure AI.

Three ideas recur in how LG Chem talks about the opportunity. The first is metal replacement, swapping a metal part for a plastic one that does the same job. A recent example the company points to is a polycarbonate compound reinforced with glass fiber that replaces aluminum in solar-panel frames at roughly half the weight with comparable durability. The same logic applies to robot frames and housings. The second is developing materials together with the customer rather than selling from a catalog. Its September 2026 agreement with Volkswagen to jointly develop next-generation lightweight materials is the car-industry version of the same idea. The third is building safety into the material itself, particularly around batteries and electronics. The robot materials it showed at Chinaplas included materials that delay battery thermal runaway, because a machine that works beside people has to contain a fault rather than let it spread. What those three ideas look like in actual parts is easiest to see anatomically, which is where the map proper begins.

A body map: which materials go where in a humanoid robot, and what LG Chem supplies

The most practical way to think about humanoid robot materials is the way an anatomy book would: one body part at a time. The table below summarizes the map: for each region, what it has to do, what the material must deliver and the material families LG Chem supplies for it. The sections that follow walk through the map from head to foot, including the metals and other materials used alongside these polymers, and each ends with a short note on LG Chem’s solutions for that part of the body.

Head and sensors: materials that let the robot perceive

A humanoid robot’s head is, in practice, a cluster of sensors with a face on it. The windows in front of its cameras and LiDAR (the laser-based depth sensor that lets a robot judge distance) need polycarbonate, a tough and clear plastic, that stays impact-resistant in the cold and, for depth sensors, lets specific infrared wavelengths through while blocking visible light. The electronics behind those windows pack a lot of heat and electromagnetic interference, the stray signals that scramble nearby circuits, into a very small space. Plastic compounds filled with carbon fiber and carbon nanotubes can block that interference, which is why they are replacing the metal shielding cans traditionally used, and why the housing resin is chosen for how much heat it can take as much as for how it looks. None of this is exotic. The same set of materials has been serving smartphone cameras and the driver-assistance sensors in cars for a decade, which is why materials suppliers already present in those markets, LG Chem among them, tend to find the robot head familiar territory.

LG Chem’s solutions for the head and sensors. LG Chem supplies transparent polycarbonate that keeps its impact strength in the cold and can be tuned to block light below a chosen wavelength, around 700 nm for example, while letting more than 90% of the near-infrared light a depth sensor uses pass through. For the lens covers of robot eyes and sensors, it offers the high-refractive materials it showed at Chinaplas, which combine glass-like clarity with impact resistance and allow curved designs. Behind the windows, its EMI-shielding compounds based on polycarbonate, polyamide and PBT use carbon fiber and carbon nanotubes, a material LG Chem also produces for conductive plastics and EMI shielding, to replace metal shielding while also spreading heat. Glass-reinforced PBT gives LiDAR and radar housings and camera modules dimensional stability and can be laser-marked and laser-welded.

Torso and frame: where metal replacement pays back

The torso is the structural problem. It has to carry the battery, the computer and the shoulders and hips, and it is the region where the choice between metal and plastic has the largest effect on total weight. The logic has already played out once in solar-panel frames: a polycarbonate or polyamide compound reinforced with glass fiber can be as stiff as aluminum at about half the density, expands less with heat than plain plastic, and needs no painting. In LG Chem’s own metal-replacement studies on robot housings, the plastic version has come in at roughly 40% lower weight and roughly 40% lower cost than the metal baseline, with the exact figure depending on the housing design.

The battery pack inside the torso is a materials system of its own. LG Chem’s advanced-materials business develops high-voltage, cost-innovative cathode materials and high-safety separators for electric vehicles, and a humanoid robot’s pack is in effect a small, oddly shaped vehicle pack that has to stay cool in less space. Around the cells, the enclosure, the insulation on the busbars that carry current between cells, and the connectors must meet UL 94 V-0, the strictest common flammability rating for plastics, which is where flame-retardant ABS and PC/ABS grades do work that ordinary resins cannot.

LG Chem’s solutions for the torso, frame and battery. For frames and housings, LG Chem offers polyamide compounds reinforced with 40–60% glass fiber, which combine high strength and stiffness with dimensional stability and flame retardancy, as well as flame-retardant PC/ABS. Inside the body, its heat-resistant, flexible PVC can carry internal wiring. Around the battery, its flame-retardant ABS reaches a UL 94 V-0 rating at 1.0 mm thickness for PCB, power and battery boxes, and it has added engineering plastics designed to delay thermal runaway and an aerogel-based thermal barrier to the same toolkit. Inside the cells, the company works on the cathode materials, separators and carbon-nanotube conductive additives that set how much energy a pack can store and how safely it does so.

Actuators and joints: precision, wear and silence

The actuator, the motor-and-gearbox unit that moves a joint, is where a humanoid robot spends most of its money and most of its energy, and it is where the most demanding polymers live. Reducer gears, bushings and thrust washers, the small parts that carry force and take the wear inside a joint, increasingly use super engineering plastics. That is the family of plastics rated for continuous use at roughly 150°C and above, and its best-known member is polyether ether ketone (PEEK), which melts at 343°C and has a density of around 1.3 g/cm³, about half that of aluminum. What makes these plastics valuable is the combination: lightness, the ability to run against steel without grease, electrical insulation and a damping effect that quiets gear noise. One tier down, POM, a hard and naturally slippery plastic, handles lower-load gears and bearings, and glass-reinforced PBT provides motor housings and connectors that keep their shape after molding, shrug off moisture and can safely carry the high voltages that pass through the connectors.

Two properties that rarely make a datasheet headline turn out to matter a great deal. One is squeak. A joint that sticks and then slips against its housing makes a noise that is tolerable in a warehouse and intolerable in a hospital corridor. Anti-squeak ABS compounds, developed to pass the stick-slip tests that carmakers use for interiors, transfer directly from dashboards to robot joints. The other is heat. A compact actuator can keep its plastic parts well above room temperature for hours, and a plastic loses stiffness as it approaches its glass-transition temperature, the point at which it begins to soften. That is why experienced teams tend to pick the grade based on how it behaves at operating temperature rather than at room temperature.

LG Chem’s solutions for actuators and joints. LG Chem’s current robotics portfolio covers the tiers below super engineering plastics. Glass-reinforced PBT is used for motor and actuator housings, where it offers high strength, low warpage, good electrical properties and flame retardancy; other PBT grades serve connectors that must resist electrical tracking, including a hydrolysis-resistant version that holds its properties in heat and humidity. POM covers gears and bearings, chosen for rigidity and resistance to friction and wear, and an anti-squeak ABS that meets an automotive stick-slip standard is offered for motor housings and other parts that rub. Precision actuation materials are also one of the three robotics categories the company has said its R&D program will build, which places the higher-performance tier on its roadmap rather than in its current catalog.

Hands and skin: the soft materials that make a robot safe to touch

The hand is the hardest part of a humanoid robot to build and the part people notice first. It needs rigid finger linkages, grip surfaces that give a little, a protective skin and, increasingly, an array of touch sensors underneath that skin. The rigid parts follow the actuator logic above. The soft parts belong to elastomers, the family of stretchy, rubber-like plastics: thermoplastic copolyester elastomers for grip pads that need both feel and wear resistance, silicone for skins, and nitrile or other synthetic rubbers for the gloves that food-handling or clinical robots replace daily.

Tactile sensing, the robot’s sense of touch, is where materials research is moving fastest. A June 2025 study in Nature Machine Intelligence embedded high-resolution touch across a robot hand using an elastomer membrane over clear silicone and showed that tactile feedback raised collision-avoidance success from 53.5% to 100%. Cambridge researchers the same month demonstrated a single-material electronic skin made from a conductive gelatin-based hydrogel that senses pressure, temperature and damage at once. What both results have in common is that the sensor is a material rather than a separate device: an elastomer or a hydrogel (a soft, water-rich gel) whose stretch, conductivity and ability to spring back define the data the AI model receives. For materials companies with synthetic rubber and elastomers in their portfolios, that is an open door rather than a finished product.

LG Chem’s solutions for hands and skin. For grippers, LG Chem offers a thermoplastic copolyester elastomer with a soft feel, high impact strength and good abrasion resistance. For the gloves that service robots go through every day, its nitrile latex, the main material in disposable medical gloves, and its polyolefin elastomers are natural starting points. Its carbon nanotubes can also make elastomers conductive or antistatic, a property that matters more as skins begin to carry sensors.

Outer shell: the surface the world touches

The shell is where the robot meets people, dust, cleaning chemicals and weather, and it is where ABS, the familiar hard plastic used in everything from appliance housings to toy bricks, and its weather-resistant sibling ASA do most of the work. The requirements come from where the robot is used rather than from the robot itself. A hospital or food-service robot is wiped with isopropyl alcohol or detergent many times a day, which makes ordinary resins crack; chemically resistant ABS grades are specified so the surface survives. The same environments want antibacterial performance that is built into the material and measured to ISO 22196, the standard test for antibacterial surfaces, rather than a coating that wears off. Warehouse and air-handling robots collect dust on vents and grilles, which anti-dust compounds with built-in antistatic behavior reduce. Outdoor delivery and patrol robots need ASA’s resistance to UV, rain and temperature swings so the shell neither yellows nor turns brittle.

Appearance is a functional property here too. ABS grades that come out of the mold with a metal-look, ultra-matt or high-gloss finish remove the painting step, which cuts cost and process steps and avoids chipped paint on a machine that bumps into door frames. Translucent ABS behind a hidden display lets status lights show through an unbroken surface.

LG Chem’s solutions for the outer shell. The shell is where LG Chem’s portfolio is deepest. Its ABS and ASA grades for robots cover each need described above: chemically resistant ABS that does not crack after exposure to isopropyl alcohol or detergent under strain; an antibacterial grade that keeps 99.99% antibacterial activity after accelerated aging tests; an anti-dust grade that uses slip and antistatic agents; paint-free metallic, ultra-matt and high-gloss finishes, with support on mold design for matt surfaces; a grade made for metal plating; translucent ABS for hidden displays; flame-retardant ABS for electronics boxes; and weatherable ASA for outdoor robots.

Feet and mobility: where rubber meets the floor

The sole of a humanoid robot’s foot takes the same punishment as a tire tread, compressed into a few square centimeters: repeated impact, sideways shear and slip on wet or dusty floors. Solution-polymerized styrene-butadiene rubber and butadiene rubber, the compounds that give car tires their balance of grip, wear resistance and low rolling resistance, are the natural starting point for foot pads and wheel treads, with polyurethane where heavier loads have to be carried. It is a small part by weight and a large part by consequence, because a robot that slips falls, and a robot that falls is the most expensive failure a deployment can have.

LG Chem’s solutions for feet and mobility. LG Chem produces both solution-SBR and butadiene rubber, the tire compounds that give treads their grip and wear resistance, and its thermoplastic copolyester elastomer is also used for robot wheels, where impact strength and abrasion resistance decide how long a tread lasts.

How materials move the numbers: function, efficiency and optimization

It is easy to list robot materials by property and harder to show how they move the numbers a humanoid robot program is judged on: runtime, unit cost, maintenance, safety approval and where the robot is allowed to work. The table below connects the two and names the LG Chem material families behind each effect.

The pattern is that materials rarely improve one number on its own. Replace a metal housing with a reinforced plastic and the part gets lighter; a lighter part needs less torque to move; less torque allows a smaller actuator; a smaller actuator costs less and gives off less heat; and less heat widens the choice of materials around it. Design engineers call this a virtuous cycle. Sourcing managers call it the reason a material decision made in the first month of a program is still being felt in the fifth year. The cycle runs in reverse just as readily, which is why the places where polymers still fall short deserve as much attention as the places where they win.

What is still unsolved

A robot materials map is only useful if it is honest about its blank spots, and there are several. Structural limbs remain metal, because even reinforced plastics do not match aluminum’s stiffness for the weight where stiffness is what matters. Molded plastic gears that hold the very tight tolerances of a strain-wave reducer (the compact, high-ratio gearbox used in many robot joints) over millions of cycles are a specialist product, not a catalog item. Tactile sensor skins have no standard design, so every hand is a custom materials project. The highest-performing polymers are expensive and made by only a few suppliers, and McKinsey notes that the humanoid supply chain is concentrated in precision gearboxes, roller screws, rare-earth magnets and force sensors in ways that materials alone cannot fix. And the data on how parts wear and tire over years of use, which aluminum and steel have accumulated over a century, does not yet exist for plastics under the particular loads of a walking machine.

None of these is a problem of physics. They are the same engineering and supply-chain problems that metal-to-plastic conversion faced in cars and aircraft, and in both cases the conversion went ahead once the first programs proved the parts and the material suppliers helped generate the data. Two of the gaps above, precision parts inside the actuator and the bonding that holds a mixed metal-and-plastic body together, are also two of the three robotics categories LG Chem has said its R&D program will build. The question for a robot developer is therefore less which material wins and more which partner can help close the remaining gaps.

Why LG Chem: how its materials support humanoid robot development

Seen from a robot developer’s side, the map above is a long list of materials to specify and suppliers to qualify. Seen from LG Chem’s side, most of it is one portfolio. Five capabilities describe how the company supports humanoid robot programs.

The first is breadth. LG Chem supplies material families for nearly every region of the body map, from lens and shielding materials in the head to synthetic rubber in the feet, so a developer can work through a large share of a robot’s polymer and battery materials with one partner, one set of test data and one technical team.

The second is tuning materials to where the robot will work. The same shell may need chemical and antibacterial resistance in a hospital, anti-dust behavior in a warehouse and weatherability outdoors. LG Chem’s robotics grades are organized around exactly those conditions, and the company supports mold design as well as material selection, which matters most for paint-free finishes.

The third is experience in replacing metal. The glass-reinforced compounds now proposed for robot frames grew out of automotive and solar-frame work, and LG Chem’s own robot housing studies put the plastic version at roughly 40% lighter and 40% cheaper than metal. Lightweight materials are also the focus of its September 2026 agreement with Volkswagen.

The fourth is depth in batteries and safety. Few parts of a humanoid robot are as safety-critical as its battery, and LG Chem works on it from both sides: the cathode materials, separators and conductive additives inside the cells, and the flame-retardant plastics, thermal-runaway-delaying compounds and aerogel barrier around them.

The fifth is technical support close to where robots are designed. LG Chem runs customer and technical service centers in Korea, China, Germany and the United States, and its stated aim is to help customers design parts and manufacturing processes rather than only supply resin, so material choices can be checked against a robot’s actual geometry and operating conditions early.

Behind these sits a group that is building humanoid robots and supplying their batteries. In practice, the largest gains come when the materials conversation starts while a robot’s overall architecture is still being decided, because housings, brackets and enclosures, whose size is set by geometry rather than peak load, are fixed early. That is the role LG Chem has said it wants to play: a materials partner in the room while the part is still being drawn. Which is another way of saying that the answer to what humanoid robots are made of is still being written, and that a surprising share of it will be written by the people choosing the materials.

Frequently asked questions

What are humanoid robots made of?

Humanoid robots are made of a combination of metals and polymers chosen by body region. Aluminum and steel still carry the main structural loads and gear trains; engineering plastics such as reinforced polyamide, PBT, polycarbonate and ABS form frames, housings, enclosures and shells; super engineering plastics such as PEEK serve actuator gears and bearings; elastomers and synthetic rubber provide skins, grips and foot soles; and lithium-ion cells built from cathode and separator materials supply power.

What is physical AI, and why do materials matter to it?

Physical AI is artificial intelligence that perceives and acts in the physical world through a machine such as a humanoid robot. Materials matter because motion has physical costs that software does not: the mass, friction, heat, noise and durability of a robot’s parts set how long it can run, how safely it can work near people and how much it costs to build. Actuators alone are 40–60% of a humanoid’s bill of materials, and their performance is largely a materials question.

Which plastics are used in robots?

Robots use a wide range of plastics. ABS and ASA are common for outer shells because of their chemical resistance, finish and weatherability; polycarbonate and PC/ABS for transparent windows, enclosures and frames; glass-reinforced polyamide and PBT for structural parts, motor housings and connectors; POM for gears and bearings; PEEK and other super engineering plastics for high-load actuator components; and thermoplastic elastomers and silicone for grips and skins.

Can plastic replace metal in a humanoid robot?

Partly. Reinforced engineering plastics can replace aluminum in housings, brackets, enclosures and frames that are sized by geometry rather than peak load, typically at about half the weight and, in LG Chem’s robot housing studies, around 40% lower cost. Super engineering plastics can replace metal in gears and bearings. Long structural limbs that are governed by stiffness per kilogram generally remain metal.

What materials are used for robot skin and tactile sensors?

Robot skins use soft materials: silicone and thermoplastic elastomers for the outer layer, nitrile and other synthetic rubbers for replaceable gloves, and elastomer membranes or conductive hydrogels as the sensing substrate. Recent research has shown elastomer-based tactile skins raising robot grasp reliability substantially, and single-material hydrogel skins that sense pressure, temperature and damage at once.

How is LG Chem involved in robotics materials?

LG Chem has named robotics materials one of its three growth businesses within a KRW 15 trillion R&D program to 2035, and is developing structural, precision actuation and bonding materials for robots. Its current portfolio covers most of a humanoid robot’s body: transparent, high-refractive and EMI-shielding materials for the head and sensors; glass-reinforced polyamide, PBT and PC/ABS for frames, housings and actuators; POM for gears; flame-retardant plastics and cathode and separator materials for the battery; ABS and ASA for the shell; elastomers and nitrile latex for grippers and gloves; and solution-SBR and butadiene rubber for the feet.

Sources and further reading

• Goldman Sachs Research forecast revision, September 2026, as reported by Humanoid Guide.

• McKinsey & Company, “Scaling the humanoid robotics supply chain into billion-dollar wins” (April 2026).

• TrendForce, “Humanoid Robots Move Toward Commercialization; Solid-State Batteries Set to Break Through Power Bottlenecks” (January 2026).

• LG Corp. newsroom, “LG Chem to Invest KRW 15 Trillion in R&D, Intensifying Focus on Semiconductor, Mobility, and Robotics Materials” (June 2026).

• LG Chem blog, “기술이 강한 컨버팅 회사, LG화학이 나아갈 미래” (July 2026) and “금속 프레임을 대신할 수 있는 플라스틱이 있다고?” (September 2026).

• LG Chem, Advanced Materials business domain page.

• Asia Business Daily, “LG Chem to Showcase Robotics, Automotive Electronics, and Medical Materials at Chinaplas 2026” (April 2026).

• LG Chem product pages: ABS, ASA, SSBR, NBR latex and carbon nanotubes.

• UPI, “LG Chem teams with Volkswagen to develop new materials” (September 2026).

• Robotics Tomorrow / PR Newswire, “LG to Unveil Its Next-Gen Humanoid Robot, Built on NVIDIA Isaac GR00T” (August 2026); KED Global, “LG Energy wins battery supply deals with top humanoid robot makers” (July 2026).

• Nature Machine Intelligence, “Embedding high-resolution touch across robotic hands enables adaptive human-like grasping” (June 2025); University of Cambridge, “Single-material electronic skin gives robots the human touch” (June 2025).

• Wikipedia, “Polyether ether ketone” and “UL 94”; ISO 22196.

This article is part of a series on the materials science behind physical AI. It describes material families and engineering principles and does not constitute a product specification or recommendation.