The first rule of dressing robots: forget everything you know about human proportions. A tailor measures a person and works from centuries of accumulated knowledge about how fabric drapes over shoulders, falls from waists, and wraps around limbs shaped by bone and muscle. None of that knowledge transfers cleanly to a machine whose shoulder joint is a brushless motor and whose "waist" is a steel yaw actuator.

Each humanoid robot platform has its own geometry, its own sensor layout, its own thermal profile, and its own movement envelope. Dressing them requires understanding all of these factors. This guide breaks down the major platforms currently in production or advanced development, with practical notes on garment design for each.

General Principles Before We Start

A few rules apply across every platform:

For foundational context on the field, see our overview of what robot couture is.

SoftBank Pepper: The Gateway Platform

Pepper is where many designers first encounter robot fashion, for the simple reason that there are more Peppers deployed in customer-facing roles than any other humanoid robot. Since its 2014 launch, Pepper has worked in banks, shops, hospitals, airports, and hotel lobbies across Japan, Europe, and beyond.

Body Characteristics

Pepper stands about 120 cm tall, roughly the height of a nine-year-old child. Its body is top-heavy, with a large head, broad chest, articulated arms, and a wheeled base instead of legs. The chest houses a 10.1-inch touchscreen display. The head contains cameras, microphones, and 3D sensors. The arms have limited but expressive range of motion.

Garment Approach

The most successful Pepper garments are vest-style or tabard-style coverings that drape over the shoulders and torso while leaving the chest screen fully exposed. Think of a barista's apron adapted for a non-human torso. The screen cutout is the central design constraint, everything flows around it.

Arm coverings work if they are loose enough to allow Pepper's gesture range, but many operators skip arm garments entirely because the arms are small and the design challenge outweighs the benefit. Head coverings are rare and generally not recommended, as they tend to interfere with facial recognition cameras and microphones.

The wheeled base is usually covered with a fabric skirt or drape, which has the side benefit of hiding the mechanical base and making Pepper look more like a character and less like a machine on wheels. This single addition, a base cover, has the highest impact-to-effort ratio of any Pepper garment.

Materials Notes

Pepper runs relatively cool and has minimal ventilation requirements, so thermal considerations are secondary. The main material requirements are washability (Pepper gets touched a lot in public settings), color fastness (the garment is often in bright retail environments), and light weight (heavy garments can affect Pepper's already-limited stability).

Tesla Optimus: The Mass-Market Contender

Tesla's Optimus (also called Tesla Bot) is the platform most likely to be produced at scale, which makes it the one most likely to drive volume demand for robot clothing. Tesla has stated its goal of producing millions of units, primarily for industrial and household tasks.

Body Characteristics

Optimus is designed to approximate human proportions more closely than most platforms. It stands roughly 173 cm tall, weighs about 57 kg, and has a body shape that is recognizably humanoid, two arms, two legs, a head, a torso. The head houses cameras and sensors behind a smooth, featureless face panel. The hands are dexterous, with individually actuated fingers.

Garment Approach

Optimus is, in many ways, the easiest humanoid robot to dress because its proportions are closest to human. Standard garment construction techniques, darting, seaming, pattern grading, transfer more directly than they do with other platforms. A designer with experience in menswear or workwear will find the transition less jarring than expected.

The main challenges are at the joints. Optimus uses visible actuators at the shoulders, elbows, hips, and knees, and these create bulges and geometry changes during movement that human joints do not. Garments need articulated panels or stretch fabric at these points. The face panel needs to remain fully exposed, and there are sensor arrays along the torso that require transparent or cutaway panels.

Because Tesla is targeting mass production, there is an opportunity for standardized sizing, something rare in the robot clothing world. If millions of Optimus units share the same dimensions, garment manufacturers can produce off-the-rack clothing for robots for the first time. This could fundamentally change the economics of the field.

Materials Notes

Optimus is designed for physical labor, and its clothing needs to withstand the same stresses. Durable, abrasion-resistant fabrics are essential. Stretch performance fabrics (think Cordura blends, canvas-stretch hybrids) are likely to be the workhorses. For household deployment, softer fabrics become viable, cotton-poly blends, jersey knits, that would not survive an industrial setting but are appropriate for home use.

Boston Dynamics Atlas: The Athlete

Atlas is the most physically dynamic humanoid robot in existence. It can run, jump, do backflips, and perform parkour maneuvers that would challenge a human gymnast. It is not a commercial product (Boston Dynamics uses it as a research platform), but it represents the extreme end of what garment designers may eventually need to accommodate.

Body Characteristics

The electric Atlas (introduced in 2024) is a significant departure from the earlier hydraulic version. It is more compact, with a highly articulated torso and limbs that can rotate in ways no human body can. The head is a sensor bar rather than a head-shaped form. The proportions are distinctly non-human, powerful legs, a relatively compact torso, and arms with extraordinary range of motion.

Garment Approach

Dressing Atlas is the hardest challenge in humanoid robot fashion. The range of motion is extreme, joints that rotate 360 degrees, a torso that can twist fully around, legs that bend both forward and backward. Any garment that restricts movement is unacceptable, and most garments will restrict movement at these extremes.

The most promising approaches borrow from competitive athletics and performance dance. Ultra-stretch fabrics, minimal coverage, strategic paneling, and garments that move with the body rather than around it. Compression-style garments that sit close to the surface and stretch in every direction show the most potential. Traditional tailored garments, jackets, shirts, trousers, are essentially impossible for Atlas in its full range of motion.

For applications where Atlas's full athletic capability is not needed (standing, walking, light manipulation), more conventional garments become feasible. The key is understanding the specific motion profile for the deployment context and designing to those requirements, not to the robot's theoretical maximum.

Agility Robotics Digit: The Warehouse Worker

Digit is designed specifically for logistics and warehouse work. It is one of the first humanoid robots to enter volume production for commercial use, with Agility Robotics targeting deployment in Amazon fulfillment centers and similar facilities.

Body Characteristics

Digit stands about 175 cm tall and has a distinctive look: backward-bending bird-like legs, a compact torso, articulated arms with simple grippers, and a head that is essentially a sensor module. The leg geometry is the most striking feature and the biggest challenge for garment design. Digit does not have a "front" of the knee in the human sense, the joint bends the other way.

Garment Approach

Forget trousers. The leg geometry makes conventional legwear impractical. Instead, designers working with Digit tend toward torso-focused garments, vests, tabards, or short-jacket styles that end above the hip joint. These can carry branding, safety markings, and protective panels without interfering with the complex leg kinematics.

If leg coverage is required (for safety visibility or debris protection), individual leg sleeves with elastic retention work better than connected garments. Think compression sleeves rather than pants. These need to accommodate the reversed knee geometry and the significant shape change that occurs during Digit's walking cycle.

The arms are simpler to dress than the legs, and arm-mounted tool pouches or utility sleeves have shown promise in warehouse settings. Digit's hands are simple grippers rather than articulated fingers, which makes glove-style coverings unnecessary.

Figure 01 and Figure 02: The Newcomers

Figure AI has moved rapidly from concept to functional prototype, and its robots represent the newer generation of humanoid design. Figure 02, announced in 2024, features a more refined human-like form factor than many competitors.

Body Characteristics

Figure 02 stands about 170 cm tall with proportions deliberately designed to approximate human body shape. The head is more traditionally head-shaped than many competitors, with cameras and sensors integrated into a face-like arrangement. The hands are among the most dexterous in the field.

Garment Approach

Figure 02's human-like proportions make it one of the more "dressable" platforms. Like Optimus, it benefits from the transferability of human garment construction techniques. The main design considerations are joint articulation (particularly at the shoulders and hips), sensor clearance on the head and torso, and cable routing along the arms and legs.

Figure's focus on general-purpose deployment means garments need to be versatile. A Figure robot might work in a warehouse in the morning and a retail store in the afternoon. Modular garment systems, a base layer with swappable outer pieces, align well with this use case.

Unitree H1 and G1: The Chinese Contenders

Unitree has made headlines with its surprisingly capable and relatively affordable humanoid robots. The H1 and the smaller G1 represent the Chinese robotics industry's push into humanoid platforms.

Body Characteristics

The H1 stands about 180 cm tall and is built for speed and agility. It holds speed records for humanoid running. The body is lean and mechanical-looking, with exposed actuators at major joints and a sensor bar head. The G1 is smaller (about 127 cm) and designed more for dexterous manipulation than athletic performance.

Garment Approach

The H1's exposed-joint design creates an interesting challenge. The actuators at the knees, hips, and shoulders are prominent visual features, and any garment needs to either accommodate them (with cutouts or stretch panels) or dramatically reshape the robot's silhouette by covering them with structured panels.

The most practical approach for the H1 is a sport-inspired aesthetic, performance fabrics, athletic cuts, stretch panels at joints. This aligns with the robot's physical capabilities and avoids fighting the mechanical aesthetic of the exposed hardware.

The G1, being smaller and more compact, suits vest and jacket-style garments similar to those used for Pepper, adapted for its different proportions and joint layout.

1X NEO: The Home Robot

1X Technologies (formerly Halodi Robotics) is developing NEO specifically for home environments. This changes the garment design calculus significantly.

Body Characteristics

NEO is designed to be soft and approachable. It uses a compliant actuator system that makes it inherently safer around people than rigid-joint robots. The body is intended to feel less mechanical and more organic in its movements. Proportions are roughly human-like but with softer, less angular lines than most competitors.

Garment Approach

Home deployment changes everything about garment design. The clothing needs to feel domestic, warm, and unthreatening. Think soft knits, natural fibers, muted colors. The aesthetic should be closer to casualwear than to workwear or athletic gear.

Washability is paramount, a home robot's clothing will get dirty frequently and needs to survive regular machine washing. The attachment system needs to be simple enough that a non-technical person (the homeowner) can change the robot's clothes. Snap buttons, simple zippers, and pull-on designs work better than the magnetic or tool-assisted closure systems used in commercial settings.

There is also a personalization angle. People will want to choose their home robot's clothing, the way they choose cushion covers or kitchenware. This creates demand for a variety of styles, colors, and designs, potentially the closest thing to a "fashion market" in the robot clothing world.

Cross-Platform Design Principles

Despite the significant differences between platforms, several design principles hold across the board:

Layering Systems Work

A base layer that fits close to the robot's body, plus an outer layer that can be swapped for different contexts, is the most versatile approach. The base layer handles sensor clearance and range-of-motion accommodation. The outer layer handles branding, safety, and aesthetics. This two-layer system scales across platforms and use cases.

Closures Matter More Than You Think

The closure system, how the garment opens, closes, and attaches to the robot, is often the make-or-break detail. Buttons are generally avoided (too fiddly for robotic or technician hands). Zippers work but can catch on cables. Magnets are popular for light garments but do not hold in high-movement situations. Snap fittings with positive-lock mechanisms are emerging as the preferred solution for most applications.

Color and Pattern Have Engineering Implications

Dark colors absorb more solar radiation, a meaningful consideration for robots operating outdoors or near windows. Reflective materials can interfere with optical sensors on nearby robots. Certain patterns can confuse computer vision systems. Color and pattern choices in robot fashion are not purely aesthetic decisions; they have functional consequences that human garment design rarely needs to consider.

Documentation Is Part of the Garment

Every robot garment should ship with a fitting guide, care instructions, and a sensor-clearance map. This is not optional. A garment that is fitted incorrectly can degrade the robot's performance or even create safety hazards. Clear documentation, ideally with photographs or diagrams, is as much a part of the product as the fabric itself.

Looking Ahead: Platform Convergence and Standardization

The current state of humanoid robot design resembles the early automobile industry: every manufacturer has its own proportions, its own joint designs, its own sensor layout. This fragmentation makes standardized clothing impossible and keeps the cost of garment development high.

Some convergence is likely as the market matures. The platforms that succeed commercially will establish de facto body standards, much as the IBM PC established a standard for personal computers. When that happens, and it may happen within the next five years, robot fashion will shift from bespoke fitting to something closer to ready-to-wear.

Until then, the garment designer's first job on any project is the same: get the robot in the room, scan it, study it, move it, and learn its body. There are no shortcuts yet. For the business context behind these platforms, see our robot clothing industry overview. For the functional arguments behind clothing robots at all, read why robots wear clothes.