Humanoid Payload & Reach: What These Robots Can Actually Lift and Carry

Defining the Physical Limits

In the rapidly maturing field of humanoid robotics, the ability to manipulate the physical world remains the primary differentiator between entertainment and utility. For RobotWale readers, the question is not merely what a robot looks like, but what it can physically do. We define "payload" as the maximum static or dynamic weight a robot can carry at its end-effector without compromising stability or structural integrity. "Reach" refers to the workspace envelope—the maximum distance from the robot’s base to the furthest point the manipulator can extend while maintaining torque control.

Unlike wheeled autonomous vehicles where range is primarily a battery metric, humanoid payload and reach are mechanical constraints dictated by motor torque, linkage geometry, and center-of-gravity management. Current industry standards suggest that for industrial logistics, a payload capacity of 15kg to 20kg is the baseline for viability, allowing the handling of standard e-commerce parcels and automotive components. Anything below this threshold often relegates the robot to a research platform rather than a production asset.

Shipping Hardware: The Ground Truth

We grade claims by shipping hardware first. This category includes robots that have been manufactured, deployed in real environments, and have published technical specifications verified by third-party observers or official press releases. Currently, only a handful of companies have moved beyond prototype stages to actual hardware deployment.

Tesla Optimus Gen 2

Tesla’s Optimus Gen 2 represents the most aggressively publicized hardware in this space. During the 2022 AI Day and subsequent updates, Tesla disclosed a payload capacity of approximately 20kg (44 lbs) with a reach extending to waist height for most tasks. The design utilizes a compact actuator suite designed for high torque density.

However, the 20kg figure is often cited for static lifting. Dynamic lifting—moving that weight while walking or maintaining balance—likely reduces the effective payload to 10-15kg to prevent tipping. The arm’s reach is limited by the torso height of 169cm (5ft 6in), placing the working envelope squarely at human waist and shoulder levels. This is suitable for factory assembly lines but less optimal for high-bay warehousing without external scaffolding.

Agility Robotics Digit

Agility Robotics’ Digit is a distinct comparison point because it is bipedal but not fully humanoid in hand configuration (though it has a gripper). Digit is designed for logistics and has demonstrated a payload capacity of 15kg (33 lbs). Its reach is constrained by its height of 162cm (5ft 3in). The critical advantage here is reliability; Digit has been deployed in Amazon’s logistics centers, providing real-world data on payload endurance rather than theoretical torque limits.

Boston Dynamics Atlas (Humanoid Mode)

Boston Dynamics’ new electric Atlas replaces the hydraulic predecessor with electric actuators. While the original hydraulic Atlas had impressive agility, the new electric version focuses on payload capacity and precision. Current specs indicate a payload capacity of roughly 18kg (40 lbs) with a reach optimized for warehouse picking. The focus here is on repeatability and torque control rather than extreme dynamic movement.

Pilot Deployments & Near-Term Commercialization

Beyond shipping hardware, we examine pilot deployments. These are units currently operating in controlled industrial environments, often with limited availability to third-party buyers.

Figure AI & Apptronik

Figure AI, in partnership with BMW, has deployed Figure 01 units in pilot programs. While specific payload numbers are often guarded, Figure AI has demonstrated the ability to lift standard 15kg boxes and place them into bins. The reach is estimated at 170cm (5ft 7in). Apptronik’s Apollo aims for similar metrics, targeting 20kg payload capacity with a focus on manufacturing support. Both represent the tier where hardware is present, but mass production is not yet confirmed.

Fourier Intelligence

Fourier Intelligence has released the H1 and P1 models. The P1 model, which is closer to a shipping-ready unit, targets industrial manufacturing. Specifications suggest a payload of 20kg with a reach extending 175cm. This model has been demonstrated in automotive assembly lines in China, providing a reference for payload endurance under continuous load.

The Indian Market Context

For Indian industries, the conversation shifts from technical capability to economic viability. India’s robotics import policy imposes a 10% Basic Customs Duty (BCD) on finished robots, with component duties varying from 0% to 20% depending on the nature of the import. This significantly impacts the landed cost of imported humanoids.

Availability and Pricing

Currently, no global humanoid manufacturer has a dedicated India sales channel for mass-market humanoids. Units must be imported via specialized robotics integrators. Estimated landed costs for Tier 1 hardware (Tesla Optimus, Figure 01, Agility Digit) range from $50,000 to $150,000 USD. Converting this to INR with current exchange rates and taxes, the price bracket sits between ₹42 Lakhs and ₹1.2 Crores for a single unit.

This pricing excludes integration costs, which often double the total cost of ownership. However, local manufacturing initiatives, such as the production of humanoid components in India under the PLI (Production Linked Incentive) scheme, could reduce this cost by 20-30% over the next five years. For now, the Indian market is restricted to pilot deployments in large enterprises like Tata Motors or Adani, where the ROI calculation supports high CAPEX.

Infrastructure Constraints

Payload and reach are not just about the robot; they are about the environment. Indian warehouses often feature racking heights exceeding 10 meters. A robot with a 169cm reach cannot service the top 40% of these racks without external lifting aids. This creates a bifurcation in the market: robots for ground-level pick-and-place versus robots for high-bay logistics. The current generation of humanoids is heavily weighted toward the former.

Technical Trade-offs

The relationship between payload and reach is inverse. As the arm extends further from the base, the moment arm increases, requiring exponentially more torque from the joints to maintain the load. This is why most humanoids sacrifice reach for payload capacity. A robot that can lift 20kg at full extension usually requires a much larger battery pack to manage the energy draw.

Energy Consumption

Lifting a 20kg load at shoulder height consumes significantly more energy than lifting at waist height. For a robot with a 4-hour battery cycle, carrying payload reduces operational time. This is a critical metric for Indian shift-based manufacturing, where battery swaps or charging infrastructure must be factored into the floor plan.

Stability and Safety

The payload limit is often dictated by the center-of-gravity (CoG) tolerance. If a robot carries an off-center load, it must use its legs to counterbalance. If the weight exceeds the motor torque limit, the robot will tip. This is why safety certifications are required for payload handling. In India, where labor costs are low, the economic case for a robot that can safely handle 20kg must be weighed against the cost of a human worker who can handle the same task indefinitely without energy constraints.

Conclusion

The current state of humanoid payload and reach is defined by the gap between announcement and deployment. While the hype cycle suggests 50kg payloads are imminent, the ground truth remains in the 15kg to 20kg range for shipping hardware. For Indian industries, the viability of these systems hinges on localized manufacturing to reduce the ₹1 Crore price tag and infrastructure upgrades to support the 170cm reach envelope.

Until these costs drop and payloads stabilize at 20kg+ with high reach, humanoids will remain specialized tools for high-value manufacturing rather than general-purpose labor. The focus must remain on hardware that ships, pilots that deploy, and announcements that verify.

References

• Tesla AI Day 2022 and 2023 Keynotes. tesla.com/ai
• Agility Robotics Product Specs. agilityrobotics.com/products/digit
• Boston Dynamics Electric Atlas. bostondynamics.com
• Figure AI Official Website. figure.ai
• Fourier Intelligence P1 Specifications. fourierintelligence.com
• Ministry of Commerce and Industry, India. Import Duty Schedule. dgft.gov.in