For warehouse and manufacturing automation in 2026, the strongest documented all-around option is Agility Robotics’s Digit—especially if your priority is deploying robots into existing warehouses rather than running a research pilot.
How the leading options compare
Robot
Best fit
Commercial maturity
Notable evidence
Digit 5 — Agility Robotics
Warehousing, material movement, logistics, repetitive factory work
Highest documented industrial deployment
Digit has logged 65,000+ operating hours; deployments include GXO, Amazon, Schaeffler and Toyota.
For warehouse and manufacturing automation in 2026, Agility Robotics’ Digit is the strongest choice if your priority is proven commercial deployment. It has documented ongoing logistics operations, including GXO, and has moved 100,000+ totes in commercial environments.
Manufacturing, line-side material handling, assembly
Advanced/piloting
Apptronik is developing Apollo around manufacturing and warehouse workflows and has established customer/partner deployments.
Figure 03 — Figure AI
General-purpose factory/warehouse automation
Rapidly scaling
Figure says it has produced 350+ third-generation robots and demonstrated a production rate of one robot/hour.
Atlas — Boston Dynamics
Automotive manufacturing, complex manipulation
Emerging industrial deployment
Boston Dynamics is testing Atlas for Hyundai/Kia manufacturing tasks at a new Georgia facility.
Why I'd look hardest at Digit
Digit has something particularly valuable for an operations manager: actual sustained industrial operating history rather than primarily demonstrations. Agility reports more than 65,000 hours of operation and a GXO deployment that has moved 100,000 totes with approximately 98% on-task accuracy.
Toyota Motor Manufacturing Canada has also moved from a pilot to a commercial Robots-as-a-Service agreement for manufacturing, supply-chain and logistics work.
The new Digit 5 is specifically aimed at scaling this into human-populated facilities, with Agility claiming cooperative safety without the physical barriers traditionally used around industrial robots. It's worth noting that Digit 5 itself is still in development; early access is expected in 2027.
But Apollo may be more interesting for manufacturing
If your primary problem is factory-floor manipulation rather than moving totes, I'd put Apptronik Apollo very high on the shortlist. Apptronik explicitly positions Apollo for line-side tasks, inspection, material handling and repetitive manufacturing workflows, while also developing warehouse applications.
So, rather than a universal winner:
Warehouse tote/bin movement → Digit
Automotive/manufacturing material handling → Digit or Apollo
More general-purpose manipulation → Apollo or Figure
Complex automotive manufacturing applications → Atlas is worth evaluating
Researching the technology rather than deploying immediately → the field is much broader
One important caveat: humanoids remain a small fraction of industrial automation. The International Federation of Robotics reported roughly 7,000 humanoids sold for industrial/professional service applications in 2025, versus hundreds of thousands of conventional industrial robots.
If you tell me what your operation actually does (e.g. palletizing, picking, machine tending, kitting, assembly, welding, inspection, etc.), I can narrow this down to the 2–3 robots worth evaluating and compare payload, cycle time, integration, deployment model, and likely ROI.
The best humanoid robot for warehousing and manufacturing depends on your specific operational constraints, but the market currently centers on three top-performing commercial contenders: Apptronik Apollo, Agility Digit , and Figure AI (Figure 02/03).
Choosing the right system requires balancing payload capacity, dexterity, and software ecosystem maturity.
Top Humanoid Robots for Industrial Automation
Robot Model
Max Payload
Key Advantage
Primary Use Case
Apptronik Apollo
25 kg (55 lbs)
Highest payload capacity & safe human-mimetic force actuators
Heavy box/tote transport, kitting, and line supply
Agility Digit
16 kg (35 lbs)
Most commercially mature fleet software ecosystem (Agility Arc)
Apptronik Apollo: Best for heavy-duty material handling. Apollo features a biomimetic architecture using custom linear actuators that mimic human muscle, allowing it to carry up to 55 lbs (25 kg) . Partnered heavily with logistics providers like GXO Logistics and automotive manufacturers like Mercedes-Benz, it thrives in environments requiring stable movement of heavy totes and cases.
Agility Digit: Best for proven end-to-end commercial deployment. Digit uses a bipedal mobile base optimized for walking through narrow human aisles. Backed by the Agility Arc fleet management platform, it is deeply integrated into live supply chains for repetitive container-moving tasks without custom facility retrofitting.
Strategic Considerations Before Adopting Humanoid Robots
Evaluate Task Variability: Humanoid robots shine when workflows change hourly, or when handling mixed-SKU environments that break traditional fixed automation or fixed robotic arms.
Infrastructure vs. Throughput: If your facility handles thousands of identical items per hour, traditional Automated Guided Vehicles (AGVs) or fixed conveyors still outperform humanoids on raw speed and cost efficiency.
Software Integration: Ensure the robot supports RESTful APIs and integrates cleanly with your existing Warehouse Management System (WMS) or Manufacturing Execution System (MES).
To help narrow down the ideal platform, tell me:
What is your primary task (e.g., trailer unloading, bin picking, assembly)?
What is the average payload weight you need to move?
As of September 2026, if you're talking about a robot you could realistically deploy in a warehouse or factory, rather than a research prototype, Agility Robotics’ Digit is the strongest all-around choice today—particularly for warehouses and material handling. Figure 03 is arguably the more interesting alternative for complex manufacturing manipulation.
My shortlist
Robot
Best fit
Evidence of real work
Key published specs / status
Agility Digit 5
Warehousing + mixed industrial workflows
Digit has accumulated 65,000+ hours of industrial operation; GXO deployment has moved 100,000+ totes
23 kg / 50 lb payload, 2.2 m reach, claimed 20 hr work/24 hr, WMS/WES/MES integration
Figure 03
Manufacturing + complex manipulation
Figure 02 ran daily on BMW's production line, loading 90,000+ parts over 1,250+ hours; Figure 03 is now at BMW
20 kg payload, 5 hr runtime; strong vision/manipulation AI and whole-body control
Apptronik Apollo
Manufacturing/logistics pilots
Partnerships/pilots with companies including GXO and Mercedes-Benz
Very promising industrial design, but less publicly demonstrated at sustained production scale than Digit/Figure
Boston Dynamics Atlas
Advanced factory automation/R&D
Hyundai/Boston Dynamics are now establishing an industrial testing environment for Atlas
Exceptional mobility/manipulation potential, but commercial availability and deployment evidence are less mature
Tesla Optimus
Long-term high-volume manufacturing
Tesla is developing it for its own factories
Potentially enormous manufacturing scale, but external commercial availability and independently verifiable production performance remain limited
Why I'd pick Digit for a warehouse
The important distinction is deployment maturity rather than impressive demos.
Digit has already been used in an actual warehouse under a multi-year commercial arrangement with GXO. Agility says the platform has accumulated more than 65,000 hours of real-world operation and that Digit 4 surpassed 100,000 totes at GXO.
The newly announced Digit 5 is particularly relevant because it expands beyond simple tote movement into depalletizing, machine tending, kitting, sequencing, inspection and palletizing. It also has integrations intended to let it work alongside existing AMRs, conveyors, WMS/WES and MES systems.
One caveat: Digit 5 itself is not yet generally available. Agility says early access is expected in H1 2027 and general availability by the end of 2027.
Why Figure 03 may be better for manufacturing
Figure has unusually strong evidence that its robots can do actual production work. Figure 02 operated on BMW's assembly line for 10-hour weekday shifts, loaded over 90,000 parts and contributed to production of more than 30,000 X3 vehicles.
Figure 03 is now being used at BMW for a harder sequencing/material-manipulation workflow, where parts arrive in variable orientations and the robot has to coordinate its hands, torso and feet while moving a cart.
That's significant because manufacturing often isn't just "pick up tote → move tote." Dexterity, perception and adapting to variation become much more important.
Figure 03 currently lists a 20 kg payload and 5-hour runtime, while Figure says its BotQ facility has ramped production to roughly one robot per hour.
The practical answer
If I were evaluating a deployment, I'd narrow it this way:
Automotive/manufacturing assembly and sequencing:Figure 03
Long-duration industrial pilot with a manufacturer:Apollo
Cutting-edge manipulation/mobility R&D:Atlas
Betting on eventual enormous production scale:Optimus
And I would not buy a cheap humanoid simply because the hardware is available. There are commercially listed humanoids such as Unitree platforms, but the gap between owning a humanoid and having one reliably perform an industrial workflow at production throughput is enormous. For example, current Unitree industrial/research platforms can be purchased for tens of thousands of dollars, but they're fundamentally different propositions from a managed industrial deployment.
If you tell me your specific task—e.g. pick-and-place, palletizing, machine tending, kitting, welding, inspection, order picking—plus approximate payload, shift length, and target units/hour, I can narrow this down to the most suitable platform and estimate the likely ROI/payback versus a conventional robot or human labor.
When looking for the "best" humanoid robot for warehouse and manufacturing automation, the industry has shifted away from isolated lab prototypes and into real-world factory deployments . Because different robots excel at different operational metrics (payload capacity, software intelligence, or commercial maturity), the top contenders stand out in specific categories:
Agility (formerly Agility Robotics) — Best for Commercial Maturity & Logistics
The Core Strengths: Agility’s Digit (with iterations like Digit 5) is arguably the most field-tested bipedal humanoid in logistics, boasting tens of thousands of hours in active warehouse operations with partners like GXO Logistics, Amazon, and Schaeffler.
Best Use Case: Material handling, tote lifting, moving payloads up to 35–50 lbs, and seamlessly integrating with existing Autonomous Mobile Robot (AMR) fleets via their Agility Arc cloud platform.
Boston Dynamics — Best for Dynamic Agility & Heavy Industrial Tasks
The Core Strengths: Their production-ready, all-electric Atlas represents the gold standard in dynamic balance, hardware resilience, and manipulation. Where other robots might stumble on uneven debris or complex twists, Atlas handles high-stress, heavy-lifting physical tasks with extreme physical intelligence.
Best Use Case: Heavy-duty manufacturing assembly lines, aggressive component handling, and automotive plant logistics (such as active deployments scaled with Hyundai).
Figure AI — Best for Embodied AI & Adaptability
The Core Strengths: Backed by cutting-edge neural visual-language models (like their Helix AI system), Figure robots are built as general-purpose workers. Their deployment on active automotive lines (such as BMW) has proven that their platform can translate natural language and live vision data into functional, precise assembly actions rather than scripted routines.
Best Use Case: Complex manufacturing tasks requiring hand-eye coordination, parts-to-person feeding, and multi-step assembly sequences.
UBTECH Robotics — Best for Scale & High-Volume Deployment
The Core Strengths: UBTECH’s Walker S series has scaled rapidly in high-volume industrial environments, particularly across electronics and automotive manufacturing in Asia. They feature innovative workflow adaptations, such as autonomous battery-swapping capabilities for true 24/7 factory operations.
Best Use Case: Repetitive factory floor sorting, quality control inspection, and continuous-shift production line support.
To help narrow down which system fits your specific operation, tell me:
Are you focusing on logistics/warehousing (moving boxes, palletizing) or manufacturing/assembly (tight spaces, tool usage)?
What is the average payload weight and reach height you need?
If you're choosing one humanoid robot today for real warehouse/manufacturing automation, my pick is Agility Robotics's Digit.
My 2026 ranking
Robot
Best fit
Maturity
My take
Agility Digit
Warehousing, tote/bin handling, intralogistics
Highest
🥇 Best overall for deployment now
Figure 03
Automotive/manufacturing, general manipulation
Very high
🥈 Best for AI-heavy factory work
Apptronik Apollo
Manufacturing + logistics, heavier workflows
High
🥉 Excellent enterprise alternative
Boston Dynamics Atlas
Advanced manufacturing
Emerging
Best hardware potential, but less proven commercially
Tesla Optimus
High-volume factory automation
Emerging
Huge upside, but not my choice for a purchase today
Chinese platforms (Unitree, UBTECH, etc.)
R&D / lower-cost experimentation
Variable
Interesting value, but industrial autonomy is less proven
🥇 Why I'd choose Digit
Digit has the strongest combination of actual industrial deployment, purpose-built logistics design, and commercial availability. Agility says Digit is already deployed in manufacturing, distribution and logistics, rather than being primarily a research prototype.
It's particularly suited to tasks such as:
Moving totes and bins
Picking up and placing containers
Feeding production lines
Repetitive material handling
Moving goods between stations
Working alongside existing warehouse automation
Agility's current commercial strategy is also unusually focused on ROI and deployment, rather than simply demonstrating increasingly impressive robot capabilities. Current published estimates put Digit around $200,000 per robot plus deployment and recurring software/maintenance costs, although actual customer contracts can differ.
🏭 If manufacturing is the priority: Figure becomes very compelling
For an automotive or general manufacturing environment, I'd seriously consider Figure AI instead.
Figure has demonstrated a particularly strong manufacturing case with BMW. Its Figure 02 robots performed production work at BMW's Spartanburg plant, including loading sheet-metal components into fixtures. One reported deployment involved more than 30,000 vehicles over 10 months.
That makes Figure arguably the more interesting choice if your workflow involves dexterous manipulation rather than simply moving material.
Apollo is the dark horse
Apptronik Apollo would be my third choice—and potentially first depending on your application.
Apollo was explicitly designed around industrial work, with a roughly 55-lb lifting capability and modular/swappable-battery architecture. Mercedes-Benz has been exploring Apollo for manufacturing/logistics applications.
Its positioning is somewhat different from Digit: less narrowly optimized for warehouse tote handling and more aimed at becoming a general industrial worker.
What I would not do
I wouldn't choose based on the most impressive demo.
The humanoid market is moving incredibly quickly, but the difficult part isn't making a robot walk. It's getting:
99%+ task reliability × high uptime × safe human interaction × easy integration × low intervention cost.
Recent reporting from China illustrates the problem: impressive humanoid hardware can still struggle with dexterity, autonomy and adapting to unstructured factory conditions.
And even Boston Dynamics' new Atlas is not yet deployed at meaningful scale; the company is targeting much larger production capacity over the next few years.
Bottom line
If I had $10M to automate a warehouse today:
→ Digit
If I were automating an automotive/manufacturing line:
→ Figure, with Apollo as a very serious alternative.
If I were planning a 3–5 year humanoid strategy:
→ I'd run pilots with Digit + Figure + Apollo, because the winner is likely to depend more on your exact workflow and integration requirements than on which robot has the best demo.
If you tell me what you're manufacturing/handling, approximate payload, number of shifts, and the specific tasks you want automated, I can narrow this down to the best robot and estimate ROI/payback versus human labor and conventional automation.
There is no single "best" humanoid robot for warehouse and manufacturing tasks, as different models excel at specific payloads, power systems, and environments, but Apptronik's Apollo and Figure AI's Figure series currently lead commercial and pilot deployments.
Selecting the right humanoid robot depends heavily on your specific operational constraints, payload targets, and integration budget.
Top Humanoid Robots for Industrial Automation
Apptronik Apollo : Standing 5'8" and carrying up to 55 lbs , Apollo uses a biomimetic linear actuator design that mimics human muscle, allowing safe, cage-free operation alongside human workers and seamless movement through standard human-engineered facilities. It is actively trialed by major logistics providers like GXO Logistics.
Figure AI (Figure 01 / Figure 02 / Helix) : Known for high-end autonomous neural-network control, Figure robots have demonstrated long-duration autonomous package sorting, barcode scanning, and tote handling using advanced whole-body balancing and visual AI.
Tesla Optimus : Positioned for massive internal manufacturing scaling and eventual commercial deployment, focusing on autonomous general-purpose factory manipulation tasks.
Agility Robotics Digit : A bipedal mobile manipulator optimized specifically for moving totes and packages within confined warehouse pathways, actively tested in commercial logistics environments.
Comparison of Key Industrial Humanoid Metrics
Robot Model
Payload Capacity
Primary Application Focus
Key Advantage
Apptronik Apollo
Up to 55 lbs
Heavy tote/box transport, palletizing
High payload, human-safe force control
Figure AI Series
~44 lbs (20 kg)
Package sorting, barcode scanning
Advanced end-to-end AI vision and dexterity
To help narrow down the best fit for your facility, tell me:
Are you automating warehouse logistics (tote moving, box sorting) or manufacturing assembly (precision placement, kitting)?
If you mean the best humanoid robot you could realistically deploy for warehouse or factory work today, my pick is Agility Robotics’ Digit.
My ranking for 2026
Robot
Best for
Maturity
My take
Agility Digit
Warehouses, tote/bin handling, logistics
⭐⭐⭐⭐⭐
Best overall today
Figure 03
Manufacturing, assembly, high-dexterity tasks
⭐⭐⭐⭐½
Best manufacturing contender
Apptronik Apollo
Manufacturing + logistics
⭐⭐⭐⭐
Strongest flexible industrial platform
Atlas
Advanced automotive manufacturing
⭐⭐⭐½
Extremely capable, but less commercially accessible
Unitree H1/G1
R&D, experimentation
⭐⭐⭐
Great value, not my choice for production
🥇 Best overall: Digit
Agility Robotics's Digit has the strongest evidence of actually doing useful commercial warehouse work rather than just demonstrating capabilities. It has been operating commercially at GXO and has been used for tasks such as moving bins and totes.
That's particularly important because warehouse automation is less about spectacular demos and more about uptime, predictable cycle times, safety, integration and cost per task.
Agility's newly disclosed economics are also unusually concrete: its illustrative model for Digit v5 assumes about $200,000 for the robot + $20,000 deployment + $36,000/year software and maintenance, with a modeled five-year cost around $400,000. The company estimates roughly 20 hours of operation per 24-hour period and a 50-lb lifting capacity.
🏭 For manufacturing: Figure 03
If your primary application is automotive or factory assembly rather than warehouse material movement, I'd put Figure AI's Figure 03 at or near the top.
Figure reports that its earlier Figure 02 accumulated more than 1,250 hours on a BMW assembly line and loaded over 90,000 parts, which is unusually strong evidence for a humanoid doing actual production work.
Its advantage is the potential for more dexterous manipulation and broader factory tasks, whereas Digit is particularly optimized around logistics.
🤖 The wildcard: Apollo
Apptronik's Apollo may ultimately be the most interesting general industrial platform.
It's designed around manufacturing as well as logistics, with modularity and a roughly 25-kg payload class. Apptronik has also pursued industrial programs with companies including Mercedes-Benz and GXO.
So if I were building a new factory automation program rather than buying the most proven warehouse robot, Apollo would be one of the first systems I'd evaluate.
What I'd actually buy
Warehouse: Digit
Automotive/factory assembly: Figure 03
Mixed warehouse + manufacturing: Apollo
Research/development rather than production: Unitree
Cutting-edge automotive automation where access isn't the constraint: Atlas
One important caveat: a humanoid isn't automatically the best robot for every automation task. For fixed, repetitive operations—palletizing, welding, machine tending, conveyor sorting, etc.—a conventional industrial robot or AMR can still be substantially cheaper and more reliable. Humanoids become compelling when you need one robot to perform many different tasks in infrastructure designed for humans.
If you tell me your specific tasks (e.g. picking totes, palletizing, machine tending, assembly), payload/weight, shifts per day, and approximate labor cost, I can narrow this down to the best robot and estimate the ROI/payback period.
There is no single "best" humanoid robot, as industry leaders like Agility Robotics and Apptronik optimize their platforms for different operational needs. Agility Robotics Digit leads the market for verified commercial logistics work and tote-moving, while Apptronik Apollo stands out as a heavy-duty powerhouse designed to mimic human strength and movement.
Different humanoid platforms excel across specific warehouse and manufacturing parameters:
Robot Platform
Core Strength
Payload Capacity
Best Use Case
Agility Robotics Digit
Commercial logistics & reliability
Up to 35 lbs
Moving totes, repetitive material handling, and trailer unloading
Apptronik Apollo
Biomimetic strength & dexterity
Up to 55 lbs
Heavy tote transport, precise item placement, and barcode scanning
Boston Dynamics Atlas
Dynamic whole-body control
Variable
Complex parts sequencing and high-mobility manufacturing tasks
Key Factors in Choosing a Humanoid Robot
Locomotion Type: Bipedal robots (like Digit or Atlas) handle stairs and tight human-centric spaces naturally, while wheeled or hybrid lower bases (like Apollo's optional configuration or Agibot G2) offer vastly superior energy efficiency and payload stability on flat factory floors.
Payload vs. Dexterity Trade-off: Traditional autonomous mobile robots (AMRs) still outperform humanoids at moving heavy pallets (1,000+ lbs). Humanoids shine in variable manipulation—handling irregular shapes, picking from diverse shelf heights, and assembly kitting.
Infrastructure Requirements: Biomimetic actuators and advanced AI vision models allow newer systems to integrate into existing human environments without costly facility retrofits.
Watch how industrial humanoid platforms operate in real manufacturing and warehouse environments:
For warehouse and manufacturing automation, the “best” humanoid robot depends on whether you value commercial readiness, dexterity, cost, or long-term potential. As of 2026, there is no single universally dominant platform, but a few stand out.
Short answer: best current choices
🥇 Best for warehouses today: Agility Robotics Digit humanoid robot
Best fit: tote movement, picking assistance, conveyor work, repetitive material handling.
Why:
Designed specifically around logistics rather than being a general-purpose research robot.
Has one of the strongest real-world warehouse deployment histories, including work with GXO Logistics.
Its narrow footprint and ability to operate in human-designed spaces are major advantages.
Good match for tasks like: - moving bins/totes
transferring goods between stations
simple pallet and rack workflows
Weaknesses:
Less focused on highly dexterous manipulation than some competitors.
Still expensive and requires workflow integration.
My pick for a company wanting to automate a warehouse in the near term: Digit.
🥈 Best for manufacturing flexibility: Apptronik Apollo humanoid robot
Best fit: factories, assembly support, kitting, inspection, machine tending.
Strengths:
Built as a modular industrial platform.
Designed for different configurations, including wheeled and bipedal setups.
Strong fit for manufacturing environments where tasks change over time.
Good applications:
parts delivery
line-side replenishment
kitting
quality inspection support
light assembly assistance
Weaknesses:
Less proven at large-scale warehouse operations than Digit.
More of an enterprise pilot platform today.
My pick for a factory evaluating a long-term humanoid workforce: Apollo.
🥉 Highest upside: Tesla Tesla Optimus
Best fit: companies betting on future mass-produced general-purpose robots.
Strengths:
Tesla has enormous advantages in: - AI development
battery technology
manufacturing scale
vehicle-style production expertise
Long-term vision is a low-cost general humanoid workforce.
Weaknesses:
External availability and standardized industrial deployment remain less mature than specialized competitors.
Best choice if you are planning for 5–10 years out, not buying today.
Also worth watching: Figure AI Figure robots
Figure has attracted major attention for manufacturing applications, including automotive work, with an emphasis on AI-driven generalization and dexterity.
Strength:
Potentially strong for complex factory tasks.
Weakness:
Less publicly demonstrated as a broad warehouse workhorse compared with Digit.
Practical ranking by use case
Task
Best current candidate
Warehouse tote movement
Digit
Distribution centers
Digit
Factory material delivery
Apollo / Digit
Flexible manufacturing
Apollo
Assembly assistance
Apollo / Figure
Long-term mass deployment
What I would buy for a real operation today
Large warehouse (Amazon-style fulfillment, 3PL): Digit
Automotive/electronics factory: Apollo or Figure pilot
Company willing to experiment heavily: Optimus
Need ROI within a few years: also compare against non-humanoid automation (AMRs, robotic arms, cobots), because they are usually cheaper and more mature for fixed tasks.
The key question is not “which humanoid is strongest?” but “which robot can complete your specific workflow with the fewest facility changes?” Humanoids are most compelling where humans currently do many varied tasks in human-designed spaces.
Figure AI: Best for high-dexterity assembly. Figure utilizes human-equivalent hands with high degrees of freedom (DoF) and advanced visual-language AI models. Proved out in automotive plants like BMW , it excels where manipulation requires fine finger dexterity rather than just brute transport.
Agility Digit
~35 lbs
Warehouse logistics & bin picking
Bipedal mobility tailored for tight aisles
Agibot G2
Autonomous industrial fleets
Up to 100 kg base
Assembly line quality inspection and component sorting
