A robot arm is not a factory-floor luxury. Lightweight arms — many designed to work right next to people — now handle real production tasks in shops with 10 to 200 employees. If your team spends hours each day on repetitive manual work, a robot arm is worth a serious look.

The shop floor is dark except for the CNC status light. The cobot picks a warm part from the chuck, sets it on the conveyor, loads the next blank, and taps the cycle-start. The spindle spools up. No one else is here yet.

Where a robot arm fits in

The strongest starting points are tasks that are repetitive, consistent, and physically simple — the same motion, hundreds of times per shift, with parts that don’t change much between runs.

Mainstream today

Machine tending — loading and unloading a CNC, press brake, or injection molder
Pick-and-place — moving parts between stations or into packaging
Palletizing — stacking boxes onto pallets at end of line
Welding — MIG/TIG on fixtured parts with repeatable joints
Sanding and polishing — force-controlled surface finishing with consistent pressure
Visual inspection — camera-equipped quality checks at production speed

Emerging applications

On-demand micro-factory — a cobot tends a CNC and 3D printer, kitting custom orders from digital files overnight
Adaptive rework stations — cobot scans a defective part with 3D vision, plans a repair path, and re-machines or re-welds without custom fixturing
Collaborative assembly — cobot handles heavy or awkward subassemblies while a human does the fine work
Lab automation — sample handling, pipetting, and test prep in quality labs and small pharma
Food and beverage prep — portioning, dispensing, and plating in commercial kitchens and commissaries
Autonomous mobile manipulation — cobot arm mounted on an AMR, moving between workstations and performing tasks at each stop without fixed installation
Massage therapy — force-sensitive cobots delivering consistent deep-tissue and recovery treatments in wellness clinics

Cobots vs industrial arms

🧩 Mini Cobot

•Tabletop-sized — fits on a workbench
•Payload under 1 kg, reach under 500 mm
•$5,000–10,000 entry point
•Ideal proof-of-concept before scaling up

Best for: Light kitting, sorting, lab tasks, education

🤝 Collaborative (Cobot)

•Works alongside people — no safety cage
•Payload typically under 20 kg
•Fast to reprogram for new tasks
•Lower total deployment cost

Best for: Flexibility, shared workspaces, first-time automation

⚡ Industrial Arm

•Full-speed operation behind guarding
•Payload from 20 kg to 1,000+ kg
•Cycle times under 3 seconds
•Reach beyond 1,700 mm

Best for: Speed, heavy payload, hazardous processes

Choose industrial when

Cycle times under 2–3 seconds are required
Parts exceed 20 kg or cobot payload ratings
Reach beyond 1,700 mm is needed
Process is hazardous — welding, grinding, painting, hot parts
High-speed pick-and-place at 60+ cycles per minute

Stick with a cobot when

Payloads stay under 15 kg at moderate speeds
Frequent reprogramming for different products
No space for safety enclosures
Shared workspace with human operators

Costs

The entry point is smaller than you think

Compact tabletop cobots — like the AgileX Piper class — start around $5,000–10,000. They handle light pick-and-place, kitting, and sorting tasks on a bench. You don’t need a $75K cell to prove that automation works for your shop. Start small, prove value, then scale.

$5–10K

Mini cobot (entry)

$25–35K

Standard cobot arm

$50–100K

Full cobot cell

<$20K

Industrial arm (entry)

A cobot arm starts around $25,000–35,000. Total cell cost typically lands between $50,000 and $100,000 once you add tooling, fixtures, integration, and training. An industrial arm may cost less than a cobot, but safety infrastructure (fencing $3–15K, light curtains, interlocks) and 30–50% more floor area push total cell cost higher. Leasing options can bring monthly payments close to what you’d pay a part-time employee.

What the payback looks like

Time the task you want to automate: hours per day, labor cost per hour. Plug the numbers in below to see when the robot pays for itself.

System tier

Hours/day on task

Labor cost/hour ($)

Days/week

Estimated payback

9 months

Annual labor displaced: ~$39,000

What else you need besides the robot

End-of-arm tooling matched to your parts
Mounting surface (table, pedestal, or floor plate)
Programming time — some cobots use teach-by-demonstration
Vision system if parts don’t arrive in the same position every time
An internal system owner — not a robotics engineer, someone willing to learn

Explore robot arms

If your team spends a significant part of each shift walking materials between stations, a mobile robot can take over that movement. These are not warehouse-scale-only machines — a single unit in a mid-sized facility can free up hours of labor every day.

The warehouse associate checks his step counter at lunch. 14,000 steps — mostly pushing the same cart between the same three stations. The AMR now runs that triangle route every 22 minutes, all shift, without a break.

Where mobile robots make the biggest difference

The highest-impact application is repetitive point-to-point material movement: raw materials to workstations, work-in-progress between process steps, finished goods to shipping. If your team is walking the same routes dozens of times per shift, that movement is directly replaceable.

Best-fit applications

Point-to-point material delivery between fixed stations
Cart replacement on high-frequency internal routes
Finished-goods transport from production to shipping dock
WIP movement between sequential process steps
Kitting and staging for assembly lines
Inventory replenishment from bulk storage to pick zones

AMR vs AGV

🧭 AMR

•Navigates with sensors, cameras, and maps
•Reroutes around obstacles dynamically
•Easy to redeploy to new routes
•No floor modifications required

Best for: Changing layouts, mixed traffic, first deployment

🛤️ AGV

•Follows fixed paths — tape, wire, or paint
•Simpler, cheaper per unit
•Highly predictable routing
•Lower compute requirements

Best for: Stable layouts, predictable routes, tight budgets

Costs

Start with one robot, one route

You don’t need a fleet to get value. A single mobile robot handling your highest-volume route can prove the concept and generate real throughput data. Most vendors offer fleet management software that makes adding units straightforward later.

$25–40K

Entry-level AMR

$5–15K

Deployment & mapping

12–18 mo

Typical payback

A single robot replacing four to six hours of daily material handling labor typically pays for itself within 12 to 18 months. Calculate your payback against actual labor hours displaced — not projected efficiency gains.

What your facility needs

Reasonably flat floors — most AMRs handle small ramps and thresholds
Aisle width of 1.2–1.5 meters minimum
Wi-Fi coverage for fleet management and status reporting
Clear, consistent pathways — biggest prep step is usually decluttering aisles

Where mobile robots struggle

Narrow aisles below one meter
Outdoor terrain with significant elevation changes
Layouts that shift daily or weekly
Tasks requiring the robot to pick, load, or align items — that’s manipulation, not movement
First few weeks — expect occasional interventions until the map stabilizes

Browse mobile robots

The buying process for a robotic system is unfamiliar territory for most businesses. It doesn’t work like ordering equipment from a catalog. This guide walks through each stage so you know what to expect, what to ask, and where first-time buyers typically go wrong.

The owner opens a spreadsheet and types: "CNC machine tending, 6-second cycle, aluminum housings, 3×4-foot footprint, $80K budget." He sends it to three integrators before lunch. Two respond with site-visit requests by end of day.

Start with the task, not the robot

Document exactly what you want to accomplish before looking at any product: the specific task, the cycle time, the parts involved, the physical constraints of your space, and your target budget. A clear task description is the single most important thing you can bring to a vendor conversation.

What to document before calling anyone

The specific task and motion sequence
Part dimensions, weight, and material
Required cycle time and daily volume
Available floor space and ceiling height
Existing equipment the robot must interface with
Target budget (including integration)
Timeline — when do you need this running?

Who to contact

🏭 OEM (Direct)

•Sells the robot itself
•May offer application packages for common tasks
•Skips integrator markup (30–50%)
•You handle installation and programming

Best for: Teams with in-house engineering, straightforward applications

🔧 Systems Integrator

•Designs and builds the complete workcell
•Handles tooling, fixtures, safety, programming
•Absorbs project risk you’d otherwise carry
•Fixed-price contracts available

Best for: First-time buyers, custom tooling, complex cells

📦 Distributor

•Resells OEM products with local support
•Often offers basic integration services
•Good regional availability
•Can connect you with certified integrators

Best for: Regional support, simpler deployments

Go direct to the OEM when

The OEM offers a pre-built application package that matches your task
Your team has engineering talent for installation and programming
The application is straightforward enough that custom tooling isn’t required
You’re comfortable taking on the integration risk to save 30–50%

Use an integrator when

Your application requires custom tooling, fixtures, or grippers
The robot needs to coordinate with other equipment on the line
The cell requires safety guarding design
Your team doesn’t have engineering bandwidth to manage the project
This is your first robotic system

Vendor red flags

Proposes a specific robot model in the first conversation
Doesn’t ask to see your parts or visit your facility
Won’t run a feasibility study with your actual materials
Can’t provide references from businesses your size
Avoids fixed-price contracts or clear acceptance criteria

Questions to ask any vendor

How many projects similar to mine have you completed?
What is your typical timeline from agreement to production?
What does your support contract include, and what does it cost?
Can you connect me with two or three reference customers at my scale?
Do you have certified integrator partners in my region?

Is this task even a good fit?

Score your task in 60 seconds

Rate 1–5 on four dimensions. Example: CNC machine tending with identical parts (5), two shifts (4), hard to staff (4), moderate tolerance (3) = 16. Totals of 16–20: pursue immediately. 11–15: explore with a vendor. Below 10: focus on process improvement first.

1–5

Input consistency

1–5

Volume & frequency

1–5

Labor difficulty

1–5

Precision needs

Not every repetitive task should be handed to a robot. If you can’t hire for the position, can’t retain people, or workers are getting injured, robotics becomes attractive even at lower volumes. Under 1,000 task-hours per year with no safety or hiring pressure, focus on process improvement first.

Understand the full cost

<50%

Robot share of total cost

$2–10K

Feasibility study

2–8 wk

Simple cell timeline

2–4 mo

Complex cell timeline

The robot itself is often less than half the project cost. Beyond the arm: end-of-arm tooling, integration engineering, safety equipment, installation, operator training, and a service contract. Ask every vendor for a three-year total cost of ownership estimate so you can compare proposals on equal terms.

Plan for the ramp-up

Systems don’t run at full capacity on day one — plan a learning period
Designate an internal system owner with dedicated time
Budget for operator training beyond the initial install
Expect edge cases to surface in the first weeks of production
Request a feasibility study ($2–10K) before committing to a full purchase

Explore the full directory

Over 50 cobot models from a dozen manufacturers. This vendor-neutral guide cuts through the noise — specs, costs, and trade-offs that actually matter.

Six browser tabs, six cobots, identical headlines: "easy to use," "fast payback," "safe around people." The spec sheets blur together.

The selection problem

The right cobot for a welding shop is wrong for an electronics assembly line. You need a framework, not another brochure.

What makes a cobot a cobot?

A cobot works safely alongside humans without full safety caging. Built-in torque sensors detect contact and stop the arm before dangerous force levels.

The trade-off: cobots sacrifice speed and payload for safety, easier programming, and smaller footprints.

Top 10 cobot applications

Pick and place
Machine tending (CNC, injection molding, press brakes)
Palletizing and packaging
Assembly (screwdriving, insertion, snap-fit)
Welding (MIG, TIG, arc)
Quality inspection
Surface finishing (sanding, polishing, deburring)
Gluing and dispensing
Lab automation

The Cobot Landscape: Payload vs Reach

Hover any dot to see the model. Click to visit its profile. Most cobots cluster in the 3–16 kg payload range with 500–1300 mm reach.

Three cobot profiles

Knowing which profile matches your application narrows the field fast.

🏋️ Payload-Optimized

•Payload above 12 kg, often 16–25 kg
•Heavier robot weight (30–65 kg)
•Reach typically 1200–1500 mm
•Higher joint torques for heavy parts

Best for: Palletizing, machine tending with heavy parts, material handling

🥏 Reach-Optimized

•Reach above 1300 mm, up to 1750 mm
•Moderate payload (5–12 kg)
•Longer arm segments, larger footprint
•Good for multi-station coverage

Best for: Large workspaces, palletizing wide patterns, machine tending at distance

🎯 Precision-Optimized

•Repeatability under 0.03 mm
•Lower payload (3–5 kg)
•Compact design, lighter weight
•Often 6+ axis for dexterous access

Best for: Electronics assembly, inspection, lab work, fine dispensing

Key specs decoded

Payload (kg) — Max weight at full reach, including gripper. A 10 kg cobot with a 2 kg gripper handles 8 kg parts.

Reach (mm) — Distance from base to tool tip. Measure your pick-to-place distance and add 20% margin.

Repeatability (mm) — How precisely it returns to the same point. Under 0.05 mm for assembly; 0.1 mm is fine for palletizing.

Degrees of freedom — 6-axis is standard. 7-axis adds a redundant joint for flexible positioning in tight spaces. 4-axis (SCARA) is cheaper but limited.

IP rating — IP54 for normal environments. IP67 for washdown, food, pharma, or machining with coolant.

The ecosystem matters more than the arm

The arm is 30–50% of total cost. Tooling, integration, safety, and training are the rest. A cheaper arm with no ecosystem costs more long-term.

Tooling (EOAT): Check for certified grippers and tools. UR has 1,100+ accessories. Some brands have fewer than 20.

Software: Pre-built app packages (palletizing wizards, vision-guided pick) cut programming from hours to minutes.

Integrators: No local support = risky bet. Check regional availability before buying.

Costs and ROI

What you’re really paying for

The arm is the visible cost. Integration, tooling, and training are the majority. Always ask for a three-year TCO estimate.

$3–10K

Desktop / education cobot

$25–45K

Standard production cobot

$50–100K

Full cobot cell (arm + tooling + integration)

12–36 mo

Typical payback period

TCO includes: arm, tooling ($2–15K), integration ($5–30K), safety ($1–5K), training ($1–3K), and maintenance (5–10%/year). Two cobots at the same arm price can differ by $20K+ in total deployment cost.

ROI calculator

Select your application, adjust the numbers, see when the robot pays for itself.

System tier

Hours/day on task

Labor cost/hour ($)

Days/week

Estimated payback

9 months

Annual labor displaced: ~$39,000

7 mistakes to avoid

Automating the hardest task first — start simple
Using a cobot when an industrial robot is better
Ignoring end-of-arm tooling costs
Underestimating integration costs (arm is 30–50% of total)
Assuming cobots are inherently safe without a risk assessment
Not involving operators early
Spec-shopping without considering ecosystem

Before you buy

Define your application and task sequence
Measure your workspace (pick distance, place distance, clearance)
Calculate total cost of ownership, not just arm price
Verify safety compliance for your industry
Check EOAT availability for your application
Request demos with your actual parts
Designate an internal system owner

Browse all cobots

Yes — if you know where to look and what to expect. Several humanoid robots are commercially available today, but the market splits into three very different worlds. This vendor-neutral guide cuts through the hype.

You watched that humanoid demo on LinkedIn — the one where it folds laundry, carries boxes, does a backflip. You Googled “buy humanoid robot.” You got press releases, waitlists, and a lot of “coming soon.”

So… can you actually buy one?

The short answer is yes — several humanoid robots are commercially available today. But what you can actually do with your purchase varies wildly depending on whether you’re a factory operator, a home buyer, or a researcher. This guide maps the landscape honestly.

What counts as a humanoid?

A humanoid robot has a bipedal, upright, general-purpose form factor designed to operate in human-built environments — doorways, stairs, workstations. The key promise: a robot that fits into spaces designed for people, without rebuilding your facility.

This guide covers humanoid robots only. Quadrupeds (like Boston Dynamics Spot) are a separate category with different use cases and a separate guide.

Three markets, three realities

🏭 Industrial Humanoid

•Enterprise-only, $100K–$250K+ range
•Built for factories, warehouses, logistics centers
•Sold via enterprise sales, often leased (Robot-as-a-Service)
•Key players: Agility Digit, Figure 02, Apptronik Apollo, UBTECH Walker S2

Best for: Facility operators piloting humanoid automation for dangerous or repetitive tasks

🏠 Home / Personal

•Consumer-facing, sub-$25K
•Designed for household tasks — cleaning, carrying, basic assistance
•Direct purchase or subscription model
•Key players: 1X NEO (~$20K), Unitree R1 (~$5K)

Best for: Early adopters who accept limitations and want a personal humanoid at home

🔬 Research / Dev Platform

•Sold to labs, universities, and developers, $13K–$150K
•General-purpose experimentation hardware
•Open SDK/API, designed for custom software development
•Key players: Unitree G1/H1/H2, Fourier GR-2

Best for: Researchers and developers building on humanoid hardware

The humanoid landscape: who builds what

Each dot is a humanoid robot in our directory, plotted by height and weight and colored by market segment. Click any dot to visit its profile. Industrial humanoids cluster heavier and taller; home robots are lighter; research platforms span the widest range.

What you’re actually paying for

Realistic first-year costs

A warehouse operator leasing an Agility Digit for material handling — expect $150K–$200K year one including integration, training, and support. A university buying a Unitree G1 for locomotion research — $16K–$30K upfront plus grad student time.

$5K–$250K+

Price range across all segments

1–3 hrs

Typical battery runtime

20–44

Degrees of freedom

3–12 mo

Enterprise deployment lead time

Humanoid pricing is still volatile. Industrial units are mostly lease-based (Robot-as-a-Service). Home robots are pre-order with consumer pricing. Research platforms are the most straightforward — you buy hardware, you own it.

You’re not ready if…

You need 24/7 autonomous uptime starting next month
You expect plug-and-play out of the box with no software customization
Your facility has no WiFi, narrow aisles, and multi-floor layouts with no elevator
You don’t have budget for integration, training, or a pilot period
Your use case is better served by a cobot arm or mobile robot — cheaper, proven, available now

Consider it if…

You have a controlled pilot environment — single zone, defined task
You’re evaluating on a 12-month horizon, not expecting ROI in week one
Your task is repetitive, dangerous, or happens in human-scale spaces
You have an innovation budget or research grant earmarked for this
You want to build internal expertise before the market matures
You’re a researcher or developer who wants hands-on humanoid hardware

The honest timeline

Shipping today: Agility Digit (warehouse logistics), Figure 02 (BMW factory pilot), Unitree G1 (research labs worldwide), 1X NEO (pre-orders open, first home deliveries 2026).

12 months out: Broader enterprise availability, second-gen models from Figure and Apptronik, Tesla Optimus likely entering limited external pilot, Unitree R1 shipping.

The honest caveat: Runtime is short (1–3 hours), autonomy is limited, and software is evolving fast. Early adopters get a seat at the table — but it’s still early. If you need a robot that works reliably for full shifts today, a cobot arm or AMR is the pragmatic choice.

Before you commit

Request a live demo or pilot from the manufacturer
Define your pilot scope: one task, one zone, one shift
Budget 2–3× the hardware cost for integration, software, and support
Set expectations with your team — this is a pilot, not a deployment
Talk to other early adopters (directory reviews, manufacturer case studies)
Compare against simpler alternatives: would a cobot arm or AMR solve this faster?

Browse humanoids