WholeReach · industry directory

Robotics

The companies that define the robotics industry, each read on the same nine-check website rubric — a 0–100 score with the exact fixes — alongside the concepts, the developments and the questions that shape the field.

See the directoryGet your free score

The companies

33 companies in the robotics industry, each graded on the nine things that decide whether a buyer or an AI assistant can find and trust them online. Open any name for the full write-up and the specific fixes.

70+ solid   45–69 gaps   under 45 costing customers

Sites we couldn’t reach

6 businesses whose website blocked our scanner — common when a site filters datacenter traffic, and not a sign the site is broken. A blocked site is never scored zero.

Who defines the robotics market

The firms, funds and institutions that set the terms in robotics — context, not local listings; never scored.

Robotics guides

Concepts

Payload

Maximum weight a robot can carry at its wrist, including the gripper and cabling — not just the part. A 10 kg arm with a 4 kg gripper moves 6 kg of product.

Reach

Maximum distance from base to wrist. Reach and payload trade against each other on nearly every arm.

Work envelope

The 3D volume the robot can physically reach. Cell layout, fixturing and guarding are all designed around it.

DOF (degrees of freedom)

Independent axes of motion. Six is the minimum to place a part at any position and orientation; a seventh adds redundancy for reaching around obstacles.

Repeatability

How closely a robot returns to the same taught point, typically ±0.02 to ±0.1 mm. The spec most often quoted — and not the same as accuracy.

Accuracy

How closely the robot reaches a commanded position in real-world coordinates. Always worse than repeatability, and it matters when programming offline from CAD.

TCP (tool center point)

The reference point at the working end of the tool that the controller actually moves and rotates about. A wrong TCP is the most common cause of a program that looks right and runs wrong.

End effector / EOAT

End-of-arm tooling — the gripper, torch, nozzle, sander or vacuum array bolted to the wrist. It is what touches the part, and where most application engineering goes.

Cobot (collaborative robot)

A robot with force/torque sensing and speed limits so it can share a workspace with people, often without full fencing. 'Collaborative' describes the application and its risk assessment, not the robot alone.

Articulated robot

The classic jointed arm with rotary axes, usually six — the most flexible and most common industrial configuration.

SCARA

Selective Compliance Assembly Robot Arm — four axes, rigid vertically and compliant horizontally, used for very fast planar pick-and-place and assembly.

Delta robot

A parallel-linkage overhead robot with very low moving mass, used for high-speed picking of light parts. Typical in food and packaging.

Cartesian / gantry robot

Moves on linear X-Y-Z axes over a rectangular envelope. Favoured for large work areas, heavy loads and straight-line motion.

Teach pendant

The handheld device used to jog a robot and record positions. Modern equivalents use tablet-style flowchart or no-code interfaces.

Offline programming (OLP)

Writing and validating programs in simulation from CAD instead of on the live machine, so the cell keeps producing while the next job is developed.

Cycle time

Elapsed time for one full work cycle. It converts directly into parts per hour and therefore into the business case.

Throughput / UPH

Units per hour actually delivered — the operations metric, distinct from theoretical cycle time because it includes faults, changeovers and waiting.

Duty cycle

Proportion of time a robot or axis runs at load. Sustained high duty cycles derate allowable payload and shorten maintenance intervals.

Singularity

An arm pose where two axes align and the robot loses a degree of freedom, causing wild joint speeds or a fault. Programs are planned to route around them.

Kinematics (forward/inverse)

Forward computes where the tool ends up from joint angles; inverse computes the joint angles needed to reach a pose. Inverse solutions are often multiple — hence 'elbow up' vs 'elbow down'.

Harmonic / strain wave gearing

A compact zero-backlash reduction gear using a flexible spline — the standard drivetrain in precision robot joints. Harmonic Drive is the originating trademark.

Cycloidal reducer

A high-ratio, shock-tolerant gear used in the heavier axes of industrial arms. Nabtesco dominates this category.

Servo drive and encoder

The amplifier controlling each joint motor and the feedback device reporting its position. Absolute encoders let a robot know its pose at power-up without re-homing.

Machine tending

Loading and unloading a CNC machine, press or molder — the single most common entry-level industrial robot application.

Palletizing / depalletizing

Stacking or unstacking cases, bags or totes in a defined pattern. High payload, low precision, and a very common first cobot project.

Bin picking

Using 3D vision to locate and grasp randomly oriented parts from a bin. Historically one of the hardest problems in industrial robotics and a major beneficiary of AI perception.

Machine vision

Cameras plus processing used to locate parts, guide the robot, read codes and inspect quality. Vision-guided means the camera supplies coordinates rather than the fixture.

Force/torque sensing

Measuring contact forces at the wrist so the robot can insert, polish, assemble by feel, or stop safely on contact. What makes compliant assembly and sanding practical.

SLAM

Simultaneous Localization and Mapping — lets a mobile robot build a map of a facility while tracking its own position, without floor markers.

AMR vs AGV

An AGV follows fixed infrastructure (tape, wire, reflectors) and stops when blocked. An AMR navigates by onboard sensing and maps, reroutes around obstacles, and can be redeployed without changing the building.

Fleet management / traffic control

The software assigning missions, deconflicting routes, managing charging and reporting across a population of mobile robots. On multi-robot sites it matters more than the individual robot spec.

Goods-to-person (G2P)

A fulfillment model where robots bring shelves or totes to a stationary picker, replacing walking time. The core of most warehouse robotics ROI claims.

ASRS

Automated Storage and Retrieval System — dense automated storage such as cube, shuttle or crane systems, often paired with robotic picking at the workstation.

ROS / ROS 2

The Robot Operating System — open-source middleware and tooling (nodes, topics, drivers, navigation and manipulation stacks) that became the de facto research and increasingly commercial foundation. ROS 2 is the production-oriented rewrite.

Digital twin / simulation

A physics-accurate virtual model of a robot, cell or facility used to design, validate and train before touching hardware, cutting commissioning risk and time.

Sim-to-real

Training a control policy in simulation and transferring it to a physical robot. Domain randomization and synthetic data are the standard tricks for closing the reality gap.

Robotic foundation model / VLA

A large pretrained model — often vision-language-action — mapping camera and instruction inputs directly to robot actions, aiming to generalize across tasks instead of being programmed per task.

ISO 10218-1 / -2

The core international robot safety standards, revised in 2025: Part 1 covers robot manufacturers, Part 2 covers integrators, applications and cells. The revision absorbed the separate collaborative-operation guidance and added cybersecurity requirements.

Power and force limiting (PFL)

The collaborative mode where contact forces are limited to biomechanically safe thresholds, so incidental contact with a person does not cause injury.

Speed and separation monitoring (SSM)

A collaborative mode where safety-rated sensing keeps a protective distance between human and robot, slowing or stopping as the person approaches.

Safety-rated monitored stop

The robot holds position with drives energized while a person is in the workspace, resuming automatically when they leave — no restart sequence.

Risk assessment

The mandatory hazard analysis of a specific application that determines guarding, safety functions and required performance level. Safety is a property of the installed cell, never of the robot alone.

Systems integrator

The engineering firm that designs, builds, programs and commissions the production cell around a robot. Most industrial robots are bought through integrators, making them a channel as well as a competitor.

Seventh axis / track

A linear rail or turntable extending a robot's reach or serving multiple stations from one arm, treated by the controller as an additional coordinated axis.

Tool changer / quick changer

A coupling letting one robot swap between multiple end effectors automatically, passing power, air and signals across the joint — the key to multi-product flexibility.

RaaS (Robotics-as-a-Service)

Paying a recurring per-robot, per-pick or per-hour fee instead of buying capital equipment. Moves robotics from capex to opex; now standard in warehouse and increasingly in welding and humanoid deployments.

OEE

Overall Equipment Effectiveness — availability × performance × quality. The plant-floor scorecard an automation project is ultimately judged against.

What’s happening

• The IFR's World Robotics 2025 report recorded about 542,000 industrial robots installed worldwide in 2024, more than double the figure of ten years earlier, with annual installations topping 500,000 for the fourth consecutive year. The 2026 edition does not publish until late September 2026, so these remain the current figures. The installed base, not the annual number, is now the story: a maturing market with an enormous service, retrofit and upgrade opportunity. 2026-08 source ↗
• China accounted for roughly 295,000 installations in 2024 -- about 54% of the global total and the highest annual national figure on record -- and its operational stock passed two million units, the largest of any country. For the first time Chinese robot manufacturers outsold foreign suppliers in their home market, taking about 57% domestic share, up from 30% in 2020. 2026-08 source ↗
• Regional concentration is extreme: Asia took about 74% of new 2024 deployments, versus 16% in Europe and 9% in the Americas. Western vendors increasingly compete on integration expertise, service coverage and reshoring incentives rather than unit price. 2026-08 source ↗
• ABB is exiting robotics by sale, not by spin-off -- a change worth getting right, because the earlier plan is still widely repeated. ABB announced in April 2025 that it would spin the division off as a separately listed company targeting a Q2 2026 listing, then abandoned that route: on October 8, 2025 SoftBank Group agreed to acquire ABB's robotics business outright for $5.375 billion, covering roughly 7,000 employees and about $2.3 billion of 2024 revenue. As of August 2026 the sale has not closed and remains subject to regulatory approval, with completion expected mid-to-late 2026. There is no separately listed ABB Robotics company and there will not be one. 2026-08 source ↗
• The revised ISO 10218-1:2025 and ISO 10218-2:2025 safety standards were published in February 2025 after roughly eight years of revision work, replacing the 2011 editions and splitting obligations between robot manufacturers (Part 1, robots as incomplete machines) and system integrators and machine builders (Part 2, complete systems). The revision folds in the previously separate collaborative-operation guidance of ISO/TS 15066, clarifies functional safety by allowing alternatives to the former mandatory performance level, and adds cybersecurity requirements -- the first time cyber has been written into core robot safety. The practical compliance driver for European-market equipment is the EU Machinery Regulation, which applies from January 2027. 2026-08 source ↗
• Venture funding into robotics has hit records. Crunchbase counted roughly $18.8 billion raised by robotics startups in 2026 as of late June, against about $15 billion for all of 2025 on the same definition -- and definitions vary widely between trackers, so the figures are directional rather than settled. Capital availability lets well-funded newcomers underprice incumbents on pilot deployments. 2026-08 source ↗
• Investor attention is rotating within the frontier, from the body to the brain. One hand-compiled tracker of disclosed physical-AI equity rounds puts humanoids at roughly 58% of year-to-date capital in 2025 falling to about 43% year-to-date in 2026, while robotic foundation models rise from roughly 15% to roughly 29% over the same windows. The dataset is small and the precision is not to be trusted, but the direction is consistent with what the largest rounds have funded: general-purpose policies that could run on many form factors rather than any single body. 2026-08 source ↗
• 'Physical AI' has become the organizing frame, built on a three-computer architecture: one system to train models, one to simulate, one on the robot to infer. NVIDIA describes exactly this split and has assembled a stack around it -- Isaac Lab for robot learning, Isaac Sim for simulation and synthetic data, Isaac ROS for CUDA-accelerated ROS 2, Isaac GR00T for humanoid foundation models and Jetson Thor at the edge -- which has made a chip vendor one of the most influential platform players in robotics. 2026-08 source ↗
• ROS has been governed since March 2024 by the Open Source Robotics Alliance, with Intrinsic, NVIDIA and Qualcomm as its inaugural platinum members and Apex.AI, Zettascale, Clearpath, Ekumen, eProsima and PickNik among the other founding tiers. The OSRF board assigned the alliance responsibility for ROS, Gazebo, Open-RMF and their supporting infrastructure. This formalizes commercial stewardship of ROS 2 and reduces the single-point-of-failure risk that long worried commercial adopters. 2026-08 source ↗
• Alongside the funding boom there has been a real shakeout. Zebra moved in December 2025 to dispose of or exit the mobile robot business it built on its $290 million Fetch Robotics acquisition, with most robotics staff gone by the end of 2025; READY Robotics shut down at the start of 2025; and Rethink Robotics, the original cobot pioneer, has now closed twice -- in 2018 and again in 2025 after a relaunch. Buyers now routinely ask about balance-sheet durability and spare-parts guarantees, and content addressing vendor longevity converts. 2026-08 source ↗
• Public-market access is opening for robotics pure-plays. Geek+ listed on the Hong Kong Main Board in July 2025, describing itself as the first publicly listed company in the AMR warehouse robotics market. Agility Robotics announced on June 24, 2026 that it would go public through a merger with Churchill Capital Corp XI at a $2.5 billion pre-money equity value with more than $620 million of gross proceeds; as of August 2026 the deal is announced but not closed, with an S-4 confidentially submitted in July and shareholder and regulatory approvals outstanding. Listings bring disclosure that gives buyers unprecedented visibility into vendor economics. 2026-08 source ↗
• Collaborative robots remain a growing minority of installations -- IFR put cobots at 10.5% of the 541,302 industrial robots installed in 2023 -- and the payload race has moved decisively upmarket. Universal Robots' UR30 is rated at 30 kg, reaching 35 kg in certain mounting and orientation conditions, on a 1,300 mm reach, and heavy palletizing-class collaborative arms are now offered by Doosan and others. Be wary of cobot market-size forecasts: published estimates vary by three to four times depending on the research firm's definition, and IFR itself declines to claim cobots are outgrowing traditional industrial robots, saying only that they will complement rather than replace them. 2026-08 source ↗
• Agricultural robotics has moved from pilots to production economics, with laser weeding and autonomy retrofits displacing chemical and mechanical methods. Carbon Robotics reports deployments with more than 200 growers across 15 countries, selling both the LaserWeeder G2 and an autonomy retrofit kit for existing John Deere 6R, 8R, 8RX and 8RT tractors under remote operator supervision. John Deere announced its second-generation autonomy kit and an autonomous 9RX for tillage at CES 2025, slated for limited release in 2026 rather than broad availability. Field robotics is the clearest case where the labor shortage, not the technology, is the binding constraint. 2026-08 source ↗
• The humanoid category is usually described as splitting into consumer-home and industrial-labor tracks, and the split is real but leaky -- Figure markets toward the home while its flagship industrial deployment is with BMW. The verifiable industrial pairings are worth keeping straight: Agility's Digit is deployed with GXO, Schaeffler and Toyota Motor Manufacturing Canada, and Agility reports more than $300 million in secured Digit v5 orders; Apptronik's Apollo works under a commercial agreement with Mercedes-Benz delivering assembly kits and inspecting components; UBTECH's Walker S runs with Chinese and other Asian manufacturers rather than the Western partners often listed alongside it. Digit's work at Amazon was a limited R&D test at a Sumner, Washington facility, not a production deployment. Pilot-to-production conversion, not demo video, has become the credibility test. 2026-08 source ↗

Questions people ask

What is the difference between a cobot and an industrial robot?

the difference is the safety strategy, not the machine class: a cobot is designed for power-and-force-limited operation so it may share space after a risk assessment, while an industrial arm is faster and stronger and normally guarded -- and any robot can still require a fence depending on the tool and the task

What do payload, reach and repeatability actually mean when comparing robots?

payload must include the end effector and any part held, reach is measured to the wrist not the tool tip, and repeatability is return-to-the-same-point precision -- not accuracy to a commanded coordinate, which is a different and usually worse number

What is the difference between repeatability and accuracy?

repeatability is consistency returning to a taught point; accuracy is how close the robot gets to a coordinate it was told to reach; explain why offline programming needs accuracy and teach-pendant work only needs repeatability

What is an end effector and why is it often the hard part?

describe EOAT as the gripper, tool or vacuum array that touches the part, and explain that it is usually custom, drives cycle time and reliability, and is where integration projects most often need rework

What is the difference between an AGV and an AMR?

an AGV follows a fixed guide path and needs infrastructure; an AMR navigates and re-plans around obstacles using onboard sensing and a map -- the practical difference is what happens when a pallet is left in the aisle

What is SLAM and why does it matter for mobile robots?

simultaneous localization and mapping -- the robot builds a map and locates itself in it at the same time; matters because it is what removes the need for tape, magnets or reflectors

What is machine tending and why is it the most common first robot application?

loading and unloading a CNC or press; it is common because the task is repetitive, the fixture already exists, the cycle is predictable, and one operator can then run several machines

What is bin picking and why is it harder than it looks?

explain the stack: 3D vision to find pose in clutter, collision-free path planning into a bin wall, a gripper that tolerates pose error, and the singulation problem -- and note cycle time and pick reliability degrade as the bin empties

What is ROS and ROS 2, and is it production software?

describe ROS as a middleware and tooling ecosystem rather than an operating system, note ROS 2's real-time and security improvements, and say it is used in shipping products but that safety functions are not implemented in it

What is OEE and how does a robot cell affect it?

availability times performance times quality; a robot usually improves performance and quality consistency, and can hurt availability early on -- which is why baseline OEE before automating is worth measuring

What is a systems integrator and why can I not just buy the robot?

explain that the arm is a fraction of a working cell -- tooling, vision, safety, fixturing, controls, guarding and programming are the rest -- and that the integrator owns the cell working, which the robot OEM does not

What is Robotics-as-a-Service?

describe it as robots delivered on a subscription including hardware, software, support and often uptime commitments -- shifting capital cost to operating cost and shifting maintenance risk to the vendor

What is the payback period on an industrial robot?

give the calculation honestly -- fully burdened labor hours displaced plus scrap and rework reduction plus throughput gain, against total installed cost including tooling, integration, safety and training -- and refuse to quote a payback figure, because it swings with shift count, wage rate and utilization the reader has

What does a complete robot cell cost, not just the arm?

break the cost into arm, end effector, vision, safety and guarding, fixturing and conveyance, controls integration, programming, installation and training, plus commissioning downtime; give the structure and say the arm is typically a minority of it, without asserting a ratio as universal

How do I calculate ROI on automation when my labor is already short-staffed?

point out the framing error: if the position is unfilled, the benefit is throughput and missed-order recovery rather than labor savings, and the model has to value output you currently cannot produce -- which requires the reader's demand data

What is a realistic timeline from purchase order to production on a robot cell?

describe the phases -- design and simulation, procurement lead times, build, factory acceptance test, install, site acceptance, ramp -- and say the schedule is usually set by tooling design and long-lead components, so ask the integrator for the critical path

Watch

Run a robotics company?

Get your free website score and the exact list of what to fix — no charge, no pitch. Claim your listing or ask for the full report.