Opens in a new tab
needsahuman
Search ⌘KGet the data
Guide · updated 2 October 2026

Will humanoid robots take physical jobs?

Where industrial and humanoid robots really stand, why hands-on work is so hard to automate, and which physical jobs to watch.

Key points

  • Factory robots are common and growing fast. Humanoid robots are still in early, small deployments.
  • Robots do best where the setting is controlled and the task repeats. Homes, building sites and hospitals are neither.
  • What is easy for people, such as seeing, grasping and moving through a messy room, is still hard for machines.
  • Physical jobs are much less exposed to today’s AI than desk jobs.
  • Watch warehouses and factories first. Trades and care work are much further off.

Two very different kinds of robot

When people hear “robots will take jobs”, they picture a humanoid walking onto a building site. The robots doing real work today look nothing like that.

Industrial robots: already everywhere in factories

The International Federation of Robotics (IFR) counts industrial robots worldwide. Its September 2026 report found that “the global operational stock of industrial robots surged 9% to a record 5 million units in 2025,” with factories installing “more than 600,000 new units” that year.

China installed “354,000” of those, “59% of global deployments.” US installations “grew by 12% to almost 38,500 units.” The IFR expects “655,000 units in 2026” worldwide.

Robot density shows how far this has gone in each country. In 2024 the global average was 132 robots per 10,000 factory employees. The US “ranks 8th worldwide with 307 units.” South Korea leads with “1,220.”

These robots weld, paint, lift and assemble. They are fast and precise, but they work in fenced cells, on fixed tasks, in places designed around them.

Humanoid robots: early commercial pilots

Humanoid robots are built to work in spaces designed for people. A few have moved from demos into paid work:

  • Agility Robotics’ Digit at GXO. The logistics company GXO called its 2024 deal “the industry’s first formal commercial deployment of humanoid robots.” Digit’s job was moving totes “from cobots and placing them onto conveyors” in a warehouse.
  • Figure at BMW. Figure AI reports an “11 month Figure 02 robot deployment at BMW Group Plant Spartanburg,” loading sheet metal parts, “placing parts within a 5-millimeter tolerance in just 2 seconds,” over “1,250+ operational hours.” This is the company’s own account.

Both are narrow, repeated tasks in structured industrial settings. That is the pattern to expect for some time: humanoids first take on the warehouse and factory jobs that fixed robots could not reach, not the varied work of a plumber or a nurse.

Why physical work is so hard to automate

Moravec’s paradox

In 1988 the roboticist Hans Moravec observed that “it is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility” (Mind Children, Harvard University Press, 1988, as quoted on Wikipedia).

Nearly 40 years later, AI writes essays and code, while folding laundry or fixing a leaking pipe in a stranger’s kitchen remains very hard for a machine. The tasks humans find easy took millions of years of evolution to perfect.

Messy, changing places

A factory floor is the same every day. A home, a roof or a hospital ward is not. Each job is a little different, objects are in unexpected places, and people move around. Robots need to handle all of that safely.

Safety, cost and trust

A robot that drops a box in a warehouse costs money. A robot that drops a patient is a disaster. Physical work around people needs very high reliability, plus insurance, rules and public trust. Robots are also expensive to buy, maintain and repair.

What AI usage data says about physical jobs

Today’s AI assistants barely touch hands-on work. Microsoft Research found the occupations with the lowest AI applicability “include occupations that require physically working with people, operating machinery, and other manual labor.” Anthropic’s first Economic Index report found the “farming, fishing, and forestry” category made up “0.1% of queries.”

Even OpenAI’s GDPval benchmark, which tests AI on real work from 44 occupations, says that “manual labor and physical tasks are not included in the current version” (Patwardhan et al., 2025).

Physical jobs still use AI for the paperwork around them: quotes, scheduling, manuals, reports. But the core work is untouched by chat assistants.

Which physical jobs to watch

Use this rough order, from nearest to furthest:

  1. Repeated tasks in factories and warehouses. Moving totes, loading parts, picking and packing. Fixed robots already do much of this; humanoids are being piloted for the rest.
  2. Driving and moving goods on set routes. Long-haul freight, yard trucks and delivery in controlled areas.
  3. Cleaning large, simple spaces. Floors in malls, airports and warehouses.
  4. Construction tasks that repeat. Bricklaying on simple walls, drilling, finishing.
  5. Skilled trades in varied settings. Electricians, plumbers, HVAC technicians, roofers. Every job site is different.
  6. Hands-on care. Nursing, home health, physical therapy. Physical skill plus trust plus safety.

How we score robots on NeedsAHuman

Every job page has a Robots and humanoids section. It shows:

  • how much of the job’s task time is physical,
  • what kind of robot each physical task would need: fixed automation, a mobile robot or a dexterous humanoid,
  • where that technology stands today.

Physical tasks have a cap on how far our Can AI do it? score can go, set by what robots can actually do. Our timeline method assumes physical work progresses more slowly than desk work.

What to do if you work with your hands

  • Lean into the varied, judgment-heavy parts of your trade: diagnosis, problem-solving, dealing with customers.
  • Learn the new tools. Diagnostic software, AI-assisted estimating and smart building systems are becoming part of trade work.
  • Look at the growing trades. See AI and trades careers.
  • If you work in a warehouse or factory, learn to work alongside robots: setup, monitoring, maintenance and repair are growing roles.

Sources

How long will your job last?

Answer a few questions about how you actually do your job and get your own Still needs a human score, with the range, the risks and three things to do next.