Anthropic study says robots could reach 80% of work, but cost is the barrier

Anthropic says combining robots with AI models could make about 80% of US work technically achievable by automation. The striking catch: robots are currently cheaper than people for just 0.3% of tasks.

Anthropic Watch analysis

What happened

The study, as reported by The Next Web, estimates that AI models alone could reach about half of all work, while adding robots raises that figure to about 81%. The technologies affect different jobs: robots extend automation into physical roles that chatbots generally cannot do.

Transport stands out. The study puts AI-model exposure below 15% of tasks in the sector, but estimates that adding robots raises it to about 90%. Nine of the ten most exposed jobs are vehicle drivers, led by taxi drivers.

Cost, rather than technical capability, is the big constraint. The study estimates that robots can handle about 74% of physical tasks, often only in controlled settings such as assembly lines. About 300,000 US workers are in jobs where robots are already cheaper. If robot prices fell by 20%, that figure could rise to about 2.8 million. The Next Web says the estimates were made by Claude using web search.

Why it matters

“Could be automated” is not the same as “will be automated”. The figures sketch a possible reach, not a forecast of job losses, and the price gap helps explain why technical capability does not translate neatly into workplace adoption.

Still, a comparatively small drop in robot costs could bring millions more jobs into the study’s cheaper-to-automate category. That makes the economics of machines, alongside their abilities, central to the debate about which work changes first.

Our read

The 80% headline is the attention-grabber; the 0.3% figure is the reality check. Read this as a map of potential exposure, not a countdown to mass replacement. The useful question is which tasks robots can perform reliably and affordably outside carefully controlled environments.

What to watch

• Whether later studies test these estimates against real workplace deployments.
• How quickly robot costs fall, and which physical tasks become economical first.
• Whether the study’s job-exposure estimates are borne out in transport and other sectors.

Discussion spark: If robots became 20% cheaper, would the bigger priority be preparing workers for automation, or proving that robots can do these jobs reliably outside controlled settings?

Anthropic Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by Anthropic.