I'm LongbridgeAI, I can summarize articles.Anthropic's study reveals a significant gap between AI capabilities and economic viability, estimating robots can perform 74% of physical tasks but remain cost-effective for only 0.3%. Full cost competitiveness may take decades, contrasting with Nvidia's optimism. While LLMs impact office work, physical automation faces hardware cost barriers. Meanwhile, Amazon (AMZN), Anthropic's largest shareholder, holds a 'Strong Buy' consensus with a $331.82 price target.
AI company Anthropic has found a large gap between what robots can technically do and what companies can afford to automate. Indeed, its new study estimates that today's robots can perform 74% of physical work tasks in the U.S., equal to about 34% of total working hours. However, robots are cheaper than human workers for only 0.3% of work today. If robot prices keep falling by their historical rate of roughly 3% per year, Anthropic estimates that reaching 10% cost competitiveness could take about 40 years.
This matters because seeing a robot perform a task does not mean replacing a worker makes financial sense. Most robots still need structured workplaces that are designed around them. Others require expensive hardware or struggle with tasks that need precise hand movements.
For example, Anthropic estimates that robots can perform 97% of the work done by packers and packagers. In that occupation, automation is already close enough to human labor costs that replacement can make financial sense.
For welders and many cleaning jobs, the economics look very different. The robot systems required can still cost several times more than human labor.
Even Aggressive Robot Progress Could Take Decades
Anthropic also modeled a much faster scenario where robot costs fall by up to four times faster than their historical pace and machines learn new tasks twice as quickly. Even under those assumptions, robots do not become cheaper than humans for half of today's physical work until around 2050. Reaching 90% would take even longer.
That timeline runs counter to some of the optimism around humanoid robots. For instance, Nvidia (NVDA) CEO Jensen Huang has predicted a major breakthrough period for robotics as better AI gives machines stronger reasoning abilities. However, Anthropic argues that intelligence solves only part of the problem. Hardware still needs to get cheaper, and robots need better physical skills.
However, office work is different. Large language models such as Claude are software, so companies can copy and distribute them far more cheaply than physical machines. Anthropic estimates that roughly half of work tasks are exposed to LLMs, which means that AI could cut the time required for those tasks by half. Still, that measures potential productivity rather than direct job losses.
Interestingly, the company's earlier research found that there was no systematic rise in unemployment among highly exposed workers since late 2022, although hiring among younger workers showed some signs of slowing down.
Is AMZN Stock a Buy, Sell, or Hold?
Currently, investors can't invest directly in Anthropic since it's still privately held. However, investors can buy shares of its largest shareholder, Amazon (AMZN). Turning to Wall Street, analysts have a Strong Buy consensus rating on AMZN stock based on 39 Buys and two Holds assigned in the past three months, as indicated by the graphic below. Furthermore, the average AMZN price target of $331.82 per share implies 34.6% upside potential. (See AMZN Stock Forecast).
