![\"\"](\"https:\/\/www.newsbeep.com\/uk\/wp-content\/uploads\/2026\/01\/https:\/\/d1e00ek4ebabms.cloudfront.net\/production\/e14713ba-5d06-47b0-b4cb-cf290db7ba92.png\")
<\/p>\nStay informed with free updates<\/p>\n
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.<\/p>\n
Barclays has published a Special Report: \u2018AI gets physical\u2019. Here\u2019s the opening pitch:<\/p>\n
AI\u2019s next frontier is physical: humanoid robots \u2013 robots in human form \u2013 are stepping out of the lab and into the real world. They could take on the tough, repetitive jobs humans increasingly avoid in manufacturing, agriculture and healthcare \u2013 working alongside people and augmenting the workforce.<\/p>\n
Authors Zornitsa Todorova and Carlos Eduardo Garcia Martinez write that a 30x drop in unit costs over the past decade means that the market for these \u201chumanoids\u201d \u2014 which they describe somewhat imprecisely as \u201ccars in miniature\u201d \u2014 could grow from $2\u20133bn today into a (\u201cmost-optimistic\u201d scenario) $200bn industry by 2035. The tailwinds include ageing populations, growing urbanisation and shifting job preferences.<\/p>\n
<\/p>\n
Obviously, the calculations underpinning that figure \u2026<\/p>\n
Estimating the total addressable market (TAM) is challenging, and projecting its future even harder given the many moving parts. To address this, we aggregated a broad range of sources from AlphaSense \u2013 company reports, industry articles, white papers, earnings transcripts, and expert commentary \u2013 to build a forecast model across three scenarios: baseline, conservative, and optimistic.<\/p>\n
\u2026 are very vague.<\/p>\n
The big breakthrough (they reckon) is AI, which (they reckon) is good enough that (they reckon) these robots now (they reckon) understand context. They are significantly more optimistic than, say, MainFT\u2019s Sarah O\u2019Connor, who wrote about this same issue earlier this week<\/a>. <\/p>\n You can read the report in full here<\/a>. We just wanted to dwell on one part. The authors write:<\/p>\n [Humanoid robots\u2019] advantage lies in near-continuous operation: even at half human efficiency, a humanoid can deliver 25% more output per day, and at parity, up to 150% more, according to our analysis.<\/p>\n They expand on this later on in a section called \u201cLabour force augmentation and productivity\u201d. The authors name some \u201croles with the greatest potential for humanoid robot adoption\u201d (high-res<\/a>):<\/p>\n And write (their emphasis):<\/p>\n While it is still uncertain whether humanoid robots can match human efficiency, their technical advantage lies in near-continuous operation. Unlike humans, a humanoid can run almost 24\/7, with only short breaks for battery charging. For example, assume a human in a logistics facility delivers 8 productivity hours per day. A humanoid operating in 6-hour cycles (5 hours of work followed by 1 hour of recharge) allows for 20 work hours in a 24-hour period. If its efficiency is only half that of a human, that equates to 10 productivity hours per day, or 25% more output than a human. If the humanoid matches human efficiency, its output jumps to 20 productivity hours per day, or 150% more than a human worker.<\/p>\n First, the technical part of this: that six-hour cycle proposal is based on the trailed \u201cpeak performance\u201d specs for Figure AI\u2019s F-03 battery<\/a>, designed for its Figure 03 humanoid, which very much does not yet exist as something a company can actually buy. Maybe it will exist someday, but this isn\u2019t exactly a hype-free sector<\/a>.<\/p>\n Now, the non-technical part. Here\u2019s another quote:<\/p>\n The prolongation of the working day beyond the limits of the natural day, into the night, only acts as a palliative. It quenches only in a slight degree the vampire thirst for the living blood of labour. To appropriate labour during all the 24 hours of the day is, therefore, the inherent tendency of capitalist production. But as it is physically impossible to exploit the same individual labour-power constantly during the night as well as the day, to overcome this physical hindrance, an alternation becomes necessary between the workpeople whose powers are exhausted by day, and those who are used up by night. This alternation may be effected in various ways; e.g., it may be so arranged that part of the workers are one week employed on day-work, the next week on night-work.<\/p>\n![\"\"](\"https:\/\/www.newsbeep.com\/uk\/wp-content\/uploads\/2026\/01\/https:\/\/d1e00ek4ebabms.cloudfront.net\/production\/088b58fc-245e-4d50-a668-a9f4a88b627f.png\")
It\u2019s notable that a job being \u201cemotionally demanding\u201d is supposed to make it *more suitable* for a robot <\/p>\n
![\"\"](\"https:\/\/www.newsbeep.com\/uk\/wp-content\/uploads\/2026\/01\/https:\/\/d1e00ek4ebabms.cloudfront.net\/production\/026aead6-9c98-4237-b2b3-96c61b639aca.png\")
<\/p>\n