Key Takeaways
- • Raj Singh, managing partner at JLL Spark Global Ventures, told Commercial Observer that robotics has had almost no impact on real estate outside warehouses and distribution centers, with floor-cleaning robots in hospitals and airports the only significant deployment so far.
- • Singh identified labor shortages as the core driver of interest, citing close to half a million unfilled U.S. construction jobs and high turnover and absenteeism in facilities management roles.
- • Singh said the shift toward multi-task robots depends on "physical AI," but a physical-action equivalent of the large language model and dexterity remain unsolved, so he declined to predict when robots will become mainstream.
- • JLL Spark has not invested in any robotics company to date; Singh said its focus is on the software and services that connect building systems to robots and coordinate mixed human-robot workforces.
Summary
Raj Singh, managing partner at JLL Spark Global Ventures, offered a cautious assessment of robotics in real estate in a Commercial Observer interview. Asked about current impact, Singh said that outside warehouses and distribution centers the effect is "almost not at all." The robotics sector is largely in a pilot phase, with cleaning robots in large, open buildings such as hospitals and airports the only use case with significant deployment.
Singh said the main driver of interest is labor. He noted that close to half a million construction jobs in the United States remain unfilled, and that facilities management roles carry low pay, high turnover and absenteeism. Robots are seen as a potential substitute for hard-to-fill roles and may be more cost-effective over time. Historically, however, robots performed single tasks, whereas human jobs combine many. Singh said "physical AI," which pairs robots with an AI system capable of multiple tasks, has changed that equation, and that tasks such as restocking, cleaning kitchens and resetting meeting rooms could become feasible.
Several barriers remain. According to Singh, robotics needs an equivalent of the large language model trained on physical actions, since LLMs draw on the internet's written material and no comparable dataset of physical tasks exists. Approaches such as world models and visual-language-action models are in development. Dexterity is also unresolved, and Singh said it is "a fool's game" to predict when it will be achieved.
On China, Singh said the country leads in industrial robot deployment and may be leading in the development of advanced robotics, partly because robotics is a national priority and supply chains are concentrated locally. He said Chinese robots have not solved the world-model or dexterity problems and are not yet performing these real-estate tasks. Construction sites, he added, are harder environments than factory floors, and even Chinese robots being trialed by companies are "not quite ready for prime time," which he sees as an opening for U.S. and European robotics firms.
Singh also raised business-model concerns, saying that if robotics companies capture all the margin, building owners gain little from deploying robots. JLL Spark has not invested in any robotics company, and Singh said the firm is not expert enough to pick winners. Its interest lies in software that translates building management system outputs, such as work orders, into instructions a robot can act on, and in orchestration software for mixed human-robot environments. Singh said he does not expect robots to replace much of the workforce soon, anticipating hybrid workplaces instead.
Singh said the main driver of interest is labor. He noted that close to half a million construction jobs in the United States remain unfilled, and that facilities management roles carry low pay, high turnover and absenteeism. Robots are seen as a potential substitute for hard-to-fill roles and may be more cost-effective over time. Historically, however, robots performed single tasks, whereas human jobs combine many. Singh said "physical AI," which pairs robots with an AI system capable of multiple tasks, has changed that equation, and that tasks such as restocking, cleaning kitchens and resetting meeting rooms could become feasible.
Several barriers remain. According to Singh, robotics needs an equivalent of the large language model trained on physical actions, since LLMs draw on the internet's written material and no comparable dataset of physical tasks exists. Approaches such as world models and visual-language-action models are in development. Dexterity is also unresolved, and Singh said it is "a fool's game" to predict when it will be achieved.
On China, Singh said the country leads in industrial robot deployment and may be leading in the development of advanced robotics, partly because robotics is a national priority and supply chains are concentrated locally. He said Chinese robots have not solved the world-model or dexterity problems and are not yet performing these real-estate tasks. Construction sites, he added, are harder environments than factory floors, and even Chinese robots being trialed by companies are "not quite ready for prime time," which he sees as an opening for U.S. and European robotics firms.
Singh also raised business-model concerns, saying that if robotics companies capture all the margin, building owners gain little from deploying robots. JLL Spark has not invested in any robotics company, and Singh said the firm is not expert enough to pick winners. Its interest lies in software that translates building management system outputs, such as work orders, into instructions a robot can act on, and in orchestration software for mixed human-robot environments. Singh said he does not expect robots to replace much of the workforce soon, anticipating hybrid workplaces instead.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.