Japan today has 29.3% of its population aged 65 or older. This is the highest rate among large nations — countries with more than 40 million inhabitants — and it continues to rise. On a global scale across all populations, only Monaco (37%) ranks ahead. By 2040, the country will need to find 570,000 additional care workers to keep its nursing homes functioning. It does not have them. So for the past decade, it has bet on robots.
This bet was worth examining closely. A pooled cross-sectional study published in January 2026 in BMC Health Services Research by a team from Keio University provides a more nuanced picture than marketing brochure promises. The care robot is neither the imagined savior nor the useless impostor. It restructures the profession without replacing it. And this distinction matters for Europe, which will face the same demographic cliff in ten years.
The Essential Points
- Japan projects a shortage of 570,000 care workers by 2040, with 29.3% of the population already aged 65 or older — the highest rate among large nations
- The Keio University / BMC Health Services Research study (January 2026), a pooled cross-sectional study over three years examining organizational factors in robot adoption in Japanese nursing homes, provides valuable insight into the conditions for deploying such equipment
- Adoption faces concrete obstacles: equipment costs, lack of training, and cultural resistance slow large-scale deployment
- The fundamental problem remains political: immigration, training, and home care are the only levers capable of absorbing the demographic shock in the long term
A Country Facing Impossible Arithmetic
Japan has been aging longer than any other developed nation. In 2025, it has approximately 36 million people aged 65 or older, according to data from the Ministry of Internal Affairs. The dependency ratio — the number of working-age people for each retiree — has fallen below two to one in some rural regions. In a standard provincial nursing home, recruitment is now a matter of chronic urgency.
The projected shortage of 570,000 care workers by 2040 is not an alarmist projection: it is the difference between needs calculated by the Ministry of Health, Labour and Welfare and the number of professionals the system can reasonably expect to train and retain. The elderly care sector is notoriously difficult: physically demanding work, salaries below the national average, high turnover. A 2024 survey by the Japan Institute for Labour Policy and Training established that a significant proportion of care workers on the job report regular musculoskeletal pain, and that lower back disorders constitute the leading cause of work absence in the sector.
It is in this context, not in any abstract desire for modernization, that the Japanese government began actively financing the robotization of care facilities. The public support program for deploying care robots, launched in 2013, mobilized several tens of billions of yen in subsidies and research support. This is not a niche experiment: by the end of 2024, several hundred facilities reported using at least one robotic device in their care provision.
What Robots Actually Do in Nursing Homes
The Keio University study, published in January 2026 in BMC Health Services Research, is a pooled cross-sectional study over three years examining organizational factors associated with the adoption of care robots in Japanese nursing homes. It distinguishes four categories of equipment: transfer and lifting robots (for moving residents from bed or chair), wearable exoskeletons for care workers, nighttime surveillance and fall detection robots, and companion robots for cognitive stimulation.
Earlier work on lifting robots showed they significantly reduce mechanical strain on staff lower backs. In facilities that adopted them regularly, the rate of self-reported lower back pain decreased and work absences for musculoskeletal disorders declined over the observed periods. Exoskeletons produce comparable results, but their adoption remains lower: they are uncomfortable to wear for extended periods, and more experienced care workers tend to reject them, perceiving them as a constraint rather than an aid.
For surveillance robots, the outcome is useful but less dramatic: they detect nighttime falls faster than human rounds, which reduces intervention time. But they also generate false alerts, and night shift care workers report alert fatigue in facilities where the configuration has not been refined.
For companion robots, the picture is more ambiguous. Earlier Japanese studies, particularly on the PARO robot (a therapeutic robotic seal, commercialized since the early 2000s), had shown positive effects on anxiety and agitation in dementia patients. These short-term positive effects have been confirmed by other work, which notes however that they tend to diminish after several weeks of exposure. The companion robot does not create a relationship: it stimulates, it occupies, it temporarily soothes. This is not nothing, but it is not relational care.
Adoption Remains Blocked by Concrete Obstacles
If the gains are documented, why does adoption remain so limited? The Keio study identifies several main obstacles, and they are not technological.
The first is cost. A quality lifting robot costs between 500,000 and 1.5 million yen, or between 3,000 and 10,000 euros. Public subsidies cover part of the investment, but small facilities, which are also the most numerous in rural areas, often lack the cash flow for the initial purchase or the technical staff for maintenance. The Japanese market comprises several thousand facilities for the elderly. Yet, according to recent data, approximately two-thirds of tokuyo-type nursing homes had adopted at least one robot by 2022, and 63% used monitoring robots — notable progress but still leaving many facilities, particularly the smallest ones, outside this dynamic.
The second is training. Introducing a lifting robot into a care team requires learning time, adapted protocols, and a break-in period where productivity drops before rising again. In facilities chronically understaffed, this training time is perceived as an inaccessible luxury. The study documents several cases of purchased equipment that was underutilized due to insufficient support in learning to use it.
The third is cultural, and it is perhaps the most interesting. Some care workers, particularly the most experienced in the profession, resist the idea that care can be assisted by a machine. They see it as a threat to their professional expertise, a dehumanization of the gesture. This resistance is not irrational: it carries a correct intuition about what care requires in terms of presence, touch, attention. The problem is that it sometimes leads to rejecting tools that primarily protect the bodies of the care workers themselves.
The Augmented Caregiver, Not Replaced
It is here that the initial framing error in the public debate deserves to be named. The question “will robots replace care workers?” is the wrong question. This is not what the data shows, and it is not what Japan is building.
What automation produces in nursing homes, when it works well, resembles what it produces elsewhere in the economy: it moves the most mechanical, most physically costly tasks toward the machine and frees the professional for tasks that require judgment, relationship and adaptation. A care worker who does not have to single-handedly lift an 80-kilogram resident has more energy to talk to them, to observe their condition, to report a change. The quality of presence changes.
This pattern is consistent with what studies on automation in other sectors document: organizations that gain from automation are those that reorganize work around new capabilities, not those that simply add machines without changing processes. In care, this reorganization remains largely to be done.
The relevant question is therefore: what skills must the care worker master in a partially robotized environment? The Keio study suggests that the highest-performing facilities are generally those that have redefined the role of care workers around coordination, clinical assessment and relationship, entrusting the mechanics of lifting and monitoring to the machine. This requires upskilling, not substitution.
The Political Problem That Robots Cannot Solve
Care robots can reduce difficulty, improve certain outcomes for residents, and help understaffed facilities cope. They will not fill 570,000 missing positions.
Japan knows this. The political debates accompanying the robotization of the sector have for several years focused on three levers that machines cannot activate on their own.
The first is immigration. Japan long resisted the idea of significantly opening its borders to foreign labor. This resistance is eroding under demographic pressure. Visa programs for foreign care workers have been expanded since 2019, with specific pathways for Filipinos, Indonesians and Vietnamese. In 2024, the Kishida government further eased long-term residency conditions for qualified foreign care workers. This is a real change, but volumes remain modest compared to needs, and countries that recruit their care workers from abroad also export part of their problems to countries of origin.
The second lever is wage revaluation. Japanese care workers are paid significantly less than the national average, a gap documented for years by the Ministry of Labour. Several wage increases have been implemented since 2020, but they remain insufficient to attract in large numbers candidates who choose less physically demanding sectors. Without substantial revaluation, robotization risks improving the working conditions of care workers who do not exist.
The third lever is home care and home-based services. The model of large nursing homes is not the only possibility. Japan has been experimenting for several years with intermediate housing arrangements and pooled home services that allow relatively independent elderly people to remain at home longer, with scheduled care visits. This model is less costly, often better experienced by residents, and less dependent on a pool of staff concentrated in one facility. It poses significant logistical coordination problems, but it changes the structure of the problem.
What Europe Can Learn From the Japanese Experience
Germany currently faces approximately 35,000 vacant nursing positions, with projections of 280,000 to 690,000 shortages in the nursing sector by 2049 according to the BIBB. France is aging at a less brutal but inexorable pace. Italy and Spain have demographic dynamics similar to Japan’s from twenty years ago.
Japan is ten to fifteen years ahead on this curve, and its experience offers concrete lessons. Care robots work when they tackle specific, repetitive, primarily physical tasks, and when their deployment is accompanied by work reorganization. They do not work as a global substitute for absent care workers. And large-scale deployment requires active public policy: funding, training, standards, support for the most fragile facilities.
The mistake would be to view the robotization of Japanese care as a technical solution to a demographic problem. It is a tool in a response that remains fundamentally political. Choices on immigration, wages, and models of organizing care are societal choices, not engineering problems. The machine can change the conditions in which these choices arise. It does not make them on behalf of decision-makers.
The real interest of the Japanese assessment may be there: it shows that the right question is not “how many robots must we buy?” but “what care profession do we want to build for the next twenty years, and what tools must this renewed profession have at its disposal?”. This question, technology cannot answer alone.
Sources
- BMC Health Services Research / Keio University (January 2026): https://pmc.ncbi.nlm.nih.gov/articles/PMC12910932/
- Japanese Ministry of Internal Affairs and Communications — demographic statistics 2025
- Japanese Ministry of Health, Labour and Welfare — projections of care personnel needs 2040
- Japan Institute for Labour Policy and Training — survey on working conditions in the care sector, 2024
- Japanese government subsidies program for sector robotization (Robot Introduction Support Project, METI), annual reports 2013-2024
- MIC Japan / Japan Times – Population 65+ in 2025: https://www.japantimes.co.jp/news/2025/09/16/japan/society/japans-elderly-population/
- Cabinet Office Japan – Annual Report on Ageing Society FY2025: https://www8.cao.go.jp/kourei/english/annualreport/2025/pdf/2025.pdf
- WEF – Japan’s longevity economy (2025): https://www.weforum.org/stories/2025/09/japans-longevity-economy/
- BMC Health Services Research – Keio Study (Tosaka et al., January 2026): https://link.springer.com/article/10.1186/s12913-026-14082-4
- MHLW via US Trade.gov – Shortage of 570,000 care workers by 2040: https://www.trade.gov/market-intelligence/japan-healthcare-caregiving-technologies
- ISVD – Structure of Japan’s care worker crisis (2026): https://isvd.or.jp/en/columns/2026-03-09-care-worker-shortage-structure
- ScienceDirect – Robots and labor in nursing homes (2024): https://www.sciencedirect.com/science/article/pii/S0927537124001623
- AMED Japan – Care robot program launched in 2013: https://www.amed.go.jp/en/program/list/12/02/001.html
- FMC Group / BIBB-IAB – Care worker shortage Germany: https://fmcgroup.com/germany-skilled-worker-shortage/
- Visual Capitalist / UN WPP 2024 – Global ranking of seniors: https://www.visualcapitalist.com/ranked-countries-with-the-highest-share-of-seniors/