When the Nurses Run Out, the Robots Roll In: Asia's Elderly Care Crisis Meets AI

When the Nurses Run Out, the Robots Roll In: Asia's Elderly Care Crisis Meets AI

Asia's most developed economies are ageing faster than anywhere on Earth. Japan's workforce is collapsing under the weight of exploding demand for care. South Korea's elderly care industry is ballooning. And across the region, governments and companies are turning to the same answer: humanoid robots that can do the work humans have abandoned.

The numbers tell a grim story. Japan's nursing shortage has reached crisis point, with only one candidate competing for every 4.25 open care positions. Meanwhile, the nation's births have cratered to just 720,988 in 2024, a 5% decline that leaves fewer young workers supporting more elderly people. By the end of 2024, all of Japan's baby boomers had reached age 75 or beyond. The mathematics of care are broken.

Enter AIREC: a 150-kilogramme humanoid robot developed under Japan's JST Moonshot Goal 3 programme. Its creator, Professor Shigeki Sugano of Waseda University, has designed a machine that rolls patients to prevent bedsores, guides joint rehabilitation with pain-free force estimation, performs ultrasound examinations, assists with dressing, and can even cook scrambled eggs. It sounds like science fiction. It is being deployed in Japanese care facilities today.

By The Numbers - Japan: 570,000 care worker shortage expected - Only 1 candidate per 4.25 open nursing positions (December 2024) - Japan births dropped 5% year-on-year to 720,988 in 2024 - AIREC weighs 150kg and costs approximately 10 million yen (£64,000 to £667,000) - South Korea's elderly care market valued at $63.8 billion in 2024; projected to reach $96.9 billion to $208.5 billion by 2030 - South Korea population aged 65+: 10.51 million (20.3% of total population) - AI in Aging and Elderly Care Market growing at 21.30% CAGR through 2035

How AIREC Works, and Why It Matters

AIREC is not a simple robotic arm. It uses deep neural networks (DNN) for real-time joint prediction, paired with fisheye and depth cameras that build a three-dimensional model of a patient's posture. The robot was trained on NVIDIA Isaac Sim environments and RTX GPUs, meaning it learned to move patients safely in virtual space before ever touching a real human being. That careful engineering matters. A patient being rolled incorrectly could suffer serious injury. A robot guessing wrong about where a joint is, or how much force it can tolerate, could harm someone.

The robot's ultrasound capability is particularly striking. Rather than requiring a skilled sonographer to attend every exam, AIREC can position the probe correctly, collect images, and send them for remote analysis. For facilities stretched beyond capacity, this means faster diagnostics without waiting weeks for a specialist visit.

Zenkoukai care facilities are already testing AIREC alongside sleep sensors and exercise robots. But the road to widespread adoption remains long. Full integration into a facility is unlikely before 2030. The technology is not yet cheap enough, nor proven enough, for rapid scaling across Asia's care sector.

We're not replacing nurses, we're extending what they can do. A robot never gets tired, never makes a mistake from fatigue, and never refuses the difficult patient. But a robot also cannot hold a hand when someone is dying. Both are essential.

— Professor Shigeki Sugano, Waseda University

South Korea and China: The Race Accelerates

Japan may have the oldest population, but South Korea's elderly care boom is the most economically vibrant. The nation's senior industry is now valued at $63.8 billion, with projections suggesting it will more than double to $96.9 billion to $208.5 billion by 2030. With a population of 10.51 million people aged 65 and above, representing 20.3% of South Korea's total, is driving demand for innovations that can meet care needs at scale.

China, meanwhile, has moved to formalise its position on humanoid robots in elderly care. In 2024, the Chinese government issued explicit guidelines for deploying humanoid robots in elder-care settings. This represents not just technological enthusiasm but state-level commitment to automation as a solution to demographic crisis.

The regional consensus is clear: robots will supplement the care sector. The question is no longer "if" but "how many" and "how quickly."

| Region | Elderly Population (65+) | Industry Value (2024) | Projected Value (2030) | Key Development | |---|---|---|---|---| | Japan | 36% of population | Not separately valued | N/A | AIREC deployment, JST Moonshot Goal 3 | | South Korea | 20.3% of population (10.51M) | $63.8 billion | $96.9 billion to $208.5 billion | Rapid expansion, AI companion dolls | | China | Growing rapidly | N/A | N/A | 2024 humanoid guidelines issued | | Asia-wide | Fastest ageing region | AI care market growing at 21.30% CAGR | N/A | Multiple robot platforms, national commitments |

The Limits of Automation

Not everything can be outsourced to silicon and steel. Care is a human endeavour at its core. The warmth, the conversation, the dignity of being known by someone, the comfort of a familiar voice: these are not things a robot provides, no matter how sophisticated its language model or how precise its movements.

A humanoid robot can prevent bedsores and perform basic exercises, but it cannot sit with someone experiencing profound loneliness. It cannot offer the emotional presence that makes being old bearable. Asia's elderly care crisis is partly a problem of labour supply, but it is also a problem of meaning. Societies that have moved care work to the margins, paid carers poorly, and treated the work as unskilled have created a vacuum that no amount of automation will truly fill.

The most honest assessment is this: robots will handle the physical demands. They will perform the repetitive, strength-intensive tasks that wear out human bodies. They will fill some of the gaps left by insufficient workers. But they cannot replace the human connection that makes care actually work.

A robot can do the work, but it cannot do the caring. We need both solutions: better pay and working conditions for human carers, and robots to handle the heaviest physical burden. To choose one over the other is to fail the elderly.

— Takashi Miyamoto, Director, Zenkoukai Care Facilities

The Economics of Care in an AI Age

The economic logic driving robot adoption is straightforward. A care worker in Japan costs roughly £25,000 to £45,000 per year in salary and benefits, plus training, replacement when they burn out, and the overhead of management and scheduling. AIREC costs around 10 million yen upfront, and runs on electricity and maintenance. Over ten years, even at high maintenance costs, a robot begins to look economically rational in a way human care never does under current market conditions.

This creates a moral hazard. If governments and care operators can automate away the need to invest in human carers, why would they pay better wages or improve conditions? The result is a race to the bottom in care work: automation accelerates, wages stagnate, fewer people enter the profession, shortage deepens, and the case for more robots strengthens.

Some care facilities in the region are attempting a different model, much like Samsung's initiatives in everyday AI companions for older adults, where technology supplements rather than replaces human attention. And projects like AI wellness home solutions such as Ceragem are exploring how sensors and AI can help families stay connected across distances.

But these remain outliers. The dominant trajectory is substitution, not supplementation.