Table of Contents
Contents
In the spring of 2025, a wave of robotics investors began making frequent visits to elderly care facilities.
This surge of enthusiasm was driven by a convergence of data points: China’s silver economy is projected to reach 9 trillion yuan in 2025 and exceed 25 trillion yuan by 2030. Meanwhile, the penetration of large language models has, for the first time, created a viable technical pathway for companionship — a need that was previously considered impossible to address. In 2025, the elderly care robotics sector attracted over 2 billion yuan in investment for the year, with the companion robot subcategory growing at an annual rate of 120%, making it the fastest-growing segment.
Yet the skepticism has not faded. “Companionship is a false need” remains a common refrain in investment circles, and whether Chinese consumers are willing to pay for emotional companionship is still an open question. The ceiling on the one-time hardware sales model is obvious, and whether subscription-based or platform-based models can prove themselves before the policy window opens around 2028 remains the sector’s biggest commercial uncertainty.
The current players are taking divergent paths. Startups are betting on vertical scenarios like companionship, tech giants are building platforms and infrastructure, and traditional care institutions are treating smart devices as efficiency tools. The three groups form a symbiotic relationship across the industry chain, even as competition intensifies in parallel.
The business of elderly care is waiting for a moment of certainty large enough to justify the bet — while racing to see who can survive long enough to reach it.
Investors Descend on Elderly Care Facilities
This spring, every robotics investor seems to be making the rounds at elderly care homes.
In 2025, China’s silver economy reached a market size of 9 trillion yuan, with projections to exceed 25 trillion yuan by 2030. According to statistics, the smart elderly care sector completed 112 financing events in 2024, totaling approximately 11.6 billion yuan. But compared to a market measured in the trillions, that figure remains modest. The elderly care business has long occupied an awkward position: a vast market that few dare to enter or invest in.
The turning point came in 2025. The large language model revolution ignited by ChatGPT finally penetrated the walls of elderly care facilities, with AI making previously impossible companionship a reality. That year, the elderly care robotics sector attracted over 2 billion yuan in investment, with a compound annual growth rate of 48%. Companion robots, while accounting for only 6% of the overall market, grew at an annual rate of 120%, making them the fastest-growing subcategory.
As the Year of the Horse began, the first wave of entrepreneurs building AI products for the elderly were already booking back-to-back meetings with investors.
Zhang Kai, founder of Xinyi Tech, was among them. He had been busy demonstrating his companion robot to investors: a 25-centimeter-diameter base, 50 centimeters tall, weighing 3.5 kilograms, with a cute pet-like appearance that invites hugging. It navigates rooms freely, charges itself, recognizes family members, initiates interactions, and — most importantly — truly understands the people in the household. Another team fielding investor visits hails from Shanghai Jiao Tong University. Founder Wu Yuanpei’s product, the XinShou Bracelet, is a counterintuitive AI device. Rather than pursuing the sleek tech aesthetic of an Oura Ring, the team chose to embed a large language model inside a silver bracelet, deliberately erasing any trace of technology to win over its target customers. “The best technology is technology that goes unnoticed,” Wu Yuanpei said. “Elderly users shouldn’t have to learn how to use a device; the device should blend naturally into their existing habits. Bracelets have been a staple of Chinese culture for decades — we’re not creating a new behavior, we’re embedding intelligence into one that’s been around for generations.”
Two radically different product forms, both targeting the same market of 310 million people. But capital’s skepticism toward AI plus silver economy has not fully dissipated.
The loudest doubts come in two forms. First: is companionship really a genuine need? “Some investors told me flat out that companionship is a false need, that people’s economic circumstances haven’t reached the level where they need companionship,” Zhang Kai said, recounting a challenge he has faced more than once during funding pitches. One investor put it even more bluntly: “People simply aren’t wealthy enough to spend several thousand yuan on companionship.”
The skepticism extends beyond investors to potential users as well. When asked whether he would buy a companion product for his grandmother, one young man born after 2000 replied: “I don’t think the elderly are as fragile as we imagine.”
Facing such doubt, Zhang Kai offered a rebuttal: “There’s a Chinese saying — ‘affection flows downward’ — meaning people naturally pour more care into younger generations than older ones. In a typical family today, you’ll see four grandparents and two parents revolving around a single child, sometimes showering the child with more attention than necessary. But flip the picture: two middle-aged adults are responsible for at least four elderly parents, sometimes even their own grandparents. When it comes to companionship for the elderly, there is a genuine gap — it’s an acute, structural pain point.” Spending a few thousand yuan on a companion robot, he argued, is not a luxury purchase. It is buying peace of mind.
The data backs up the pain point. According to the Ministry of Civil Affairs, by the end of 2025, more than 118 million elderly people in China were living alone or in empty-nest households — 38% of the total elderly population.
Wu Yuanpei’s response to the skepticism was direct: “We need to distinguish between entertainment-oriented companionship and survival-critical companionship. In an aging society, the problems of elderly people living alone, chronic disease management, and sudden health emergencies are not psychological needs — they are structural realities. AI companionship is no longer a luxury. It is infrastructure.”
In Wu Yuanpei’s view, the elderly and the young have fundamentally different needs when it comes to AI companionship. Young people’s emotional needs tend toward expression, resonance, and social connection. For the elderly, the core need is closer to a sense of security, of being noticed, of dignity. Many elderly people don’t lack someone to chat with — what they lack is a sense of certainty: that someone will know if something goes wrong, that someone is paying attention to how they’re doing day to day, and that their daily life retains a sense of dignity and decency.
The second line of skepticism concerns the silver economy’s reputation as a large but hard-to-monetize market, where building brand loyalty among elderly users is notoriously difficult and marketing remains an uphill battle.
Zhang Kai was unsurprised by the “silver economy doesn’t make money” critique, but he sees a generational shift creating opportunity. “The previous generation of elderly — those born around 1955 who lived through the great famine — are genuinely frugal. But the new generation, those born after 1965, are completely different. They’re active seniors with transformed spending habits, willing to invest in their own quality of life. Then there’s another cohort: the children in their forties and fifties who buy products for their parents. They have both the purchasing power and the filial motivation. In 15 to 20 years, this group will themselves become the elderly. So there’s no need for debate — acceptance of AI robot products will only grow.”
“Looking at Japan’s silver economy trajectory as a reference, we believe 2028 to 2030 will definitely see an inflection point,” Zhang Kai said with evident patience. “What we need to do is refine our product, push it to market fast, sell enough volume so that our AI capabilities can keep iterating. If the company stays steady and survives long enough, the wave will come naturally by 2028 or 2030.”
Big Tech Enters the Game with a Different Playbook
While startups hustle for investor meetings outside elderly care facilities, China’s tech giants have been quietly positioning themselves in the silver economy. But their approach is fundamentally different from the startups’.
Big tech builds platforms and infrastructure, earning service fees and technology licensing revenue. They don’t care how much a single piece of hardware sells for — what matters is how many devices connect to their platform. Startups build products for vertical scenarios, earning margins on hardware and premiums in niche markets. Their goal is to achieve excellence in a specific use case. Traditional care institutions operate ground-level service networks; for them, smart technology is a tool to boost efficiency and reduce costs — a caregiver who once looked after five elderly residents can now manage eight with the help of smart devices.
The three are not competitors but collaborators along the industry chain. This ecosystem is taking shape, and for startups, it carries dual implications.
The good news: they no longer have to fight alone. There are platforms to integrate with, channels to partner through, and technology stacks to leverage. The bad news: competition is also intensifying. A startup that merely produces a simple hardware product risks being replaced by a big tech platform feature, or having its prices squeezed by institutional bulk procurement.
The market itself is also stratifying. From an investment perspective, the structure of the elderly care robotics market is shifting. Rehabilitation and nursing products command 94% of market share, but their growth rate has plateaued. Companion products account for just 6%, yet their 120% annual growth rate makes them the fastest-expanding segment.
This contrast reveals that the established market has entered a zero-sum game, with players competing for B2B clients — elderly care facilities and rehabilitation centers — in an already fierce price war. The emerging market, by contrast, is experiencing volume growth. The base is small, but the ceiling is high. Whoever captures the consumer market stands the best chance of breaking out.
Investment logic is also shifting. A fund partner who has backed multiple elderly care tech projects described the change: “From 2020 to 2023, investors focused on technical moats and hardware capabilities. They’d ask what makes your robot more advanced, how precise your sensors are, what your algorithm recognition rate looks like. But starting in 2025, investors care more about user retention and real-world scenarios. They ask whether elderly users actually use the product, whether they renew subscriptions, whether adult children are willing to keep paying. The entire industry’s investment logic has pivoted from technology-driven to demand-driven.”
User profiles are also becoming more granular, with three distinct customer segments emerging.
The first is active seniors born after 1965: they have pensions, purchasing power, and familiarity with digital products, and they’re willing to pay for quality of life. The second is middle-class adult children aged 35 to 50: they feel both filial duty and anxiety, with parents living alone in their hometowns while they work in distant cities. Buying a smart device is both an act of filial piety and a way to buy peace of mind. The third is families caring for disabled elderly — the true hardship market. These families need life-saving functionality: monitoring, emergency calls, and nursing assistance. But their ability to pay is often limited, requiring support through long-term care insurance, government subsidies, and other funding mechanisms.
The silver economy plus AI sector is transitioning from pure functionality to a blend of functionality and emotional connection.
This evolution mirrors the trajectory of the smartphone. In the early days, phones satisfied the basic need for voice calls — the Nokia era sold durability and signal strength. Then the iPhone arrived, adding social media, entertainment, and gaming. The phone ceased to be merely a communication tool and became a lifestyle. Elderly care technology is now undergoing a similar transition: rehabilitation and nursing represent the voice-call era, with a relatively mature market. Companionship and emotional connection represent the social-and-entertainment era — still in its infancy, but growing faster, signaling where demand is heading.
The Policy Window Opens as New Models Take Shape
Changes in the external environment are injecting new certainty into the industry.
In 2025, long-term care insurance pilot programs expanded from 49 cities to 70, covering a population of 180 million. What does this mean? In the past, when an elderly person suffered a fall or became disabled, the family bore the full cost of hiring caregivers and purchasing equipment — expenses of thousands to tens of thousands of yuan per month that many families simply could not afford. Now, in cities covered by long-term care insurance, certain smart devices are beginning to qualify for reimbursement: smart mattresses, emergency call devices, and anti-wandering wristbands. Eligible elderly individuals can have a portion of their costs covered. This doesn’t just reduce the financial burden on families — more importantly, it lowers the barrier to market education. Previously, startups had to convince users on their own why a wristband was worth several thousand yuan. Now, the government has effectively endorsed the product. With half the cost reimbursed, the user’s decision threshold drops immediately.
But policy is only an external condition. Whether a company survives ultimately depends on whether its business model works.
The traditional one-time hardware sales model has too low a ceiling. Selling a single device for a few thousand yuan in profit, with no recurring revenue, means that when a user loses interest after three months and shelves the product, the company has no incentive to keep updating or iterating. This model works in consumer electronics because of upgrade cycles — people replace their phones every two years, their earbuds every year. But elderly care devices are different. Seniors don’t swap devices frequently; a single device might be used for years. If a company relies solely on hardware sales, it will quickly exhaust its growth potential.
New models are being explored. One direction is the subscription model: monthly fees of 99 to 299 yuan that include both hardware access and a service package. The hardware might be provided free of charge or at a nominal fee, with the real revenue coming from the monthly service. What’s in the service package? Round-the-clock emergency response — when an elderly user presses the SOS button, the call center must respond within three minutes, contacting family members or dispatching someone to the home if necessary. Regular health reports — a weekly summary of the user’s activity data, sleep patterns, and heart rate trends, pushed to adult children. AI companionship conversations — not just a robot talking, but a psychology team in the background analyzing the elderly user’s emotional state and flagging anomalies for early intervention. The advantage of this model is that continuous user payments provide the company with steady cash flow and the incentive to keep improving its services.
Platform models are also emerging. Rather than building hardware, these companies build connections — linking elderly users, their adult children, service providers, and care institutions, and taking a commission on each transaction. For example, an elderly user orders a bathing assistance service through the platform, which dispatches a nearby care worker. After the service is completed, the platform takes a 15% to 20% cut. Hardware is the entry point, services are the core, and data is the asset. This model offers enormous potential, but it demands exceptional operational capability — sufficient service providers on the supply side and enough users on the demand side. Both sides must reach scale for the model to work.
Government-funded models are also innovating. These go beyond simple procurement to combine government subsidies with individual co-payment. For instance, one city launched a plan in which the government subsidizes 60% of costs and individuals pay 40%, lowering the user threshold while ensuring reasonable profit margins for companies. There is also a “service voucher” model: the government issues vouchers to eligible elderly individuals, who can use them to purchase smart devices or services. Companies then redeem the vouchers with the government for payment. The beauty of this approach is that government spending is targeted precisely, companies gain a stable revenue source, and users receive tangible benefits.
