“Honestly, robots simply aren’t selling right now. The only way to move them is to rent them out,” said Cheng Zhijun, head of sales at a leading robotics company.

For now, robots remain at a very early stage. Most are still hand-built prototypes and are difficult to deploy in real-world settings.

The industry’s only apparent answer is “leasing plus insurance.”

Since 2025, PICC, Ping An and CPIC, three of China’s top insurers, have all moved into the market.

Qingtianzu, the first robot leasing platform, has even taken out a 200 million yuan policy for its 2,000 robots.

Is this really the only path for robots to reach the market?

The Deployment Problem

Cheng Zhijun is convinced robots cannot follow the simple buyout model of refrigerators or color TVs. “I’ve visited dozens of factories and met more than 100 supermarket and mall owners. They simply don’t dare buy robots outright.”

Cheng said the reason is straightforward: robots are too new. No one knows whether they work well, or what happens when they break.

Robots are fundamentally different from phones and home appliances. Between users and manufacturers lies a huge “trust gap.”

If a phone breaks, you lose connectivity. If an appliance breaks, you lose convenience. But if a robot “breaks,” it could become a disaster.

First, after-sales service for robots is a major challenge.

Phones and home appliances are static or semi-static devices with extremely stable structures. Robots move.

Under the principle of “antifragility,” the more precise and advanced an instrument is, the more fragile it tends to be, with a shorter service life and a higher failure rate.

Wang Zhongyuan, president of the Beijing Academy of Artificial Intelligence, once said that of 10 robots they purchased from a certain brand, five broke within one or two months.

A group of robots made by Songyan Dynamics appeared on the Spring Festival Gala. The smallest of them, called “Little Bumi,” sells for around 10,000 yuan.

After the gala, Little Bumi reportedly triggered a buying rush.

“Reservations have reached the thousands,” a Songyan Dynamics salesperson said.

But the rush was quickly questioned by industry veterans.

“For robots priced below 10,000 yuan, many components can only be low-end versions. If such robots enter homes and interact with children and pets, how long can they really last?” Many industry insiders expect this buying rush to be followed by a string of after-sales problems.

If any robot is sold outright, the manufacturer must build a nationwide repair force with extremely fast response times. For startups, the logistics and after-sales costs would be hard to estimate.

Second, robots iterate extremely quickly. For traditional appliances or hardware, delivery is the finish line. Once a refrigerator is bought, its functions are fixed.

For the robotics industry, delivery is the starting point.

Unitree founder Wang Xingxing said that in just a few months in 2025, the company’s algorithms had already gone through several iterations.

Users who pay heavily for an outright purchase may end up with hardware that is “obsolete three months after launch.”

That pace of technical depreciation can create deep frustration and brand resistance among the first wave of individual buyers.

Third, liability for robots remains unresolved.

If a robot breaks an antique in your home or injures a pet, is the fault with the sensor hardware or the vision large model software?

Current law has not yet assigned liability for “AI behavior.”

Under a buyout model, users and manufacturers could fall into endless disputes.

At that point, a third party is needed to cut through disputes and uncertainty and take responsibility.

Cheng said candidly that because of these objective problems, robot sales have been moving extremely slowly. What can be done?

There has to be a way to sell robots. Otherwise, valuations cannot be sustained, and this bubble will burst.

Insurance Plus Leasing

The industry’s ultimate answer is leasing plus insurance.

In 2025, Beijing, Shenzhen and Ningbo were among the first to roll out robot insurance subsidies.

After this round of policy guidance, insurers responded with notable enthusiasm.

In 2025, a group of established insurers rushed into robot insurance, each trying different approaches while feeling their way forward.

The insurers now in the market include PICC, Ping An, CPIC and Dajia Property & Casualty Insurance. Publicly disclosed cumulative coverage is about 223 million yuan.

Once insurance, a powerful financial tool, entered the scene, it did solve the core problem. The biggest pain point in robotics today is that customers are afraid to use the machines.

Robot repairs are too expensive, often costing 30,000 to 300,000 yuan, trapping companies in a cycle of “not daring to use them, fearing damage and being unable to pay.”

The solution becomes clear once the numbers are laid out.

Take PICC Property and Casualty’s Wuhan plan as an example: a 5,000 yuan premium provides up to 500,000 yuan in coverage. With that financial backstop, companies do not have to worry that a single machine “falling” will wipe out their economics.

For now, this insurance backstop is spreading across the industry, from factories in Jiangsu to enterprise customers such as Soulmate Technology and the Shanghai Chuansha Nursing Home, which signed the “first elderly-care robot liability insurance policy.”

Insurance has become a prerequisite for bulk robot procurement.

As the head of the nursing home put it, “Only with a backstop do we dare run a pilot.” But insurance alone cannot solve the problem.

The robotics industry is severely short of data, making it difficult to price risk for insurance products.

And because robots iterate so quickly, “residual value management” is also difficult.

In the past, the residual value of items such as phones and cars declined in a linear way and was easy to calculate. An iPhone, for example, might lose 20% of its residual value each year.

Robot residual values, by contrast, can fall off a cliff.

There is a famous saying in actuarial circles: “Unpredictable risk is gambling; only quantifiable risk is insurance.” Cheng believes that in robotics, an extremely fragile and fast-iterating industry, insurance alone is like building a house on a swamp. The foundation remains unstable.

Leasing, another powerful tool, must be added. First, after-sales service.

A robot breaks?

The leasing company can send a new unit from the nearest warehouse, while the broken one is shipped back to a central facility for batch repair.

Leasing turns fragmented claims into scaled asset maintenance. Only then will insurers dare take the business; otherwise, premiums would be so expensive that no one could afford them.

Second, the data problem.

The biggest problem with the current insurance model is the lack of data.

But robotics companies all know that data is their core asset, and they are reluctant to share it.

Without data, insurers can barely move.

In a leasing scenario, however, robot operating data, such as when machines break and which parts are fragile, flows centrally to leasing companies and insurers.

That allows data to accumulate gradually.

Third, residual value.

The residual value problem can also be partly mitigated through leasing.

When leasing companies set rental prices, they need to factor in residual value forecasts.

For example, if a robot may depreciate within six months, early rental fees can be set very high to extract more value before depreciation hits.

Qingtianzu, the first robot leasing platform, has even taken out a 200 million yuan policy for its 2,000 robots.

The Only Right Answer?

The “leasing plus insurance” model turns robots’ technical risk into a financial cost.

That is the real power of finance.

Many people in the industry therefore predict that future robots will most likely be rented rather than bought outright.

And “leasing plus insurance” appears to be the only clear path for robots to cross the “valley of death.”

But several leading players do not see it that way. “Why should I hand my most critical data to an insurance company?” He Jia, founder of a robotics company, told Guiji Jianglin.

Isn’t insurance just algorithms?

“My algorithms patrol a thousand times per second. I understand risk better than the insurance companies do,” He Jia said. In the robots he recently sold, he embedded a “lifetime maintenance subscription package.”

In other words, it is a form of alternative insurance developed by his own company.

It removes the insurance company’s commission and keeps risk-pricing power directly in the company’s hands.

“I don’t need insurers to take the key position. I’m the one dealing the cards,” He Jia said. His goal is to eliminate the insurance companies.

Several insurers have also noticed the ambitions of leading robotics companies, but they dismiss them.

Cars have been around for so many years and still rely on auto insurance. Eliminate insurance companies? Fantasy. But He Jia pointed to the key difference. Traditional auto insurance exists because cars are “dumb devices.”

Where car owners drive, how they drive and what road conditions they face cannot be fully monitored in real time by insurers or manufacturers.

A robot, however, is a data center that is online 24 hours a day. Manufacturers can have a “God’s-eye view.”

They know when a machine is likely to break, and whether an accident was caused by faulty code or hardware fatigue.

Insurers want in? Manufacturers can charge high “data retrieval fees” or simply refuse to provide data.

Without data, insurers are “blind” in front of robots and can only be led by manufacturers.

“To be honest, at this stage we do need to rely on leasing plus insurance. But over the long run, I think each robotics company will spin out its own insurance platform to take on this part of the business,” He Jia said. The mantis stalks the cicada, unaware of the oriole behind it.

Finance has become an accelerator for robot deployment, and even a bottleneck technology in its own right.

But in turn, robotics, as a disruptive industry, also wants to remake finance.

When an industry can predict risk in real time, create value in real time and recover assets in real time, shouldn’t we say that it is itself the most efficient financial system?