What opportunities does the investment from major internet companies bring to embodied intelligent robots?

Over the past decade or so in the era of mobile Internet, numerous Internet giants have emerged and thrived, gradually becoming an important part of our daily lives and the operation of systems. For society, it can be said that these enterprises have increasingly become like the "capillaries".

In 2025, behind the booming embodied intelligent robot track, the presence of these enterprises is also evident:

The investment department of JD.com, which had been quiet for some time, made consecutive investments in six robot - related companies from May to July this year.

Meituan has continued its investment rhythm in the robot track over the past year or more, participating in different rounds of financing and investing over 500 million yuan.

Looking at Alibaba, which had a fierce food - delivery battle with Meituan this summer, they have also been active, frequently making investments in companies upstream of the robot industry.

Together with companies like ByteDance and Tencent, by the autumn of 2025, it is obvious that behind the leading embodied intelligent robot companies, there are basically more than one Internet company.

Why do these Internet companies with abundant funds and resources favor this track? And what opportunities and uncertainties will their investments bring to the players in this track?

Why Internet companies don't want to miss this investment boom

In the past decade, as the expanding Internet giants' business forms and boundaries gradually approached and clashed, when facing potential businesses and tracks, they either entered the market themselves or made strategic investments.

The former mainly includes software, online virtual businesses, and those with an offline foundation. For example, the popular general large - models at the beginning of the year, and the competition between "logistics - food delivery" capabilities during the hot summer. To sum it up, the logic is that the business is highly relevant, the industrial chain is relatively short, and it can reuse some of the giants' existing capabilities.

For the latter, it usually involves hardware - based businesses. When it comes to large - scale industrial manufacturing, large companies are always more cautious about entering the market directly.

Furthermore, why are the giants making heavy bets on the robot track this year?

Firstly, there is, of course, the consideration of financial returns.

After being highly regarded and hyped, since the second half of this year, several leading robot companies have spread the news of planning to go public.

For example, Unitree Technology, one of the few profitable companies in this track for consecutive years, if it successfully goes public, Meituan, which holds about 8% of its shares (through "Hanhai Information Technology Co., Ltd."), will gain good investment returns.

Who would want to miss the next Unitree? Therefore, before the industry takes off and when the valuations are still "reasonable", finding a way to enter the market has become the goal of the investment departments of large companies.

Secondly, it is about the synergy effect that these robot products can bring to their own businesses in the future.

In July, when JD.com invested in Zhujidongli, it stated that it hoped that the company's technology, products, and services could form sufficient synergy with JD.com's existing supply - chain business system and provide support in areas such as warehousing logistics and retail robots.

Large companies like Alibaba and Tencent, which have native business capabilities in cloud computing and AI, hope to intervene through investment so that their related businesses can have a cooperation platform and suitable usage scenarios, thereby expanding the coverage and influence of their business ecosystems.

Thirdly, they want to secure a position in the service scenarios that robots can penetrate in the future.

For example, in the past few years, drones and automatic delivery vehicles have gradually appeared in small quantities in the business scenarios of logistics and food - delivery companies. As robots develop at an accelerated pace in the future, their penetration rate into different scenarios such as industry, service, and even private life is expected to continue to increase.

Who knows, at some point, Internet giants may hope that the robot companies they have invested in can shine in such scenarios, and their own services and products can also be naturally extended.

In addition, considering the relative short - coming of AI and other technologies in terms of "intuitive perception", and since robots are a "visible" and even "tangible" field, investing in and being associated with the robot industry and track will also somewhat help these companies elevate the narrative of "technology" to the strategic level.

What kind of stage do Internet companies provide for robots

Most domestic robot manufacturers are now emphasizing that they hope their products can follow the path of "first application, multiple applications, then generalization", first let robots enter the scenarios, and solve problems such as "intelligence" during the application process.

So, besides writing checks, what kind of scenarios and stages can Internet giants provide for robots at present?

In a recent interview with robot industry companies by the financial media "LatePost", the interviewees summarized several characteristics of the usage scenarios they hope to obtain in the future:

On the premise of relatively low safety requirements, firstly, they hope the scenarios have certain non - structural characteristics, such as different varieties and layouts of items in different scenarios. Currently, most of the leading intelligent humanoid robots are in the process of basically meeting simple and fully structured scenarios and seeking breakthroughs in adapting to non - structured scenarios.

Secondly, the scenarios should have a certain scale, which is equivalent to providing enough different "test papers" for robots to practice, so as to ensure the collection of enough real - world data.

Thirdly, it depends on whether the marginal cost of replication ability can be gradually reduced through large - models to accelerate the generalization ability of the models.

For example, last year, Meituan invested in Galaxy General in this track and cooperated with it on the first "smart pharmacy solution". Robots can achieve functions such as 24 - hour unattended operation, autonomous replenishment/retrieval of goods, identification of products, and delivery to customers.

Coincidentally, recently, Unitree Technology, a leading enterprise in the industry, launched its new wheeled humanoid robot G1 - D. I noticed that in the expected commercial scenarios introduced on its official website, there are also similar commercial retail scenarios.

It can be seen that embodied intelligent robot companies themselves have a strong desire to cooperate with Internet retail companies in scenarios.

(The expected commercial scenario of the G1 - D launched by Unitree on its official website)

The main logic of this cooperation based on "consumer retail enterprises - robot investment" is that large companies as investors can provide real offline scenarios and data for robots to help the invested enterprises accelerate iteration. And the invested robot companies use technology to help large companies improve efficiency.

In addition, factors such as high costs have also become bottlenecks restricting robots from entering more scenarios for "training".

Some practitioners have mentioned that they hope to see the stage where a certain link in the industrial operation chain is completely replaced by robots as soon as possible, because this means that in that link, embodied intelligent robots have completely surpassed humans in terms of economy.

In the past, other forms of robots or robotic arms have proven the value of industrial automation to varying degrees in the flexible assembly of components in 3C, electronics, and new - energy vehicles, as well as in material handling and sorting.

In the future, if Internet companies can provide rich and suitable scenarios in warehousing, logistics, etc., enabling embodied intelligence to achieve cross - task generalization ability in a single scenario through rapid configuration and learning from examples, and cooperate with components such as algorithm capabilities and sensors in relevant tracks to form a cost - effective and efficiency - enhancing solution, it will greatly help the large - scale commercialization of embodied intelligence.

What other robot opportunities are there in this year's industry hotspots

If we broaden our perspective and do not limit the thinking of the development path of humanoid robots' scenarios only to the shift from "structured industrial scenarios" to "unstructured service scenarios", we may have other discoveries.

In the past, with the continuous improvement of AI capabilities, some functional robots that we were familiar with in some life - service scenarios have also been quietly evolving towards intelligence, but they have received less attention due to factors such as scenario usage or appearance.

Do you still remember the public - relations drama of "Luo Yonghao vs. Xibei" that exploded on social media two months ago?

Riding on the wave of that event, cooking robots also gained some popularity. I think that "robots" could be a solution for diners who want "wok - hei" dishes instead of pre - made ones and for the catering industry aiming to reduce costs and improve efficiency.

In fact, in recent years, the new - generation cooking robots have become much more "intelligent" compared to their "wooden predecessors" that could only stir - fry dishes.

For example, in commercial scenarios, some cooking robots can already achieve the function of "multi - store data monitoring", helping the catering industry optimize supply - chain procurement plans and improve operational efficiency. In the home environment, they can not only automatically identify the types of ingredients and weigh them for cooking but also achieve a closed - loop of "cooking - procurement" by linking with the refrigerator and automatically pushing ingredient - procurement links.

However, these intelligent robots seem "less attractive" in the commercial narrative because they do not have a "humanoid" body.

At the Bund Robot Conference in September, Lingbo Technology, a subsidiary of Ant Group, showed off its R1 multi - modal interactive service robot through the "cooking" demonstration.

As a giant that mainly invests in downstream companies in the humanoid robot track and mostly plays the role of a "supplier", the R1 launched by Alibaba demonstrated its end - to - end planning ability to handle complex tasks through the cooking scenario.

In addition to the kitchen scenario as a demonstration point, this service robot will have more generalization ability expansion in the future, and it has been disclosed that it will conduct more exploratory extensions in scenarios such as healthcare companion and nursing robots.

This information, combined with Ant Group's major move of upgrading its original "Digital Healthcare Department" to the "Health Business Group" in early November, leaves a lot of room for imagination.

With the infrastructure - building and popularization of Internet companies' services in China, it is believed that in the future, these technology - driven companies will provide more scenarios and possibilities for the penetration of robots.

This article is from the WeChat public account "Caiguan Erjie", author: Erjie, published by 36Kr with authorization.