Big tech companies are "crazy about" investing in robots, and there is a consensus behind it.
In 2025, the embodied intelligence track, which had the greatest differences before, has recently become extremely popular. Recently, the financing of embodied intelligence has been accelerating continuously. On July 8th, Meituan led the investment in Tashizhihang, which secured a financing of $122 million.
Looking further, from July 7th to 9th alone, four companies including Xinghaitu, Xingdong Jiyuan, Yunshenchu, and Tashizhihang raised a total of over 2.5 billion yuan in financing. If we extend the time frame to June, companies such as Yinhe Tongyong (which received a financing of 1.1 billion yuan led by CATL on June 23rd) and Yushu Technology (which completed its Series C financing on June 19th with a valuation of over 10 billion yuan) also performed remarkably well. According to data from IT Juzi, in the first five months of 2025, there were 114 investment events in the field of embodied intelligence, with a financing amount of 23.2 billion yuan, both exceeding the totals for the whole of 2024.
Photo/Yinhe Tongyong
From the perspective of investors, large enterprises and industrial capital have become the main forces in this round. Giants such as Meituan, CATL, Tencent, and Alibaba have become the core promoters. Meituan has intensively invested in Tashizhihang and Xinghaitu in the past month and participated in projects such as Yushu Technology and Yinhe Tongyong. As of now in 2025, Meituan has made six investments in the field of embodied intelligence. Its investment logic revolves around "computable scheduling in the physical world" and covers the entire chain including robot bodies, large AI models, and scenario applications.
By leading the 1.1 billion yuan financing of Yinhe Tongyong, CATL combines embodied intelligence technology with new energy scenarios to explore the implementation of humanoid robots in fields such as smart pharmacies and logistics warehousing. Tencent, Alibaba, and Ant Group are deploying in the hardware of humanoid robots and core components such as dexterous hands by investing in companies such as Yushu Technology and Taihu Robotics.
Actually, from a technical perspective, the current enthusiasm for financing around robots doesn't seem to show significant progress compared to the heyday of Yushu. So, after the disagreements, why are large enterprises still so eager to place their bets? And what exactly is the charm of this "ticket" that seems difficult to cash in the short term?
01 Embodied Intelligence Has Gained Certainty
Why has capital suddenly shown greater interest in embodied intelligence than before? Perhaps we can get a glimpse from the example of Tashizhihang.
According to a report by LatePost, in 2024, when Tashizhihang started financing shortly after its establishment, an insider revealed that the company's direction was not fully determined yet, and the current team members were mainly doctoral students from Tsinghua University and Fudan University. By March 2025, before Tashizhihang's products were officially launched, it completed a financing of $120 million, setting a record for the largest angel-round financing in China's embodied intelligence industry.
Starting to raise funds without a clear direction and obtaining huge financing before the product was launched directly indicate that capital attaches great importance to the background of the technical team and its long-term potential. For example, the founding team of Tashizhihang consists of Chen Yilun, the former CTO of Huawei's autonomous driving, and Li Zhenyu, the former head of Baidu's Intelligent Driving Business Group. This "star team + technological barrier" model has become the core factor attracting investment.
From another perspective, large enterprises and investors may value the certainty that has emerged even more. Gao Yang, the chief scientist of Qianxun Intelligence and an assistant professor at the Tsinghua Institute for Interdisciplinary Information Sciences, revealed in an interview that regarding embodied intelligence, there is now a certain consensus in academia because some paths have been proven false and impossible to achieve. The convergent part is to do better in engineering. Although it doesn't mean that the academic topics have completely shrunk to one, they may have shrunk from 500 to 100, which means that each topic may have been explored by someone.
The decrease in problems and the increase in consensus directly lead to a clearer development path for embodied intelligence. "Since 2023, I've had this feeling: Embodied intelligence hasn't fully converged, but it's in the process of converging. From a social and engineering perspective, convergence is good because we can really start to enjoy the benefits brought by technology." Gao Yang is optimistic about the prospects after the technological convergence.
The defined path in the industry is as follows: Level 0 is industrial robots with little intelligence; Level 1 has intelligence in a single task; Level 2 can complete a few tasks in an office scenario, such as pouring a cup of coffee. Level 3 can perform 70% - 80% of human tasks in a physical scenario; Level 4 can do all the things a human can do in a single scenario, just like Waymo, which can go anywhere within San Francisco. Level 5 is not limited to a single scenario and can do anything in any scenario. Currently, the industry has reached a level close to Level 2.
Photo/Yinhe Tongyong Galbot G1 working autonomously in a smart pharmacy
According to Gao Yang's estimation, achieving Level 2 and reaching Level 3 can be seen within five years.
More interestingly, in terms of methodology, the emergence of VLA has gradually become the path for embodied intelligence to make a leap. This has not only been recognized by entrepreneurial scientists in the robot track but also to a certain extent in the automotive field. Li Xiang, the founder of Li Auto, mentioned in an interview that Li Auto's VLA model is a vivid practice of this trend. By integrating visual, language, and motion intelligence, it transforms the car into an intelligent agent capable of autonomous driving and natural interaction with users, perfectly interpreting the core concept of "embodied intelligence".
Liu Fang, the former head of Xiaomi's intelligent driving, also mentioned that if VLA works, autonomous driving will become a sub - problem of embodied intelligence. It's not difficult to understand why a large number of talents in the field of autonomous driving have joined the embodied intelligence track. Chen Yilun, the founder and CEO of Tashizhihang, was the former CTO of Huawei's autonomous driving; Li Zhenyu, the company's chairman, was the former senior vice - president of Baidu Group and the head of Baidu's Intelligent Driving Business Group (IDG); Ding Wenchao, the chief scientist, led the R & D of Huawei's ADS end - to - end decision - making network.
From another dimension, when the certain technology and certain practice routes are gradually met, talent has become a must - compete area. Gao Yang of Qianxun Intelligence, Wu Yi, the founder of Bianse Intelligence, Chen Jianyu of Xingdong Jiyuan, and Xu Huazhe, the chief scientist of Xinghaitu, are known as the "Four Returnees from Berkeley", and each of them has received financing.
In other words, large enterprises' scramble for embodied intelligence undoubtedly lies in the certainty it offers: a definite timeline, a definite technological path, and definite talent. On the other hand, embodied intelligence seems to be able to bring different levels of improvement to large enterprises from Level 1 to Level 5. In other words, the current investment can be a source of empowerment at any time.
02 Large Enterprises Catch Up and Overtake through Investment
From the current perspective, industrial capital and large enterprises are the most optimistic about the embodied intelligence track. Among Internet giants, Meituan and ByteDance are undoubtedly typical representatives.
"AI will greatly promote the development of robot technology. Meituan will pay attention to and invest in start - up companies in this field. We hope to deepen business cooperation and technological exchanges with these young companies and combine robot technology and AI more closely." Wang Xing said at Meituan's Q4 2024 earnings conference call. He emphasized that Meituan's strategic positioning in the AI field is offensive rather than defensive. He stressed that Meituan will continue to play the role of a technological "bridge" connecting the offline physical world and the online world. It can be seen that Meituan, whose main business is retail services, is more concerned about the implementation of AI in offline scenarios.
In fact, Meituan's investment in embodied intelligence is also accelerating. On July 8th, Meituan's strategic investment department led a new round of $122 million financing for Tashizhihang, an embodied intelligence start - up. On July 9th, Meituan made another move, leading the Series A4 and A5 strategic financings of Xinghaitu, with a total financing amount of over $100 million for the two rounds.
Photo/Tianyancha
In the past year, Meituan has also invested in Yushu Technology, Yinhe Tongyong, and Zibianliang Robotics and has become the largest external shareholder of Yushu Technology and Yinhe Tongyong. According to statistics, Meituan has invested in 30 robot - related enterprises in the past three years.
Meituan's interest in embodied intelligence is not only due to its optimism about the future but also for its application in current business. A typical example is the integration of Yinhe Tongyong, invested by Meituan, with Meituan Pharmacy, which improves efficiency and enables the business to adapt to more scenarios.
Looking further, according to the view of Titanium Media, compared with similar Internet companies such as Alibaba, Tencent, and Baidu, Meituan's progress in AI and large models is not obvious. So far, Meituan has neither launched online AI applications like Wenxin Yiyan and Tongyi Qianwen nor a general large model like GPT. Embodied intelligence may be another opportunity for it to overtake on a curve.
This may explain from another angle why, after Meituan's large - scale investment, JD.com and ByteDance have quickly followed suit.
According to a report by Xinzhike, the Seed team under ByteDance is massively recruiting "No. 1 positions" in robot - related directions, including robot product managers, robot engineering technology managers, and heads of embodied intelligence large models. According to a report by LatePost, ByteDance's robots have been mass - produced with a cumulative output of over a thousand units. Currently, the robots mass - produced by ByteDance are wheeled logistics robots, also known as Autonomous Mobile Robots (AMR). Without mechanical arms for sorting, their main purpose is to transport packages and parts in warehouses and production lines. They can learn autonomously, plan routes, and move to destinations.
JD.com, Alibaba, etc. are also quite optimistic about this track.
Of course, different from the logic of Internet giants eager to overtake, manufacturing giants' increased investment in embodied intelligence is more like making up for a deficiency.
Last year, Geely's Jikr 5G smart factory introduced the industrial - version humanoid robot Walker SLite from Ubtech. After 21 days of training, the robot mainly performs tasks such as material box handling. At the beginning of this year, Geely's Lynk & Co. Chengdu factory also started to introduce Ubtech's humanoid robot Walker S1. Ubtech is the world's first listed company for bipedal robots and has entered the production lines of many automobile companies such as BYD. The company optimistically expects to deliver thousands of humanoid robots in 2026. To some extent, Geely is more like a follower.
Photo/Ubtech Robotics
The relevant person in charge of Yinhe Tongyong, which received investment from CATL, said: "The leading investor, CATL, will continue to deepen industrial collaboration with the company and provide key support for the technological implementation and large - scale application of the company's embodied intelligence large model in the industrial field."
Whether it's overtaking or making up for deficiencies, both Internet giants and manufacturing giants seem to see something useful in embodied intelligence. From a broader perspective, perhaps both sides are looking forward to a more imaginative blueprint.
03 The Future Blueprint Is Comparable to Tesla
At the end of June, a Tesla Model Y started its autonomous driving journey from a factory in Texas, USA, passing through highways, urban areas, and other scenarios on its way to complete the delivery to the customer. Tesla's FSD has been installed in 1.7 million vehicles in North America with zero accidents in two months, which has become a sensational event. Elon Musk has gradually led a car - manufacturing company into multiple fields such as robotaxis and robots. Undoubtedly, the same underlying logic also attracts Chinese large enterprises. Whether they are manufacturing enterprises like CATL and Geely or Internet enterprises like JD.com and Meituan, they all seem to be looking forward to a new growth model and imagination.
This new growth model essentially involves the reconstruction of the ecosystem. When embodied intelligence crosses Level 2 and moves towards Level 3, robots will no longer be tools for a single scenario but "participants" that can understand the physical world and coordinate digital systems like humans. Behind this lies a more ambitious goal than Tesla's single - field breakthrough: not to transform a single industry but to connect every node in the physical and digital worlds, forming a "super - intelligent network" covering production, circulation, and services.
For Meituan, the starting point of this network is the instant retail scenario, but the end is far from that. When embodied intelligence robots can autonomously plan routes through crowded vegetable markets, adjust the delivery temperature according to user needs, and even complete complex tasks such as "picking up items at the door + after - sales communication" on behalf of riders, Meituan's service boundary will expand from "connecting people and goods" to "connecting people and all - scenario needs". This transformation from a "transaction platform" to an "intelligent service hub" is enough for Meituan to open up a much larger new market in addition to its existing business.
