Alibaba competes for global robotics giants

Author | Chai Xuchen

Editor | Zhang Xiaoling

At the intersection of technology and capital, a new storm is brewing. This time, the center of the storm is no longer the codes and algorithms of the virtual world, but the bridge connecting AI and the physical world—embodied intelligence. In this trend, Alibaba, the leader in the AI field, has also entered the fray.

Recently, Lin Junyang, the technical head of Alibaba's Tongyi Qianwen, announced that he has formed a team focused on embodied intelligence within Qwen. This marks Alibaba's official entry into physical AI. In fact, just a month ago, Alibaba Cloud led a $140 million financing round for the domestic robotics startup Zibian Robotics.

From investment to direct involvement, Alibaba's ambitions are evident; it is fully committed to finding a "body" for AI that can perceive and transform the physical world. Behind this is a profound era background: we are in a peculiar "discontinuity" era of AI.

On one hand, the "brain" of AI—large language models represented by Tongyi Qianwen and GPT-4—are evolving at an unprecedented speed, demonstrating astonishing cognitive and generative capabilities. On the other hand, the "body" of AI appears clumsy. In the physical world, apart from industrial robotic arms in specific scenarios, there are hardly any intelligent physical entities that can autonomously understand their environment and collaborate with humans.

Such a discontinuity also contains enormous opportunities. Last year, Eddie Wu bluntly stated that the greatest imagination of AI is to connect with the real world. At the annual shareholder meeting in June this year, Jensen Huang emphasized that AI and robotics technology are the two biggest growth opportunities, representing trillions of dollars in growth potential.

Currently, giants like OpenAI, Google, SoftBank, and NVIDIA are all focusing on the embodied intelligence track. A war to "find a body" for AI has quietly begun, and for Alibaba, which is at a turning point of transformation and value reassessment, this will also be a crucial step in realizing its grand vision of ASI and becoming "great again."

From Cloud to Reality

On October 8, Lin Junyang released a significant message: the team has officially established the "Robotics and Embodied Intelligence Group."

Lin Junyang pointed out on Twitter that multimodal foundational models are transforming into foundational Agents, which can utilize tools and memory to perform long-term reasoning through reinforcement learning, "they absolutely should move from the virtual world to the physical world."

This means that Alibaba is determined to let Tongyi Qianwen transition from the world of language into the physical world. Just recently at the Yunqi Conference, Alibaba Group CEO Eddie Wu asserted that the industry is entering the "autonomous action" phase of AI. Eddie Wu stated that in the future, there may be more intelligent agents and robots than the global population working and living alongside us.

For Alibaba, pushing AI from the cloud into the physical world is a crucial step in its grand blueprint for super artificial intelligence (ASI). In fact, Alibaba has long been laying the groundwork for its foray into embodied intelligence Since 2024, Alibaba Group has successively invested in a number of embodied intelligence companies, including ZhiJi Power, Xingdong Jiyuan, Yushu Technology, Xinghaitu, and Lingxin Qiaoshou. Alibaba Cloud led a $140 million financing round for ZhiBian Robotics in September this year.

Insiders pointed out that in the first half of last year, Alibaba began to focus on embodied intelligence, "investing in almost all embodied intelligence companies." This is because robots are the ideal carriers for large model capabilities to transition from the digital world to the physical world. By investing in companies like Yushu, ZhiJi Power, and ZhiBian Robotics, which have different technological routes, Alibaba is seizing the track.

Clearly, Alibaba's layout in embodied intelligence is not a spur-of-the-moment decision. To understand its motivation, it must be placed within Alibaba's current internal and external environment and its judgment on the future of technology.

First, this is a strategic necessity: the endpoint of AI must be the physical world. As the capabilities of large language models become more refined, their further value release will inevitably require them to perceive, understand, and transform the physical world. An AI that can only operate in a dialogue box has a visible commercial ceiling, but an AI that can interact with the environment through robotic entities has unlimited potential.

When viewed in the context of the global AI landscape, Alibaba's actions are not isolated.

In the same week, SoftBank Group announced a $5.4 billion acquisition of ABB's industrial robotics business; Jensen Huang, CEO of NVIDIA, also mentioned again that AI and robotics represent "trillions of dollars in long-term growth opportunities." Elon Musk sees his humanoid robot "Optimus" as the core of Tesla's future value, aiming to first disrupt its own factory production model.

Global giants are entering the arena, and physical AI has become the next battlefield. For Alibaba, the foundation of its commercial empire—e-commerce, logistics, and new retail—is closely linked to the physical world. Whether it is Cainiao Network's smart warehousing spread across the globe or Hema Fresh's automated supply chain, there are numerous scenarios that require interaction with physical entities.

It can be said that these scenarios are the most direct and broad application testing grounds for embodied intelligence. Therefore, "embodying" AI capabilities is an endogenous demand for upgrading its core business.

In the face of such fierce competition, if Alibaba is absent, it would be tantamount to handing over the future entry point to the physical world to others. This is Alibaba's urgent need to tell a new growth narrative to the capital market and inject a "shot in the arm."

In recent years, Alibaba's core e-commerce territory has faced challenges from new business formats like Pinduoduo and Douyin. At this time, a new imaginative story is urgently needed to reignite investor confidence. Cloud computing and AI are undoubtedly the best protagonists of this new story, and the addition of embodied intelligence adds the most hardcore chapter to it.

It extends Alibaba's capabilities from online to offline, from virtual to reality, sending a clear signal to the capital market: Alibaba is evolving into a technology platform company that can define the future shape of the physical world.

Industry researchers pointed out to Wall Street that by directly participating in robotics research and development, Alibaba can both validate the adaptability of large models in real physical environments and drive the evolution of models in areas such as noise processing and causal reasoning; it can also expand data collection dimensions, extending training scenarios from text and images to industrial-grade applications like mechanical control and sensor feedback This transformation from business model innovation to hardcore technology innovation is key to reshaping its market valuation and opening a second growth curve.

Bridging the Industry Gap

"Now AI can write and create art better than 99.99% of people. But truly getting AI to work is still a barren land." A month ago, at the Bund Conference, Wang Xingxing, founder of Yushu Technology, said this during a roundtable forum.

It can be said that we are currently in a technological "discontinuity" zone. On one side, the wisdom of AI "brain" large models is exploding at an exponential rate, while on the other side, the clumsiness of AI "bodies" and the complexity of the physical world form a stark contrast. This gap limits the potential of AI.

Players at the table have reached a consensus: enabling AI to have a body and enter the physical world to "work" is a vast blue ocean.

Morgan Stanley predicts that by 2050, a cumulative total of 1 billion humanoid robots will be deployed globally, with a market size reaching $5 trillion. This figure is approximately twice the total revenue of the 20 largest automobile manufacturers in the world in 2024. According to estimates from China International Capital Corporation (CICC), the potential market space for humanoid robots in China is expected to reach 22.8 trillion yuan by 2050, with a compound annual growth rate of 24.7% from 2024 to 2050.

Such certainty is pushing the global embodied intelligence market to the brink of explosion. Various players and capital are scrambling to seize this historic opportunity. Companies like UBTECH, Yushu Technology, Galaxy General, and Zhiyuan Robotics have all announced plans to mass-produce thousands of humanoid robots this year.

In this heated atmosphere, institutional hot money has also rushed into the embodied intelligence track this year. Industry investors have noted that there have been frequent single financing rounds exceeding 500 million yuan in the embodied intelligence field this year, including Yushu, Galaxy General, Qianxun Intelligent, Xingdong Era, Zhujidi Power, and Tashizhi Hang.

A new trend of explosion is gradually emerging. However, the process of gold mining is not achieved overnight; how to fully integrate large models into embodied intelligence is the key bottleneck currently hindering AI's landing in the physical world.

And Alibaba aims to be the "shovel seller" that bridges this gap.

Previously, Wang Xingxing has repeatedly stated that compared to hardware, the greater challenge for embodied intelligence is the shortcomings in large models, "the biggest problem is that embodied intelligence AI is not sufficient."

Many large models have performed brilliantly in virtual scenarios, but once they enter the physical world, they become out of place. This is because traditional large models rely on massive amounts of internet text data for training, while embodied intelligence requires interaction data from the physical world, which is less in quantity and harder to obtain.

Moreover, Gao Fei, head of embodied intelligence at Alibaba Cloud Intelligent Group's public cloud division, pointed out to Wall Street News that in the coming years, the volume of industry data will steadily increase, and the corresponding pressure for cleaning, storage, and training will also amplify. "Relying solely on the efforts of embodied intelligence companies, it is very difficult to withstand such pressure."

Fortunately, Alibaba is well-prepared.

Currently, Alibaba's Qwen3-Max has surpassed one trillion parameters, excelling in complex task decomposition, capable of supporting robots in handling multi-step tasks such as "unpacking deliveries - classifying items - organizing and storing"; Qwen3-VL has enhanced 3D spatial perception and dynamic object tracking, allowing for precise identification of object locations and judgment of motion trajectories "Recently, I visited about 30 embodied intelligence companies and learned from in-depth discussions with their CEOs that most of them are using the Qwen-VL model for post-training," Gao Fei told Wall Street Insight.

The commonality behind this choice is backed by sufficient technical reasons. Qwen-VL's capabilities in spatial perception, dynamic visual understanding, and path planning precisely meet the core needs of embodied intelligence for a "brain."

In addition, Wang Xuwen, head of embodied intelligence solutions at Alibaba Cloud Intelligent Group's Public Cloud Division, pointed out that over the past few years, Alibaba has accumulated a wealth of infrastructure and practical experience by "stepping into pitfalls" while serving leading large model companies and new automotive forces. Because embodied intelligence and intelligent driving have a high degree of similarity in engineering architecture, this experience can be reused.

This full-link support capability from the model layer to the application layer gives Alibaba Cloud an almost unreplicable competitive advantage in the embodied intelligence track. It can be said that Alibaba Cloud is helping the entire industry clear obstacles on the road to mass-producing robots. When these capabilities shift from software simulation scenarios to the real world, embodied intelligence will truly enter an explosive moment.

Re-evaluating Alibaba: Full-stack AI Closed Loop

Thus, Alibaba's ambition is gradually becoming clear: the Tongyi Qianwen model serves as the general "brain," responsible for task understanding and planning; the hardware companies invested by Alibaba act as dedicated "execution units," responsible for action implementation. Tian Feng even believes that Alibaba has the potential to integrate its business with Tianshu Chip Technology. In the future, there may be opportunities to build an App Store-like application ecosystem in the robotics field.

In this global game concerning the future, each player holds different cards. In fact, Alibaba has already combined and cultivated a unique set of underlying cards—a "full-stack AI" closed loop covering computing power, cloud, models, and application scenarios. This is not only its capital for participating in competition but also the core of reshaping value and retelling the narrative of "greatness."

Looking back at the market, we can see two distinctly different growth paths for giants. Take SoftBank Group as an example; its founder, Masayoshi Son, is an outstanding "hunter" who quickly pieced together a global "information revolution" ecosystem through external acquisitions and investments.

In contrast, Alibaba's path resembles a form of "farming," relying on its vast business ecosystem to cultivate the soil for technology growth internally. The advantage of this model lies in the natural connection between technology and scenarios, which creates a unique flywheel effect in the era of embodied intelligence.

This "full-stack AI" closed loop can be clearly deconstructed. Its foundation is Alibaba Cloud, as a global leading cloud computing platform, providing massive and flexible computing resources for AI research and operation. Its brain is the Tongyi Qianwen large model, developed based on powerful computing power and refined in multiple internal scenarios at Alibaba. Its body consists of the robotic hardware and control algorithms that are being supplemented through internal team formation and external company investments.

The most core and difficult-to-replicate advantage of this closed loop is its unparalleled testing ground—rich application scenarios. From Cainiao Network's millions of square meters of warehouses to Hema Fresh's automated stores, and to the massive product processing of Taobao and Tmall, Alibaba possesses the largest and most complex real physical commercial scenarios in the world These four levels together form a closed loop: real scenarios generate massive amounts of data; data is input into Alibaba Cloud for computation and model training; the optimized "Tongyi Qianwen" brain directs robots; robots execute tasks in real scenarios and generate new data, which in turn further optimizes the model.

Once this flywheel starts turning, its iteration speed and efficiency will be astonishing. It allows AI models and robot algorithms to rapidly evolve in response to feedback from the real physical world, far surpassing competitors that can only be tested in laboratory environments.

Therefore, betting on embodied intelligence is Alibaba's inevitable choice in its attempt to "be great again" through AI narratives. Alibaba is not only a successful commercial company but also a hardcore technology company with core underlying technologies.

Industry investors point out that only a full-stack approach can make AI a true infrastructure, allowing AI applications to be more easily adopted across various industries. Under the vast imaginative space of full-stack AI, Alibaba's valuation logic is further reinforced.

Recently, Goldman Sachs significantly raised its forecast for Alibaba's capital expenditures for the fiscal years 2026-28 to 460 billion yuan. Goldman Sachs believes that the transformation of AI capital expenditures is reshaping Alibaba's growth expectations, stating that "the breakthrough progress in Alibaba's AI cloud computing capabilities and international expansion potential provides new upward momentum for its stock price."

The wheels of history roll forward, and every transformation of computing platforms gives birth to new kings. Today, we stand on the threshold of a new era defined by AI and robots. Alibaba, a giant that once defined a commercial era in China, is attempting to seize the reins of this transformation by building a full-stack AI closed loop from the cloud to the ground.

This is undoubtedly a high-risk gamble, but for a giant eager to "be great again," it may be the only correct path. The world is reassessing Alibaba, and the answer may lie within those "bodies" that are about to be endowed with intelligence