Maschs Robotern-Projekt kann nicht länger verborgen bleiben.
During Tesla's third - quarter earnings conference call, Elon Musk hardly talked about cars. Instead, he completely steered the conversation towards the humanoid robot Optimus. He announced that Optimus V3 will be released in the first quarter of 2026 and planned to start a production line with an annual capacity of one million units by the end of next year.
Behind the future of robots outlined by Musk lies a fierce global competition in the field of Embodied Intelligence. In the first half of 2025, there were already 141 financing and investment events in the Chinese market, including 51 with a single - transaction value of over one million yuan. The total financing amount was already more than twice that of the entire previous year.
Previously, the public's perception of Optimus was still firmly at the level of "cutting - edge technology R & D". Even when the second model showed better mobility at the World Robot Conference in early 2025, most analysts didn't expect mass production until after 2026.
But Musk's remarks have broken this wait - and - see attitude. The annual production target of one million units is based on Tesla's experience in supply - chain management and mass production in the electric - vehicle field, as well as the efficiency improvement in AI training thanks to the supercomputer Dojo.
Most importantly, if production exceeds the one - million - unit mark, the price of Optimus could drop to $20,000. This price would completely break through the access thresholds for the consumer and industrial markets, similar to how the Model 3 once transformed the electric - vehicle market, thus initiating the breakthrough of humanoid robots.
This means that the underlying logic of this change is driven by technological breakthroughs and market demands.
Among international companies, Tesla has completed the iteration of Optimus from prototype to production model, Figure AI has implemented core processes in a BMW factory, and OpenAI has achieved the implementation of cognitive technologies in application scenarios through investments in hardware carriers. The competition is also fierce in the Chinese market: Unitree has introduced the new humanoid robot R1, and Zhipu has developed a solution for Embodied Intelligence thanks to its strengths in large - language models and has received eleven rounds of financing worth billions of yuan in the two years since its establishment.
01
Why are humanoid robots so popular?
"In the past three years, the popularity of the Embodied Intelligence sector has reached a new high almost every six months. Now we are on the verge of turning technological concepts into commercial value." An analyst who has been following the robot sector for a long time told the magazine "New Vision".
If we take 2022, when Tesla first introduced Optimus, as the starting point, the development of Embodied Intelligence in the past three years can be clearly divided into three phases. The technological breakthroughs and the strategies of market participants in each phase have jointly promoted the breakthrough of humanoid robots.
Specifically, the period from 2022 to 2023 was the phase of motion control. The core feature was "enabling the robot to balance and walk". In this phase, the industry's technological focus was on optimizing the mechanical structure and developing basic mobility. The players in this phase were mainly technology - oriented startups and technology companies. The financing was concentrated in the seed and angel rounds. The industry consensus was "to achieve technical feasibility first and then talk about commercialization".
From 2023 to 2024, it entered the phase of multimodal perceptual fusion. The key point was "enabling the robot to see and hear". With the maturity of multi - sensor technology and the iteration of AI algorithms, humanoid robots began to be able to interact with the environment. In this phase, the number of market participants increased sharply, and the financing was concentrated in the Series A rounds. Companies started small - scale application tests, with industrial assembly and logistics transportation being the preferred areas.
From 2024 to 2025, it came to the phase of cognitive awakening. The core change was "enabling the robot to think and make decisions". In particular, the commercialization of super - models like GPT - 4.5 provided strong cognitive support for Embodied Intelligence. A typical feature of this phase was the sharp increase in financing volume and frequent large - scale financings. International companies and top Chinese companies started planning for mass production, and the competition shifted from "technological competition" to "commercialization competition".
In this global competition, there are obvious differences between the international and Chinese markets for Embodied Intelligence. In terms of the technological route, international players attach more importance to "full - stack self - development" and "in - depth application - case orientation": Tesla uses its technological experience in the automotive sector to achieve vertical integration of FSD chips, vision systems, and motion algorithms. Figure AI focuses on industrial application scenarios and has established a deep partnership with BMW to improve the product through practical application.
In contrast, Chinese players are characterized by "point - based breakthroughs + ecosystem cooperation". They continuously achieve technological breakthroughs in core components such as grippers and sensors, but the ability for whole - machine integration still needs to be improved. Most companies choose to partner with application companies in the lower part of the value chain and open up the market through "customized solutions".
Behind these differences are technological accumulation, industrial foundation, and market demands.
Thanks to their long - term experience in AI, automotive manufacturing, and industrial automation, international companies have the ability to integrate the entire value chain. The foreign industrial market has a higher demand for high - precision robots and is willing to pay a premium for new technologies. The Chinese market, on the other hand, is restricted by the dependence on imported core components and the lack of unified standards. The cost of whole machines remains high, and the consumer market is price - sensitive. This drives companies to first achieve point - based technological breakthroughs in niche markets.
In addition, policy orientation also affects market development. China's support measures for Embodied Intelligence focus more on "the development of the entire value chain", while foreign countries tend to focus on "technological innovation and free competition in commercialization". These differences further shape the strategies of international and Chinese players. Although international companies adopt the full - stack strategy and Chinese companies aim for point - based breakthroughs, they all lead to the same result: the technological barriers for humanoid robots are dropping rapidly, and the commercial value is becoming more and more obvious. This is the core reason for the continuous popularity of humanoid robots.
02
What big game is Musk planning?
Every move of Tesla's Optimus affects the industry. Behind this fact lies not only the technological progress of the product itself but also Musk's strategic re - orientation for Tesla's future.
If we peel off the technological shell of Optimus, its uniqueness becomes apparent: it is not just a single robot product but the core carrier of Tesla's "AI + manufacturing" ecosystem. It is the cornerstone of Musk's vision of "bringing artificial intelligence into the physical world". If we look at the development of Optimus, we can see that every important phase involves strategic decisions.
When Optimus was first introduced in 2022, the prototype was clumsy in its movements, and the public even doubted the maturity of the technology. But Musk decided to introduce it to the public. The core of this decision was not product presentation but "positioning" - by attracting the industry's attention early on to draw resources from the supply chain and technical personnel, while at the same time sending a signal to the market that Tesla is transforming from an "automobile manufacturer" to an "intelligent - agent company".
In 2023, Tesla transferred its FSD self - driving technology to Optimus. This strategic reuse was like a "move from a higher level": the FSD chip provides the high computing power required for real - time environmental modeling of the robot, and the autopilot vision system solves the perception problem. This not only reduces R & D costs but also accelerates the maturity of the technology. This "technology transfer" is the core competence of Optimus.
In 2024, Musk launched the expansion plan for the supercomputer Dojo and invested huge sums in parallel robot training to shorten the learning cycle for new skills to 24 hours. This decision directly targets the core problem of mass production - the efficiency of AI training. Only by strengthening supercomputers can the robot be quickly adapted to different application scenarios and lay the foundation for mass commercialization.
During the third - quarter earnings conference call in 2025, Musk announced the annual production target of one million units and the release plan for Optimus V3. This marks the shift of strategic focus from "technological iteration" to "commercialization". The strength for all this is Tesla's experience in mass production and supply - chain management in the electric - vehicle market. In other words, Musk's expectations for Optimus go far beyond what a "product" can be.
He has publicly stated several times that in the future, about 80% of Tesla's value will come from Optimus robots. His ultimate goal is to create "general - purpose humanoid robots" that can be used in various application scenarios such as industrial production, household services, and medical care. In Musk's vision, Optimus will become the next - generation intelligent end - device after the personal computer and the smartphone and change the interaction between humans and machines. Tesla will build a closed business cycle of "hardware + AI services" through the large - scale deployment of robots and thus completely free itself from dependence on the electric - vehicle business.
It is worth noting that these expectations are not castles in the air. A report from Morgan Stanley predicts that the global market for humanoid robots will reach a volume of $5 trillion by 2050. If Tesla can capture 10% of the market, the Optimus segment alone would generate a revenue of $500 billion, far more than the peak value of the current electric - vehicle market business.
With the acceleration of Optimus' mass production, the Embodied Intelligence market is entering a "differentiation phase", and the differences in the strategies of different players are becoming more obvious. Musk has chosen the strategy of "mass production + general applicability": by leveraging Tesla's manufacturing strengths, he first achieves mass production in industrial application scenarios, reduces the cost per unit, and then gradually penetrates the consumer market. In the end, he wants to create a general - purpose robot. The core of this strategy is "success through scale", and essentially, he is copying the successful logic of the Model 3 in the electric - vehicle market.
In contrast, most Chinese players choose the strategy of "niche markets + customization". Unitree focuses on industrial logistics and special tasks and optimizes the robot's load - bearing capacity and adaptability to the environment to meet the specific needs of certain industries. Zhipu, on the other hand, focuses on large - language models and provides AI solutions for traditional robot companies to avoid the risks of capital - intensive whole - machine manufacturing.
The reason for these strategic differences lies in resources: Tesla has world - leading manufacturing capabilities, supercomputer resources, and brand recognition and is able to bear the risks of R & D and mass production of general - purpose robots. Chinese companies, however, are restricted by the high cost of core components and the lack of manufacturing experience and therefore choose niche markets as breakthrough points.
03
Warning about bubbles in Embodied Intelligence
The popularity of humanoid robots reminds us of previous boom phases in the technology industry. However, historical experience teaches us that there are often bubbles behind the euphoria. From the hype around industrial robots around 2010 to the capital frenzy for service robots in 2018 and the excitement about Metaverse robots in 2021 - in each boom phase, there was an increase in financing and an influx of companies, but only a few players could survive the period.
Compared with previous phases, the current hype around Embodied Intelligence is greater, and the technological iteration is faster, but the problems that have emerged are also more obvious.
Data shows that in 2025, 74% of the financing for Chinese Embodied Intelligence companies was concentrated in the Series A round or earlier. Over 80% of the companies have not yet achieved large - scale revenue, and most products are still in the "prototype presentation" or "small - scale testing" phase. There is still a long way to go to achieve a real closed business cycle.
Looking at the progress of Embodied Intelligence worldwide, a common feature of the industry emerges: technology is ahead of commercialization. Internationally, although Tesla has set an annual production target of one million units for Optimus, as of the third quarter of 2025, the trial production was still below 1,000 units. The pressure on the electric - motor supply chain due to export controls on rare earths and the difficulties in adapting AI algorithms to complex application scenarios could delay mass production. Although Figure AI has made progress in its application in the BMW factory, it only covers four core processes and depends on customized environmental transformation, making it difficult to quickly transfer to other application scenarios.
Chinese companies are still in an earlier phase. Although Unitree's H1 model has achieved an improvement in mobility, it...