With Tesla and EX Robots posing a strong threat, how long can Unitree maintain its lead with its "hardware breakthrough"?

If you follow the tech circle, many people have probably heard of a company that started with quadruped robots and became a sensation on social media at home and abroad with a product resembling a "robot dog," even being frequently reported by media outlets such as CCTV and the BBC. It is Unitree Technology. Since its establishment in 2016, Unitree Technology has been on a rapid growth trajectory. It has not only become the global leader in consumer-grade quadruped robots but also plans to officially submit an IPO application in the fourth quarter of this year, aiming to be the "first stock in the humanoid robot sector."

This is not a coincidence but a long - planned breakthrough.

Starting with quadruped robots, dominating the global consumer and industrial markets with "robot dogs," and now fully betting on humanoid robots, Unitree Technology's ambition is not just to "walk" but to "run" at the forefront of the entire humanoid robot industry. According to the latest data from the International Federation of Robotics, the global humanoid robot market size increased by 120% year - on - year in 2024 and is expected to exceed $100 billion by 2030. A report from the China Electronics Society also points out that in 2025, the scale of China's robot industry will account for more than 45% of the global total, with the humanoid robot segment leading the growth in the entire industry.

At the forefront of the trend and surrounded by giants, Unitree Technology's IPO may seem like a natural progression, but in fact, it is fraught with perils.

From "Science Fiction" to "Hot Trend"

While many robot companies are still relying on "humanoid - inspired PPTs" and rendering videos to raise funds, Unitree Technology's slogan "First quadruped, then humanoid" brought the wild fantasies back to reality. They not only strive for a human - like appearance but also care more about how to create robot products that are truly usable, marketable, and can be continuously iterated right now.

Since the launch of the first - generation quadruped robot XDog in 2016, Unitree has been focusing on the core technology of "high - dynamic motion control," constantly pushing the limits in motors, reducers, and control algorithms. In 2023, they launched the humanoid robot H1, which has a total of 44 degrees of freedom, a maximum running speed of over 5 meters per second, and can stably trot and go up and down stairs. It can even maintain balance when pushed, making it one of the very few robots in the world that can truly achieve dynamic bipedal walking.

What's even more impressive is that Unitree can turn technology into tangible products. While many laboratory robots are still at the "demonstration - only" stage, Unitree has achieved mass production of quadruped robots. The price has dropped from hundreds of thousands of dollars per unit at the beginning to $10,000 - $20,000 now, and they are truly sold to universities, research institutions, technology enthusiasts, and even corporate customers. This rapid closed - loop of "technology - product - market" has enabled Unitree to successfully navigate the most difficult first stage of robot entrepreneurship.

What is most impressive is Unitree's iteration speed. From motors to complete machines, and from quadruped to biped robots, they almost maintain a rhythm of "a major upgrade every year." For example, their self - developed M107 joint motor has a torque density of up to 220 Nm/kg, far exceeding many off - the - shelf components purchased by competitors. In terms of software, the motion control algorithm has evolved to the fifth generation, supporting closed - loop debugging from simulation training to real - machine learning, which significantly shortens the development time.

Unitree's core strategy is to highly emphasize "usability" rather than simply "showing off skills." Whether it's the anti - fall structure, waterproof performance, noise control, or battery life, they have carefully refined every detail that affects actual use. As a result, their robots are not only "able to walk" but also "easy to use and durable." In contrast, many similar products are still at the stage of "capable of demonstration but not practical" or "able to walk but not user - friendly."

The "Valuation Paradox" of Unitree's IPO

Although Unitree Technology has received a lot of investment from major institutions such as Sequoia, Meituan, Shenzhen Capital Group, and Matrix Partners, its valuation has been growing rather slowly recently. After the Series C+ round in 2023, its valuation was about $1.3 billion. Although there have been rumors of an IPO this year, there has been no new follow - up financing. Currently, the market generally believes that its pre - IPO valuation will be between $1.5 billion and $1.8 billion. This growth rate is significantly different from Zhipu Robotics' momentum of surging from $500 million to $2 billion in just one year.

Actually, not only Unitree, but several other companies are also going all out. Tesla has set up the production line for Optimus at its Texas Gigafactory this year and plans to start using it within its own factories in 2025. Relying on Huawei's Ascend and HarmonyOS ecosystems, Zhipu Robotics has signed pilot orders with several manufacturing enterprises. Dreame has directly applied humanoid robots to the household cleaning scenario, advocating the slogan of "one robot per household." Yuejiang is quietly promoting a composite robot solution of "collaborative robotic arm + mobile chassis" based on its existing light - industry customers. They may not be more geeky than Unitree, but they have stronger financial chains, deeper relationships with corporate customers, and more complete artificial intelligence teams, which are all essential elements for large - scale commercialization.

What's more troublesome is that in the long - run competition of humanoid robots, it is essentially a "hardware - software integration" battle. The real decisive factor is not the hardware but the "brain," that is, how AI can more intelligently command the body, adapt to the environment, and handle various tasks. Currently, Unitree does lead in motion control, but it has not yet established a similarly obvious advantage in AI capabilities such as environmental perception and reasoning decision - making. Players like Tesla and Zhipu Robotics are advancing on two fronts: "large - scale models + hardware." If Unitree cannot quickly make up for its AI deficiency, it may end up being just a "hardware supplier" and miss the opportunity to become a platform - level company.

On the other hand, customers are also realistic. The ability of a robot to "move" is just the first step. The reason they are willing to pay is that the robot can be "useful." Currently, there are still very few practical scenarios for humanoid robots. Industrial inspection, scientific research and education, exhibition demonstrations, and logistics handling are all still in the exploratory stage. Even if Unitree's products are cheaper and more open - source, if customers cannot see a clear return on investment within one or two years, they are likely to be put into the category of "wait - and - see."

Moreover, the "open - source and open" approach is not without risks. As the industry matures, more and more companies may choose to "close - source both hardware and software" and make profits through ecological barriers. If Unitree continues with the open - source route, although it can attract developers, in the long run, it will have to face the challenges of establishing a sustainable business model and building a moat for after - sales service.

Unitree Can't Just "Walk"

If Unitree fails to achieve commercial breakthroughs within three years, it may decline from a technology leader to a supporting role in the industry. First, its financial endurance will be tested: the R & D cost of humanoid robots is extremely high, and if it cannot form a sales closed - loop, it will become increasingly difficult to raise funds. Second, once companies like Tesla and Zhipu Robotics complete their ecological closed - loops, even if Unitree has better technology, it may be forced to withdraw from the mainstream market due to a lack of application scenarios.

Therefore, Unitree must quickly adjust its strategy. In the short term, it should continue to deepen and strengthen the application of quadruped robots in fields such as inspection, security, and education. These are the businesses that can generate cash flow at present and provide continuous support for humanoid robot R & D. At the same time, instead of working in isolation, it can consider opening up its technology platform and cooperating with automobile manufacturers, industrial groups, or AI companies to build an ecosystem and find application scenarios together.

In the medium and long term, Unitree must make up for its deficiencies in AI and interaction. It can either set up its own AI team or form a strong partnership with domestic large - scale model companies to develop a truly autonomous decision - making robot system. Additionally, it can consider launching cheaper and more lightweight models, such as upper - body - only or desktop - level humanoid robots, to lower the usage threshold and collect real - world data more quickly.

Although the humanoid robot industry is still in its early stages, the competition is already extremely fierce. The companies that will ultimately survive may not be the ones with the most advanced technology but the ones that can best balance technology, products, ecosystems, and business. The rise of companies like Unitree, Zhipu, Dreame, and Yuejiang marks that China's robot industry has moved from "imitation and follow - up" to "original innovation and leadership." Especially in underlying technologies such as motors, control systems, and motion algorithms, Chinese companies are now capable of competing with global giants.

The humanoid robot companies that will truly stand out in the future must be "all - around players" who understand hardware, are strong in algorithms, and can accurately define application scenarios. Unitree must accelerate its transformation from being "hardware - driven" to "technology + ecosystem - driven" and upgrade from a "motion control expert" to an "overall solution provider."

Otherwise, when Tesla Optimus starts mass - production and price cuts, when Zhipu Robotics' operating system gradually forms an ecosystem, and when Dreame rapidly expands its market through its channel advantages, Unitree, which has achieved a single - point overtaking with "ultra - efficient motors," may be forced to slow down before rushing towards its IPO.

This article is from the WeChat official account "Financial Insights" (ID: jrwaican), written by External Financial Observation and published by 36Kr with authorization.