Humanoid robots sold for 800 million, but UBTECH hasn't reached the profit-making stage yet.

By | Guanchaotech Pro Gao Heng

In 2025, UBTECH delivered a financial report characterized by “structural changes that are greater than scale growth.”

For the full year, the company achieved operating revenue of RMB 2.001 billion, up 53.3%. Among this, revenue from full-size embodied intelligent humanoid robot products and solutions reached RMB 820 million, up sharply 2203.7% year over year. The revenue share increased to 41.1%, making it the largest source. At the same time, the company’s overall gross margin increased to 37.7%, up 9 percentage points year over year.

But on the other side, the picture is equally clear: while revenue grew and the structure improved, the company still recorded a net loss of RMB 789 million for the full year. Although this narrowed compared with RMB 1.159 billion in 2024, the company still did not achieve profitability.

This means that a key change has already taken place: humanoid robots are no longer just demonstrative technology—they have begun to become a business that can bring in revenue. But another more realistic question has arisen at the same time: when humanoid robots become the core of revenue for the first time, has UBTECH reached a commercial inflection point, or is it still in the validation stage?

01 Growth is being rebuilt by a single business

In 2025, UBTECH achieved operating revenue of RMB 2.001 billion, up 53.3%. In terms of total scale, this is a typical high-growth financial report. However, when you break down the structure, you can see that this round of growth did not come from the synchronized expansion of existing businesses, but from a restructuring of the revenue mix.

The financial report shows that the company’s revenue mainly consists of four parts: revenue from full-size embodied intelligent humanoid robot products and solutions; revenue from non-embodied intelligent humanoid robot products and solutions; revenue from other intelligent robot products and solutions; and revenue from other intelligent hardware equipment.

(Chart source: financial report)

Among these, the most significant change is “full-size embodied intelligent humanoid robot products and solutions.” In 2025, this business generated revenue of RMB 820 million, accounting for 41.1% of total revenue, up 2203.7% year over year. In 2024, revenue from this business was only RMB 35.619 million, accounting for about 2.7%. In one year, this segment jumped from an edge business to the company’s largest revenue source, directly changing the company’s revenue structure.

The financial report explains that, as the acceleration of large-scale scenario applications progresses, the company converts technological accumulation into order delivery, driving the business to rapidly expand its revenue. This implies that humanoid robots have moved from the previous technology validation stage into the actual delivery and application stage.

By contrast, the other business segments show clear differentiation.

“Other intelligent robot products and solutions” generated revenue of RMB 629 million in 2025, accounting for 31.4% of total revenue, down 16.9% year over year. The financial report explains that the revenue decline in this segment is mainly because, for logistics intelligent robots and related solutions, some projects had not completed delivery and acceptance by the end of the year, resulting in reduced revenue recognition.

“Other intelligent hardware equipment” generated revenue of RMB 499 million, accounting for 24.9% of total revenue, up 6.4% year over year, and remained relatively stable overall. “Non-embodied intelligent humanoid robot products and solutions” generated revenue of RMB 47.963 million, accounting for 2.4% of total revenue, up 15.3% year over year. Its scale is still small, and its contribution to overall growth is limited.

From a structural standpoint, in 2025, UBTECH is no longer driven jointly by multiple businesses, but has clearly shifted toward being pulled by a single segment: humanoid robots account for most of the incremental growth, while the other businesses either have limited growth or experience declines.

In other words, the key change in this year is not that revenue exceeded RMB 2 billion, but that the source of growth has shifted. The company is moving from “multiple businesses running in parallel” to “single-point drive by humanoid robots.”

02 High gross margin has not brought profitability

If you look only at revenue and gross margin, UBTECH’s 2025 financial performance already shows clear improvement. But if you break down the chain of “revenue—gross margin—expenses—impairment—profit,” you can see that profitability has not been established in parallel.

First, look at the gross profit side. In 2025, the company achieved gross profit of RMB 754 million, up 101.6% year over year. Gross margin increased to 37.7%, up 9 percentage points year over year. The financial report explains that full-size embodied intelligent humanoid robot products and solutions became the largest revenue source, improving the revenue mix and thereby lifting the overall gross margin.

That is to say, the improvement in gross margin comes more from “what they sell has changed,” rather than “efficiency has already improved.”

When you move to the expense side, this improvement does not continue to transmit.

In 2025, the company’s selling expenses were RMB 471 million, accounting for 23.5% of revenue; management expenses were RMB 336 million, accounting for 16.8%. The two expense ratios decreased compared with the prior year. The financial report attributes this to a reduction in share-based payment expenses.

By contrast, R&D expenses remain high. In 2025, R&D expenses were RMB 508 million, accounting for 25.4% of revenue. The financial report explicitly states that this is mainly due to the combined impact of the company’s R&D investment in embodied intelligence artificial intelligence and robotics technology and a reduction in share-based payment expenses.

This means that while revenue was amplified, the company did not reduce its intensity of technology investment; R&D remains the core direction of spending.

Beyond expenses, impairment factors further drag on profit. In 2025, the company accrued credit impairment losses of RMB 151 million, mainly arising from delayed payment on accounts receivable from certain government-related customers. Under the prudence principle, it accrued impairment provisions.

As a result, despite revenue growth, gross profit improvement, and some reductions in expense ratios, the company still recorded a net loss of RMB 790 million for the full year. This narrowed compared with RMB 1.160 billion in 2024, but the company has not yet achieved profitability.

If you connect this chain together, you can reach a clearer conclusion: UBTECH’s profitability situation has indeed improved, but this improvement mainly comes from changes in the revenue mix and reductions in some expense ratios—not from the commercial model already having stable profitability.

In other words, the company has proven that humanoid robots can generate higher-quality revenue; but it has not yet proven that this revenue can cover ongoing high investment and steadily translate into profits. Profitability is improving, but the business closed loop has not yet been established.

03 Commercialization is still in the early stage

When you look at growth and losses together, UBTECH’s current situation can be summarized as: growth is in place, but profitability is not. What truly needs to be answered is not the financial results, but the stage the company is at.

From the technology and product pathway, what UBTECH is building is not a single robot product, but an entire set of system capabilities centered on embodied intelligence. The financial report shows that the company has formed a full-size embodied intelligent humanoid robot product system represented by Walker S, Walker S1, and Walker S2, mainly targeting industrial manufacturing scenarios. At the same time, at the technology level, it has built a multi-layer architecture including the trillion-parameter foundation model Thinker, the Thinker-WM world model, and the brain network (BrainNet) and collaborative agents (Co-Agent).

Well-known technology and industrial analyst Peng Duyu told us the following analysis: based on this combination, the company is not merely making the robot itself. Instead, it is trying to integrate “robot hardware + AI models + multi-robot collaborative systems” into one. Its core path is to form a closed-loop capability of “data—model—application” through continuous data collection, model training, and scenario deployment. This also determines that its business model naturally relies on long-term technology investment rather than short-term hardware sales.

But in terms of actual deployment progress, this system is still in the early stage of application.

The financial report discloses that in 2025, the company sold 1,079 units of full-size embodied intelligent humanoid robots, up 35,866.7% year over year. Annualized production capacity exceeds 6,000 units. Meanwhile, related products have already been applied in typical workstations such as material handling, sorting, and quality inspection, and have entered scenarios including automobile manufacturing, intelligent manufacturing, smart logistics, and embodied intelligence data centers.

Li Rui, Executive Director of Beijing Qishijie Technology Co., Ltd., said: these data show that humanoid robots have moved from the laboratory to production environments and have begun to have some capacity for scaled delivery. But it is also necessary to see that, at present, the applications are still concentrated in specific workstations and localized processes. Fundamentally, this is single-point substitution rather than a systematic replacement of the entire production system.

This is also reflected in the way the financial report describes the current stage. The company defines its current stage as “officially stepping from the training stage toward commercial application,” rather than already entering a mature commercialization stage. This means that its commercial rollout is still in a transitional process from “it can be used” to “it can be replicated.”

On this basis, looking at the company’s future plans allows for a clearer understanding of the logic behind its current investments. The financial report shows that UBTECH will continue to advance R&D and mass production of new models in the Walker S series. At the same time, it plans to launch embodied intelligent humanoid robots for commercial and educational scenarios, expanding application scenarios such as guided tours, exhibitions, events, and scientific research. On the technology side, it will continue to iterate the BrainNet, collaborative agents, the foundation model, and the world model.

From these statements, it can be seen that the company’s current focus is not to expand the sales volume of a single product in the short term, but to obtain data from industrial scenarios, train models, optimize the system, and then expand into more scenarios—gradually promoting large-scale rollout of embodied intelligence.

Overall, what UBTECH is currently in is not a mature profitability stage, but a typical early commercialization stage: the technology pathway is already clear, products are starting to be rolled out, and the business model is being validated. However, there is still a gap from large-scale replication and stable profitability.

Conclusion

When you look at the business structure, profitability, and technology pathway together, the stage signals provided by UBTECH’s financial report are already relatively clear.

On the one hand, by scaling up full-size embodied intelligent humanoid robot products and solutions, the company has completed validation at the revenue level for the first time: humanoid robots are no longer merely demonstrative technology. They have become a business that can enter industrial scenarios, generate orders, and complete delivery. The change in the revenue mix essentially indicates that this path has successfully taken the first step.

On the other hand, judging by profit performance, expense structure, and actual rollout, this business has not yet formed a stable commercial closed loop. High R&D investment, impairment pressure, and continued losses all point to the same thing: the company’s current revenue growth is more driven by period-based scaling up than by sustainable profitability under a mature business model.

In other words, UBTECH has already proven one thing: robots can be sold. But it has not yet proven the other, even more critical thing: that this business can make money steadily.

Peng Duyu said that what will decide the company’s valuation and its stage next is no longer whether the technology system is complete, but two more specific questions: first, whether the industrial scenarios that have already been implemented can move from single-point applications to replicable large-scale solutions; second, whether capacity, delivery, and cost control can support sustained scaling up and ultimately convert it into stable profits.

For a company that is still in the stage of “moving from training toward commercial application,” this step is the true watershed.

Massive information and precise interpretation—only on the Sina Finance APP