Shangwei New Materials' stock price doubled in 12 days: Zhiyuan-affiliated Peng Zhihui takes control, robotics business lost 422 million yuan in the first three quarters
Shangwei New Materials, now under the control of the Zhiyuan group, saw its stock price double over 12 trading days, with its market value briefly approaching 100 billion yuan on November 11 and its trailing price-to-earnings ratio reaching as high as 1,767 times. Cumulative output of the two consumer-grade robots, Qiyuan Q1 and T1, reached 10,000 units, with the latter 5,000 units produced in just 14 days, but the robotics business had yet to recognize revenue in the first three quarters, and the company posted a net loss attributable to shareholders of 422 million yuan. Over the same period, Unitree reported revenue of 1.817 billion yuan and net profit of 387 million yuan, with the two companies going head-to-head in humanoid robot shipments and the market.
Body · AI translation
Unitree Technology's recent stock price has already gone crazy enough.
On October 15, it fell to 404.03 yuan during intraday trading, and in less than a month, on November 10, it surged to 809 yuan, with total market value exceeding 320 billion yuan, and the stock price nearly doubled.
But in the humanoid robot sector, there is another stock that has risen even more sharply than Unitree.
SWANCOR Advanced Materials.
On October 23, it hit an intraday low of 103.8 yuan, and on November 9, it reached a high of 216.02 yuan, doubling in 12 trading days. On November 11, it climbed again to 244.20 yuan, with total market value approaching 100 billion yuan.
However, just as it rose sharply, it also fell frighteningly. On November 12, SWANCOR Advanced Materials plunged 11.36%, evaporating about 11 billion yuan in a single day; Unitree also fell 8.38%.
Unitree has Wang Xingxing, robot dogs and Spring Festival Gala performances, and is already a household name. So why has SWANCOR Advanced Materials risen so fiercely?
The answer cannot avoid one person: "Zhihui Jun" Peng Zhihui.
This former Huawei "genius youth" co-founded AgiBot after leaving. Last year, the AgiBot group gained control of SWANCOR Advanced Materials, and Peng Zhihui became chairman. A chemical company that originally produced wind turbine blade resin thus boarded the robot express train.
And AgiBot happens to be one of Unitree's strongest rivals.
According to a Huafu Securities research report, in 2025, Unitree ranked first globally with shipments of about 5,500 humanoid robots, followed closely by AgiBot with 5,168 units. In the first half of this year, AgiBot disclosed shipments of 9,700 units, which according to third-party statistics has surpassed Unitree.
SWANCOR Advanced Materials has not been idle either. On November 8, the company announced that cumulative production of its Qiyuan Q1 and T1 consumer-grade robots had reached 10,000 units, with the latter 5,000 units taking only 14 days.
The faster robots are built, the more the stock price surges along with it.
But opening the financial reports reveals another picture.
In the first three quarters of this year, Unitree earned 387 million yuan, while SWANCOR Advanced Materials lost 422 million yuan.
Although SWANCOR Advanced Materials is not the same as the entire AgiBot, a listed company whose robot business is still in its infancy and still burning large amounts of cash once saw its market value approach 100 billion yuan.
Unitree has already made money. So why does the AgiBot group make the capital market so crazy?
AgiBot and Unitree begin head-on confrontation
In the domestic humanoid robot sector, Unitree and AgiBot are already the two giants that cannot be avoided. But tracing the starting paths of the two companies, they took completely different business routes.
Unitree made its name with quadruped robots and entered the university, research institute and developer markets very early, especially accumulating many customers overseas.
Data shows that in 2025 Unitree's overseas revenue reached 732 million yuan, with overseas income accounting for more than 40%.
After that, its product line gradually extended to scenarios such as education, industry and commercial performances, and its appearance on the Spring Festival Gala stage helped it successfully "break out of the circle."
However, the ceiling of the research and developer market is clearly visible. If Unitree wants to continue pushing up shipment volumes, it must find major commercial customers capable of bulk procurement.
In contrast, AgiBot has attached greater importance to commercial applications since its birth, striving to push robots out of laboratories and directly into factories, warehousing and service front lines.
This road is not easy to walk. When industrial customers buy robots, they must consider not only whether the technology is advanced, but also stability, efficiency and cost; but the opportunity lies in the fact that once a product is recognized, customers may shift from purchasing a few units for testing to bulk orders of dozens or hundreds of units.
AgiBot's shipment growth largely comes from continuous development of these application scenarios. According to industry statistics cited by Huafu Securities, in the first half of this year, AgiBot's global market share was about 43%, ahead of Unitree's 31%.
However, being first in shipments does not mean being first in technology, revenue and profitability, and there are also differences in product structure and statistical calibers among different companies.
But this set of data at least confirms one point: AgiBot already has the scale to confront Unitree head-on.
And AgiBot is clearly not satisfied with this. Through SWANCOR Advanced Materials, it has also set its sights on the personal and household market.
This is a market completely different from industrial robots.
Factories buying robots calculate economic accounts: how much labor can be replaced in a year and how much production line efficiency can be improved; ordinary consumers buying robots look at the experience account: whether it is fun, whether it is useful, and whether it is worth spending money on.
The Qiyuan Q1 and T1 both start at 19,999 yuan, already much cheaper than humanoid robots that used to cost hundreds of thousands of yuan.
More importantly, SWANCOR Advanced Materials is trying to scale up the production of personal robots.
According to previous introduction by Gao Quanzhong, co-president of Qiyuan Robotics, when the production line started, daily capacity was only a dozen units, later raised to more than 200 units, with one robot rolling off the line every 2.5 minutes on average.
Previously, building 5,000 units took about a month, but the next 5,000 units took only 14 days.
The company also disclosed a set of user data: monthly active rate exceeded 98%, average daily online time reached 3 hours, and each user interacted with the robot more than 100 times per day on average.
These data at least show that early users have a certain enthusiasm for using the product. However, because the company has not disclosed the complete statistical sample and user composition, it is still impossible to judge ordinary households' acceptance based on this.
Just as AgiBot was pushing into the C-end, Unitree was also deepening its presence in the industrial market that AgiBot originally valued.
In early November, market regulatory authorities disclosed that Unitree planned to jointly establish a joint venture with companies under CNPC and others to carry out industrial embodied intelligent robot business.
On one side, Unitree is accelerating penetration into major industrial and commercial customers; on the other, AgiBot is using SWANCOR Advanced Materials to seize the consumer-grade entrance. The originally parallel business trajectories of the two giants are accelerating toward intersection.
There is another special aspect to AgiBot's layout: although SWANCOR Advanced Materials is controlled by the AgiBot group, it is a different operating entity from AgiBot Innovation.
The former mainly develops consumer-grade robots, while the latter focuses on industrial and commercial scenarios. AgiBot Innovation's shipments and revenue will not automatically be included in SWANCOR Advanced Materials' financial reports. This means that how much of AgiBot's scale advantage in the robotics industry can ultimately translate into SWANCOR Advanced Materials' performance still needs time to answer.
After all, Unitree has already started making money from robots, while SWANCOR Advanced Materials is still investing heavily in order to enter this market.
After the AgiBot group took control, co-founder Peng Zhihui became chairman of SWANCOR Advanced Materials, and CEO Tian Hua also has a Huawei background, having worked in product and business management for more than 20 years.
At the end of October this year, SWANCOR Advanced Materials adjusted its executive team again, appointing Zhao Gang as president and COO, Gao Quanzhong as co-president and CTO, and Liu Dongfang as co-president.
The three also come from Huawei. A group of Huawei veterans thus regrouped around consumer-grade robots.
Second is money.
A Huafu Securities research report shows that SWANCOR Advanced Materials' consumer-grade robot business has a total of 7 projects under research, with planned total investment of about 405 million yuan, accounting for 90.9% of the planned investment in all projects under research at that time. As of the first half of this year, about 171 million yuan had been invested.
Entering the second half of the year, the pace of spending accelerated further.
In the first three quarters of 2026, SWANCOR Advanced Materials' R&D investment reached 384 million yuan, a year-on-year increase of 1,182%, of which about 360 million yuan was directed to the consumer-grade robot business.
With this money spent, profits naturally do not look good.
In the same period, SWANCOR Advanced Materials' revenue was 1.209 billion yuan, down 5.51% year on year; net profit attributable to the parent company was a loss of 422 million yuan, down 796.94% year on year.
Interestingly, the company's original materials business is actually still making money, with net profit attributable to the parent company of about 63.52 million yuan in the first three quarters, up 4.91% year on year.
It is just that the huge expenditure on the robot business has forcibly dragged overall performance into a loss.
What is more awkward is that a lot of money has been spent, but robot revenue has not yet caught up.
The third-quarter report shows that Qiyuan Q1 and T1 officially went on sale on September 20 and were delivered gradually in October, and no robot business revenue was recognized in the first three quarters.
However, this does not mean there are no orders. As of the first half of this year, the company's robot business had received 210 million yuan in advance payments.
But advance receipts are not equal to revenue, and off-line volume is not equal to final delivery. How much performance the orders can translate into still awaits verification in subsequent financial reports.
By comparison, Unitree has already made money.
In the first three quarters of this year, Unitree's revenue was 1.817 billion yuan, up 55.71% year on year; net profit attributable to the parent company was 387 million yuan, up 270.54% year on year.
One company made 387 million yuan, and the other lost 422 million yuan, a profit gap of more than 800 million yuan.
Of course, SWANCOR Advanced Materials cannot be equated with the entire AgiBot group, and the two have different financial calibers. But for investors who buy stocks with real money, whether they can ultimately make money still depends on SWANCOR Advanced Materials' own profitability.
And more noteworthy than the loss is whether there is enough money to spend.
In the first three quarters of this year, SWANCOR Advanced Materials' net cash flow from operating activities was about 28.15 million yuan, down 74.59% year on year.
In the same period, net cash inflow from financing activities was about 489 million yuan, including 450 million yuan in related-party borrowings for robot R&D, supply chain construction and market promotion.
In other words, although the original business can still provide some cash flow, to support the rapid expansion of the robot business, SWANCOR Advanced Materials is currently highly dependent on external funding.
Even according to brokers' previous optimistic forecasts, this funding pressure will not disappear soon.
A Huafu Securities research report in October predicted that from 2026 to 2028, SWANCOR Advanced Materials' revenue will reach 2.509 billion yuan, 3.476 billion yuan and 5.786 billion yuan respectively, of which robot business revenue will be 620 million yuan, 1.488 billion yuan and 3.720 billion yuan respectively.
If the forecast comes true, by 2028 robots may become SWANCOR Advanced Materials' most important source of revenue.
But the cost is not small.
Huafu Securities expects that in 2027 and 2028, the company's R&D expenses will reach 466 million yuan and 860 million yuan respectively, totaling 1.326 billion yuan over the two years.
Even if it achieves 258 million yuan in net profit attributable to the parent company in 2028, net cash flow from operating activities may still be an outflow of 204 million yuan, net cash flow from investing activities an outflow of 504 million yuan, while net cash flow from financing activities is expected to be an inflow of 1.171 billion yuan.
That is to say, even if it becomes profitable in the future, SWANCOR Advanced Materials may still need continuous financing.
More noteworthy is that this research report was published before the latest third-quarter report, and some forecasts have already deviated significantly from actual conditions.
The research report at the time expected SWANCOR Advanced Materials to lose about 83 million yuan for the full year of 2026, but in the first three quarters alone the company had already lost 422 million yuan.
To achieve the original forecast, it would have to earn back about 339 million yuan in the fourth quarter, and the difficulty is self-evident.
And the capital market has already given a fairly high valuation.
On November 11, SWANCOR Advanced Materials' total market value once approached 100 billion yuan. The company's previous risk warning showed that, according to the statistical caliber at that time, its rolling price-to-earnings ratio was as high as 1,767 times.
The capital market can price the future of robots in advance, but whether consumers are willing to pay is another matter.
For a robot priced at nearly 20,000 yuan, why on earth would an ordinary family buy it?
This is probably more worth pondering for AgiBot and SWANCOR Advanced Materials than who builds the 20,000th robot first.
Robots have not yet learned to work, but prices have already fallen first
For AgiBot and Unitree, breakthroughs have already been made in robot mass production, but getting them to actually do work is still not easy.
For example, picking up a cup from a table requires almost no thought for a human, but a robot must not only identify its position, but also judge how much force to use and how to grasp it so it does not slip.
If the task is switched to folding clothes or tidying up clutter, the difficulty increases further.
One of the hurdles is training data.
At present, robot training relies heavily on visual data. Video lets robots see how humans work, but it cannot fully tell them how much force is needed to grasp something or what tactile feedback occurs when touching an object.
Without this information, even if a robot understands the action, it may not be able to imitate it accurately.
What is more troublesome is that high-quality data is not cheap.
According to industry research, unlabeled first-person video may cost only a few dozen yuan per hour, but once processed into high-quality data, the price may rise to several hundred yuan. If purchasing teleoperation data from real robots, it can even cost over a thousand yuan per hour.
Moreover, spending the money does not mean the robot will learn.
On November 4, Apple's research team disclosed the robot autonomous control system HALO and conducted real-environment tests on the Unitree G1.
Facing 6 unfamiliar environments that had not been previously scanned, the robot completed 360 single-skill tests, with an average success rate of 91%.
But when switched to complex tasks requiring the continuous execution of multiple actions, the success rate dropped to 70%.
This experiment cannot directly represent the capabilities of Unitree's or Qiyuan's commercial products, but it at least shows that teaching a robot one action and getting it to independently complete a task are still two different things.
Technical problems remain unresolved, yet competitors are increasing.
Among them, the most threatening is probably still Tesla.
Tesla is advancing mass production of the Optimus robot, planning to use part of the production lines at its Fremont factory to manufacture robots, and a dedicated factory in Texas is also under construction.
Musk has also previously proposed long-term capacity targets of one million units or even larger scale.
By then, AgiBot and Unitree will not only have to compete over who can build faster, but also over who is cheaper and more useful.
In fact, price pressure in the robot market has already begun to show.
The most typical example is the commercial performance rental market.
According to a Jiemian News report on November 6, in early 2025, the daily rental price of some humanoid robots was still as high as about 20,000 yuan, but by 2026, daily rental quotes in the thousand-yuan range had already appeared.
Some robots are mostly idle most of the time, and it is not uncommon for them to be powered on for fewer than 10 days a month.
The second-hand market has also shown obvious depreciation.
Some old engineering prototypes priced at 300,000 to 800,000 yuan in 2024 to 2025 now may have second-hand quotes of only 30,000 to 60,000 yuan, with the residual value of some models shrinking by nearly 90%.
Of course, different models and usage conditions vary greatly, and it cannot be concluded from this that the entire robot market is depreciating.
But it at least shows that robots that could once sell at high prices simply because of scarcity are becoming less rare.
In the past, a humanoid robot that could dance appearing in a shopping mall was enough to attract many onlookers. Now that robots are increasingly common, dancing, greeting guests, and taking photos are no longer worth as much.
Excitement can attract an audience, but it may not sustainably attract paying customers.
Industrial customers buying robots can calculate how much labor they save and how much efficiency they improve. But for household consumers, if a robot bought for 20,000 yuan can only chat and dance, and after being played with for a while is left in a corner, it will probably be hard to attract a large number of buyers.
Moreover, many of Qiyuan's current early users may be tech enthusiasts and fans of Zhihui Jun, who were already willing to pay for new technology.
These people can become the first customers, but they do not necessarily represent broader consumer demand.
This also means that although AgiBot and Unitree are both expanding the market, the tests they face are not entirely the same.
Unitree needs to prove whether its existing profitability can continue to strengthen as shipment scale expands; SWANCOR Advanced Materials needs to prove that personal robots are not only a novelty demand, but can also generate stable sales and revenue.
Over the past dozen or so trading days, SWANCOR Advanced Materials' stock price has already demonstrated the capital market's enormous expectations for robots.
But stock prices can rise in advance, while consumers will not pay in advance for features that have not yet been realized.
The rollout of 10,000 units proves that SWANCOR Advanced Materials already has mass production capability. Next, it still has to prove how many of these robots can actually be sold and how much money can be earned back.
The capital market has already written a script for SWANCOR Advanced Materials with a market value of 100 billion in advance, but whether this battle can ultimately be won still depends on whether the AgiBot system can truly turn robots into a profitable business.
References:
Huafu Securities, "SWANCOR Advanced Materials Builds a Leading Domestic C-End Embodied Intelligence Brand"
Yien, "Selling 2 Billion in Half a Year! This Company Rises to Global No. 1"
Robot Intelligence Era, "The 10,000th Qiyuan Robot of SWANCOR Advanced Materials Rolls Off the Line: A Victory in Production Capacity, or the Beginning of Inventory?"