VC Money Is Chasing Entrepreneurs with Autonomous Driving Backgrounds, as Embodied Intelligence Investment Shifts Toward Commercialization

[[**Core Summary**

The embodied intelligence primary market is eagerly courting entrepreneurs with autonomous driving backgrounds; multiple new teams have secured hundreds of millions in funding within months of founding. Capital focus has shifted from generic narratives to commercialization and data flywheels, while industry investment has cooled and screening standards have become stricter.

**Key Points**

**1. Autonomous driving "veterans" become the main fundraising draw for embodied startups**

Kunlun Xing Robotics completed three funding rounds and reached a valuation of over one billion dollars in less than 90 days of existence; founders of new teams such as Feimo Yi Technology and Moqi Intelligence mostly come from top autonomous driving systems including Huawei, Li Auto, and Horizon Robotics.

**2. The window for autonomous driving talent to switch fields has opened due to multiple factors**

**3. The capital narrative has shifted from universal myths to commercialization and data flywheels**

New companies place greater emphasis on vertical scenarios, mass-production self-sustainability, and real data closed loops; both Zhengqi Future and RIVR obtained funding or acquisitions based on the "data flywheel" logic.

**4. Investment has cooled, and investor screening standards have tightened**

Exit uncertainty increased in the second half of the year; the approaching deadline of the Hong Kong Stock Exchange special channel and rumors of IPO tightening disturbed the market, with early-stage companies bearing the pressure first; investors focus on team mass-production experience, scenarios, capital structure, and technological ceilings.

**5. Autonomous driving giants are a potential variable, and startups still have a window of opportunity**

If Huawei, Horizon Robotics, NIO, XPeng, and Li Auto enter the market, they will have advantages in capital, computing power, and scenarios, but technology routes have not yet converged and they lack large-scale revenue, so they will not rush in in the short term; startups need to build data and time barriers before giants enter.2026-10-11 15:07 VC money is chasing autonomous driving "veterans" — ChinaVenture quick overview

This article is from the WeChat public account: ChinaVenture, author: Wang Manhua, editor: Wang Qingwu

If you look carefully at the current embodied intelligence primary market, you will find a strong sense of dislocation.

On the one hand, several leading companies reported to have listing plans are falling into panic over "tightening humanoid robot IPOs"; on the other hand, a batch of newly emerging startups are coming on strong, and their fundraising pace, amounts, and investor backgrounds are all astonishing.

The most typical example is Kunlun Xing Robotics: it completed three rounds of funding in less than 90 days of existence, with its valuation directly exceeding one billion dollars, successfully joining the unicorn ranks. According to media reports, the company's first funding round was snapped up by top institutions less than three days after its registration.

Behind the rapid rise of Kunlun Xing Robotics, the hardcore backgrounds of the founding team clearly played a decisive role. CEO Ren Geng has held positions at Huawei, Alibaba, and ENN Group, with a career spanning hardware engineering, AI innovation, hundred-billion-level business operations, and global operations; another co-founder, Lang Xianpeng, was the "employee No. 1" of Li Auto's autonomous driving, built Li Auto's autonomous driving R&D system from scratch, led the delivery of end-to-end + VLM solutions, and before leaving Li Auto was also the head of the hardware body for Li Auto's humanoid robot.

In the words of investor Gaorong Ventures, this is a truly scarce team in the current embodied intelligence market.

If you spread out the list of startups that have recently obtained large funding, you will find that teams with autonomous driving backgrounds, like Kunlun Xing Robotics, have almost become standard — Huang Qingqiu, Yang Ruigang, Liu Nianqiu, Zhang Ying... all are "veterans" with rich technical experience in the autonomous driving field.

It is worth mentioning that along with this wave of entrepreneurship, the "taste" of money has also changed. In the past, the primary market was willing to pay for the endgame imagination of "general-purpose robots" — the bigger the story, the fancier the technology, the higher the valuation; but in the narratives of this wave of new companies, the most frequently appearing keywords have been replaced by commercialization, data flywheels, and vertical scenarios.

Whether it is the scramble for autonomous driving "veterans" or the emphasis on commercialization and data flywheels, to some extent both reflect the current capital market's shift in attitude toward embodied intelligence — compared with grand endgame stories, they now value "certainty" and "sense of security" more.

Autonomous driving "veterans" collectively turn to embodied intelligence

Let's start with the most recent news.

Last month, embodied intelligence startup Feimo Yi Technology announced the completion of nearly one hundred million dollars in first-round funding, led by IDG Capital, with Yunqi Partners, Didi Chuxing, Fosun RuiZheng, Yaotu Capital, and Hesai Technology participating. The company was founded earlier this year; founder Liu Nianqiu was formerly a technical partner at DeepRoute.ai, where he participated in promoting the implementation of DeepRoute.ai's mapless autonomous driving solution and the evolution of its technology system toward end-to-end models and VLA large models.

On the same day, media reported that Zhang Ying, head of Horizon Robotics' HSD, had resigned to start a business, and that the new company had completed multiple funding rounds just two months after its founding. Although the report described it as a Physical AI company, according to my understanding, this project also falls within the embodied intelligence field.

Going back further. In early July, Moqi Intelligence, founded at the end of 2025 by Huang Qingqiu, former head of Huawei's autonomous driving AI department, announced the completion of over one billion yuan in angel-round series funding, with both Alibaba and Tencent participating, and more than ten institutions including BlueRun Ventures, Legend Capital, Gaorong Capital, and Source Code Capital also appearing on the shareholder list.

It is worth mentioning that when the funding was officially announced, the company's first product had not even officially debuted, but its valuation had already exceeded seven billion yuan.

In addition, Niuwa Robotics, founded in February this year by Yang Ruigang, former director of Baidu's autonomous driving and robotics laboratory and former CTO of Inceptio Technology, also announced two consecutive funding rounds within three months, with investors including Plug and Play China Fund, Blue Lake Capital, Butong Capital, and Gongqingcheng Puyi Investment.

What has been listed above is only the projects newly established in the past six months. Earlier ones, such as Tashi Intelligence founded by Chen Yilun, former CTO of Huawei's Intelligent Automotive Solution BU autonomous driving, and Zhijian Dynamics founded by Jia Peng, former head of Li Auto's intelligent driving R&D, have also completed multiple large funding rounds this year.

And more projects are still under the water. Hazel, a practitioner in embodied intelligence, told me that not long ago, on a business trip to Beijing, they organized an industry gathering, and during the meal they heard indirectly: recently, quite a few people have left the intelligent driving industry to start embodied intelligence ventures, many of them from leading intelligent driving companies such as Huawei, Horizon Robotics, and "NIO, XPeng, and Li Auto." "The industry is really getting more and more competitive," Hazel sighed.

The more samples there are, the more it shows that this is not an aimless chase after a hype cycle, but a collective migration in a highly consistent direction. So why now?

In the view of Zhang Jiaqi, executive general manager of Chentao Capital, this is to some extent a "timely turn" after weighing the options.

He further explained that, at present, the technical routes of intelligent driving as a whole are rapidly converging, and leading players in urban NOA are successively entering the engineering implementation stage of mass production and volume ramp-up. For a core figure who has mass-produced both end-to-end and VLA, staying in the original system to go through another version would no longer yield high marginal career returns, and might also bring career uncertainty from a potential switch in technical route. "At this point, while the primary market still maintains enthusiasm for embodied intelligence, coming out to start a business can instead garner decent attention and valuation," Zhang Jiaqi said.

On the other hand, for intelligent driving talent, their capabilities can be transferred almost losslessly. Cao Jishan, founding partner of Yundao Capital, said that over the past decade, autonomous driving may have been the largest-scale and most demanding industrial training in the field of physical intelligence. It operates on open roads, facing torrential rain, hail, pedestrians suddenly darting out, e-bikes going the wrong way, and long-tail events that are almost impossible to exhaustively enumerate; at the same time, it must also meet low latency, high reliability, and fail-safe requirements. More importantly, it has already achieved large-scale terminal deployment and continuously generates data in real environments.

English text is being prepared.

In addition, in an article previously published by ChinaVenture, "Exclusive | Intelligent Driving 'Veteran' Liu Nianqiu Starts a Business, Secures Nearly $100 Million in First Round," Liu Nianqiu, founder of Feimoyi Technology, also mentioned a key point.

In the wave of embodied intelligence entrepreneurship two years ago, because the supply chain was not yet complete, building embodied intelligence at that point required a great deal of effort starting from basic components. Now the situation is very different—there are already a large number of motors and reducers on the market optimized specifically for robots, the difficulty of building the body has been greatly lowered, and at the same time, the supply of AI infrastructure is far more abundant than two years ago.

"This means we no longer need to spend one or two years 'reinventing the wheel'; instead, from day one, we can focus on thinking about the product itself and on building a closed loop of continuous iteration across product, data, and models," Liu Nianqiu said.

When timing, capability, and infrastructure all come together at the same time, the window for migration opens.

From the "universal myth" to the "data flywheel," capital begins to inspect the goods

It should be made clear that intelligent driving people have always been the main force in embodied intelligence entrepreneurship; this is not a phenomenon unique to this round. For example, in the earlier batch, Gao Jiyang, founder of Galbot, also came from the intelligent driving system. What has really changed is the story that this batch of newly emerging teams has told the outside world since the end of 2025, as well as the definition of a "good company" behind the capital.

In the past, the fanciest narrative in embodied intelligence entrepreneurship was the endgame imagination of general-purpose humanoid robots. At that time, as long as the team's background was strong enough and the story had flexibility, route adjustments within one or two years could be tolerated. But the startups that have secured funding in this round have changed their focus—they are more willing to clearly explain which real scenarios they will enter first, for whom, what problems they solve, how the unit economics work, and to emphasize "commercialization," "real data," and the "data flywheel" as key points.

For example, an embodied intelligence company I wrote about in July, Zhengqi Future, whose founder Wu Shulin once served as vice president of Baidu's intelligent driving business group before coming out to start a business. This company takes short-distance travel as its entry scenario, collects "real data with human intent," and forms a data flywheel through the QUORRA DoorMind world model and the QUORRA FLAT motion execution platform. The company completed three funding rounds within eight months, with investors including CDH VGC, Linear Capital, SAIC Hengxu, Bairui Capital, Zhengxuan Investment, and Chow Tai Fook Strategic Investment, among others.

Even a company like Kunlun Xing, which writes "general-purpose humanoid robots" into its positioning, is no exception; what the founder repeatedly says in public interviews is also "first get mass production and self-sustaining cash flow working."

In Cao Jishan's view, this is not entirely investors or startups "coining terms," but rather that the industry stage has changed. In the early days, "embodied intelligence" was more of an academic perspective. Everyone talked about papers, about humanoids, about a slightly sci-fi endgame. At that time, the story could be big, implementation could be far away, and the capital market was willing to give time. But as the concept moved from embodied intelligence to today's Physical AI, what corresponds is not a copywriting upgrade but a shift downward in focus: from telling stories to delivering results.

"If you put it at this point in 2026 and reassess whether a company is worth investing in, the ruler has to be shorter and harder. When the upper-layer large models and world models are already crowded to the point where giants have taken sides and entered the stage of consolidation and capitalization, the place where ordinary early-stage institutions can still pick up high odds is instead in the 'harder' industrial crevices," Cao Jishan said.

Zhang Jiaqi has a similar impression. He recalled that from the end of last year to the first half of this year, the market still clearly favored general-purpose model companies—investing in general-purpose could yield dozens of times returns in less than a year, while investing in solutions might yield only two or three times, so at that time a large number of scenario-solution companies had trouble raising funds; by 2026, the general-purpose track had already become "packed into a ball," and while competition raised risks, companies' commercialization pace also generally faced greater tests. "Under such circumstances, some capital shifted to investing in solutions. We are also very willing to invest in such companies, because the commercialization path and PMF are core competitiveness at the company level."

In addition, Zhang Jiaqi also offered a reference point: RIVR, which was acquired by Amazon in March this year.

This is a wheel-legged delivery robot company headquartered in Zurich. Among its early investors were Linear Capital and Sequoia China, and later Bezos's Bezos Expeditions and Amazon also placed bets one after another. What the company tells is precisely a "real world + data flywheel" story: every door-to-door delivery generates real data, and the data is then fed back into the model through large-scale reinforcement learning and self-supervised training, rolling toward general-purpose physical AI.

In his view, such targets being taken over by giants with real money has, to some extent, also driven domestic teams and capital to pay attention and place bets.

Less money, stricter ruler

Although many recent cases of large financing were mentioned above, after the overall discussions, everyone's consensus is: entering the second half of the year, the primary market's investment in embodied intelligence has begun to cool down.

A major reason behind this comes from the growing uncertainty of exit channels. For a period of time, the approaching deadline of Hong Kong's special listing channel and rumors of "tightening IPOs for humanoid robots" undoubtedly disturbed investors' sentiment; and regulators' statement of "no barbaric growth, return to value creation" also led many institutions to tighten their pace of deployment.

Once the "engine" driving the trend stalls because of listing expectations, those first affected are often early-stage companies that have just started. And when money becomes less, selecting "people" begins to become stricter.

Cao Jishan said that when a Physical AI company comes to raise funds, he mainly looks at four indicators:

In the early stage, first look at the team. Whether the founder has done truly complex projects, whether core members have cross-disciplinary capabilities in technology and commercialization, and whether the team's ceiling is high enough.

Second, look at the scenario. Whether the market is large enough, whether demand is real, and whether the vertical scenario entered today can expand to more customers or adjacent industries.

Also look at the capital structure. Embodied intelligence often requires long-term R&D, hardware iteration, and customer validation; it is a business that needs continuous financing. Team tension, shareholder structure, and subsequent financing capability will all affect whether it can reach commercialization.

Finally, look at the technical ceiling. Scenarios will change, models will upgrade, and today's solutions may quickly become outdated. The team must not only solve immediate delivery but also have the ability to keep up with changes in underlying paradigms.

As for those autonomous-driving talents transitioning into the embodied industry, Zhang Jiaqi also has a set of screening logic. In his view, a high-potential top embodied autonomous-driving team must simultaneously possess four capabilities—1) practical experience in implementation and mass production, which must come first; 2) having trained end-to-end models with sufficiently large parameter counts; 3) having processed massive amounts of driving data; 4) having used computing clusters at the level of nearly ten thousand GPUs.

"Screened by this standard, there are probably only about 20 people across the entire industry, and if enterprise management and operations are also taken into account, the number may be even smaller," Zhang Jiaqi said.

During the discussions, many investors also mentioned a major variable that could change the future landscape of the embodied industry, namely the autonomous-driving giants. The technical advantages of the autonomous-driving industry's transition to embodied intelligence have been discussed above, and compared with startups, once giants such as Huawei, Horizon, NIO, XPeng, and Li Auto enter the stage, in addition to so-called technical advantages, they also hold large amounts of cash, computing power, and scenarios, and their impact is self-evident.

However, they also stressed that the technical route for embodied intelligence has not yet converged, and certain large-scale revenue has not yet appeared, so in the short term giants probably will not rush in.

This means startups still have a window period for now. And what they need to do is build higher data and time barriers before the giants enter.

AI industry signal channel: Frontier Technology. This content comes from the internet and the original link. The views represent only the author and do not represent Huxiu's position. If there is any copyright issue, please contact hezuo@huxiu.com, and we will verify and handle it promptly. Everyone who is changing and wants to change the world is on the Huxiu APP[]