This group of people are successively getting financing.

Recently, there have been continuous financing news in the embodied intelligent robot track, highlighting the strong momentum of the industry's development and the continuous optimism of capital towards this field.

Zhipingfang completed a new round of Series A financing led by Shenzhen Capital Group, with Shenzhen Capital Group's single investment exceeding 100 million yuan. Meanwhile, Boundless Power (Beijing) Technology R & D Co., Ltd., a startup founded by Zhang Yufeng, the former vice - president of Horizon Robotics (a listed company in the intelligent driving chip field) and the former president of the intelligent vehicle business unit, will complete its first - round financing soon.

This series of developments highlights the industry trend that after senior practitioners in the intelligent driving field cross - over into the embodied intelligent track, their enterprises can quickly gain the favor of capital.

Journalists from Venture Capital Daily noticed that in recent years, many top experts in the autonomous driving field have left their original positions and entered the embodied intelligent track to start businesses, and completed large - scale financing in a short period. A new wave of financing frenzy in the embodied intelligent track is expected to emerge in 2025. These entrepreneurs are injecting the engineering experience, data iteration systems, supply - chain management capabilities, mature product thinking, and business models accumulated in the automotive industry into the field of embodied intelligence, promoting the transformation of robots from "feature phones" to "smartphones".

In addition, the popularity of the embodied intelligent track reflects the huge opportunities brought by technological integration and innovation. The accumulation of intelligent driving technology in perception, decision - making, and control is highly relevant to the technology required for embodied intelligent robots, enabling these cross - over entrepreneurs to quickly find entry points in the new field.

Completed 7 Rounds of Hundreds of Millions - Level Financing in Half a Year

It has become the norm in the industry that startups in the field of embodied intelligence founded by elites in the autonomous driving industry can quickly gain the favor of capital and complete large - scale financing in a short period.

In this round of financing, Zhipingfang not only received continuous investment from its old shareholders, Dunhong Assets and SDIC, but also attracted Bloomage Biotechnology and a large retail enterprise to participate as industrial capital. The funds will be used for the continuous iteration of Zhipingfang's GOVLA large - scale model and AlphaBot series of robots, the expansion of self - built production lines, and the expansion of the global market.

This is one of the 7 rounds of hundreds of millions - level financing that Zhipingfang has completed in the past half - year. Previously, the company successively received Pre - A+++ round financing led by Guozhong Capital and Series A+ round financing led by Puhua Capital, and also introduced investments from institutions such as Fortune Capital, Cornerstone Capital, and Nanshan Strategic Emerging Industry Investment.

When asked why they chose the "high - frequency + multi - round" financing method, a relevant person in charge of Zhipingfang told journalists from Venture Capital Daily that this is mainly due to the enthusiasm of the capital market for embodied intelligence and the recognition of Zhipingfang's technology and business implementation capabilities. "We choose the high - frequency and multi - round financing method. On the one hand, it is because investment institutions show high initiative in attention. On the other hand, it allows us to obtain resources and support more flexibly during the development stage."

Zhipingfang has launched three generations of robot products, namely AlphaBot, AlphaBot 1S, and AlphaBot 2, all of which are wheeled and liftable humanoid robots.

During the 2025 World Robot Conference, journalists from Venture Capital Daily saw on - site that the AlphaBot robot could move and stack boxes in a simulated factory, make coffee and ice cream in a soft - drink bar, and play the drums on the stage.

Currently, the number of orders for Zhipingfang's robots this year has exceeded 500. Regarding the sources of order customers, the above - mentioned person in charge introduced that the current orders mainly come from industrial service fields such as automobile manufacturing, general semiconductors, and biotechnology, as well as some public service scenarios. Now it is in the intensive delivery period, and the self - built production line will be officially put into use this month. The annual production capacity exceeds 1,000 units, which can fully complete the closed - loop of "independent R & D + independent manufacturing + independent delivery" and ensure that orders are completed on schedule.

"We will make dynamic plans according to market demand and order situations, and the production capacity is matched with the delivery rhythm. Currently, the capacity of the self - built production line can meet the demand, and there are no plans for contract manufacturing for the time being."

In terms of the technical path, as one of the first domestic enterprises to propose the end - to - end VLA (Vision - Language - Action) technology, Zhipingfang released the world's first global whole - body VLA large - scale model (GOVLA) in April this year.

When it comes to the industry positioning in the next three years, whether Zhipingfang wants to become the "Tesla of the robot industry" or the "Nvidia", the above - mentioned person in charge said that Zhipingfang positions itself as a general intelligent robot service provider integrating hardware and software. The company attaches importance to both the large - scale manufacturing of hardware and the continuous evolution of large - scale models and systems.

Collective Migration of Intelligent Driving Talents

Since 2024, a wave of senior executives starting businesses has emerged in the field of embodied intelligence. Industry leaders such as Liu Fang, the former head of Xiaomi Auto's autonomous driving technology, Chen Yilun, the former chief technology officer of Huawei's Vehicle BU's autonomous driving system, and Yu Yinan, the former vice - president of Horizon Robotics, have successively entered this track, forming a unique phenomenon of "migration of intelligent driving talents".

In March 2025, Zhang Yufeng founded Boundless Power, focusing on B - end industrial scenarios and creating dual - arm wheeled robots. The company is about to complete its first - round financing, and its core team has taken initial shape, including technical executives from Didi.

Zhang Yufeng is not an isolated case. At the end of 2024, after Yu Yinan, the former head of intelligent driving at Horizon Robotics, left his position, he joined hands with Zhao Zhelun, the former director of intelligent driving products at Li Auto, and Song Wei, the former chief architect of the software platform at Horizon Robotics, to jointly found Vita Power, also choosing the robot track.

Zhipingfang was founded in April 2023. Guo Yandong, the founder and CEO of the company, used to be the chief scientist and R & D executive at XPeng Motors and OPPO. The various intelligent systems he led have been put into use in hundreds of thousands of intelligent vehicles, hundreds of millions of consumer electronics terminals, robots, and Microsoft MaaS.

It is generally observed in the industry that talents with a background in intelligent driving account for a considerable proportion in this wave of embodied intelligent entrepreneurship. From Baidu Apollo, XPeng's autonomous driving to Waymo and other enterprises, the collective shift of technical backbones is reshaping the industry landscape.

Overall, the embodied intelligent track (including humanoid robots and large - scale embodied intelligent models) has become a gathering place for automobile enterprises, enterprises in the automobile industry chain, and top experts in the autonomous driving field. With their technological, resource, and experience advantages, this group has become the most competitive faction in the current embodied intelligent track. Currently, there are mainly three ways for players in the intelligent driving field to enter the embodied intelligent track:

Automobile enterprises directly enter the market by establishing special teams for humanoid robots through internal incubation;

Senior practitioners with a background in intelligent driving leave technology giants or automobile enterprises and independently start businesses to establish enterprises related to embodied intelligence, entering the track in the form of startups;

Leading enterprises in the intelligent driving industry chain choose to start from the upstream. They either enter the core component field of embodied intelligence (for example, RoboSense focuses on sensors, and CATL explores power systems), or reach strategic cooperation with mature humanoid robot companies and participate in the track competition through technological collaboration or supply - chain support.

Same Technological Origin of Intelligent Driving and Embodied Intelligence

The migration of intelligent driving talents to the robot field is not accidental. There are commonalities in technology between intelligent driving and embodied intelligence. Both require core capabilities such as environmental perception, decision - making and planning, and motion control.

An executive in the field of embodied intelligence told journalists from Venture Capital Daily that the underlying technology of both autonomous driving and embodied intelligence is AI. In essence, intelligent vehicles and embodied robots are extensions of the same set of intelligent hardware systems in different scenarios: the former installs the "body" on wheels, while the latter makes the "body" in a humanoid shape. They not only share core components such as lidar, high - precision sensors, computing units, and operating systems, but also use perception, decision - making, and control algorithms and cloud - based data closed - loops.

Therefore, the transition from autonomous driving to embodied intelligence is not a cross - over, but a natural migration and upgrading of AI capabilities on "movable physical terminals". Autonomous vehicles have solved the problem of unmanned driving on structured roads, and embodied robots further extend the tentacles of AI to the unstructured and full - scenario physical world, naturally becoming the next - generation intelligent terminals after intelligent vehicles.

Journalists from Venture Capital Daily learned that the high overlap of the technology stacks between intelligent driving and embodied intelligence is the underlying logic for talent cross - over. The core technologies accumulated in the autonomous driving field are being reused in the field of embodied intelligence.

For example, in intelligent driving, the fusion algorithm of lidar, cameras, and millimeter - wave radar has become mature. Chen Yilun, the CEO of Tashii, previously revealed that many engineering practices in today's embodied intelligence are migrated from autonomous driving systems. For example, the underlying computing chips, software architectures, communication middleware, asynchronous reasoning systems for fast and slow thinking, and data closed - loops can all be directly reused.

He said that autonomous driving and embodied intelligence are of the same origin in technology, and there are many connections in development experience. Autonomous driving is a key sub - task of embodied intelligence, representing the ability of movement and navigation. Many commercial robots today still use a modular hierarchical approach similar to that of sweeping robots, that is, using SLAM + rule - based planning to complete tasks. In fact, they should switch to the end - to - end AI implementation of autonomous driving.

"Secondly, autonomous driving has stepped on a huge pit in advance for embodied intelligence - AI needs to be defined in a 4D space - time, rather than within a 2D image. The end - to - end system of autonomous driving actually unifies the perception, decision - making, and planning in space - time. All modalities, whether the input vision or the output actions, are unified in the space - time coordinate system." Chen Yilun believes.

It is worth mentioning that the massive annotated data and simulation training experience accumulated in the autonomous driving field are playing a role in the field of embodied intelligence. For example, 51Sim has migrated synthetic data from autonomous driving to robot simulation.

This article is from the WeChat official account "Venture Capital Daily", author: Li Mingming, published by 36Kr with authorization.