A post-90s AI genius raised 2.5 billion in just one month.

Investment community learned that today (April 16th), Jijia Vision officially completed a nearly 1.5 billion yuan Series B1 financing. The lineup of investors behind it remains luxurious:

It was jointly invested by a well - known technology giant, multiple top national - level funds, Jianling Capital (the CVC of Yili Group), Puhua Capital, Huafu Investment, Yida Capital, New Industrialization Fund, Qianrong Capital, Shengjing Jiacheng, Turing Asset Management, Kaiyang Capital, Wuhan High - tech Investment Group, Guiyang Jin Investment Group, Shandong Industrial Investment and other top - tier state - owned platforms, industrial capitals, and dual - currency financial institutions. Multiple old shareholders such as Huakong Fund, Huamin Investment, Yicun Capital, and Lingyang Investment continued to over - invest.

This means that Jijia Vision has raised 2.5 billion yuan in just one month, an impressively fast pace. So far, Jijia Vision's latest valuation has exceeded 10 billion yuan, making it the first domestic unicorn in the world model field valued at over 10 billion yuan.

The person at the helm is a post - 90s Tsinghua University doctor, Huang Guan. He has worked at Horizon Robotics and Jianzhi Robotics successively, and has also worked at Microsoft Research Asia and Samsung China Research Institute. In just three years, he has led Jijia Vision to rapidly rise as a global leading enterprise in the world model field. This scene is a vivid portrayal of the booming AI industry.

Raised 2.5 billion yuan

A new unicorn valued at over 10 billion yuan is born

This is another phenomenon - level enterprise in the venture capital circle.

Going back to its early days, Jijia Vision received several tens of millions of yuan in seed - round financing from Chentao Capital. After that, the financing pace was rapid. In September 2024, it completed two consecutive rounds of angel and angel + financing totaling nearly 50 million yuan, invested by institutions such as BAIC Industrial Investment, Miracle Plus, Huamin Investment, Longding Investment, Qingzhi Capital, and PKSHA Algorithm Fund.

Then in August 2025, Jijia Vision announced the completion of two consecutive rounds of Pre - A and Pre - A+ financing worth hundreds of millions of yuan. The Pre - A round was led by Guozhong Capital, with Zifeng Capital and old shareholder PKSHA Algorithm Fund following. The Pre - A+ round was invested by CICC Capital, Guangzhou Industrial Investment, Yicun Songling, and Huaqiang Capital.

Soon, Huawei also stepped in. In November of the same year, Jijia Vision completed a new round of Series A1 financing worth hundreds of millions of yuan, jointly invested by Huawei Hubble and Huakong Fund. When talking about this round of financing, Huang Guan once said bluntly: "Huawei has listed the world model as the top of the ten technological trends for the intelligent world in 2035. This is also the underlying logic for investing in Jijia Vision." Besides the investment, Huawei has also promoted strategic cooperation with Jijia Vision in multiple business lines.

One month later, Jijia Vision completed a 200 million yuan Series A2 financing. It was led by Fortune Capital, co - led by old shareholder Huakong Fund, and followed by well - known institutions such as Shoufa Zhan Venture Capital, Puyao Xinyie, Caixin Capital, Huajin Capital, Zhangke Yaokun, and Fuzhuo Venture Capital. Old shareholder Hedinggong Capital over - invested. This means that Jijia Vision completed 4 consecutive rounds of Series A financing totaling 500 million yuan in just three months.

Not long ago, the Pre - B round of financing came to light. In early March this year, Jijia Vision announced the completion of a nearly 1 billion yuan Pre - B round of financing. The investors included top - tier chip and automotive industry capitals such as SMIC Juyuan, Shanghai Semiconductor Industry Investment Fund, Linxin Capital, Xingyuan Capital, and Wanlin International, as well as heavy - weight state - owned platforms and well - known financial institutions such as CICC Capital, Suzhou Venture Capital Group, Huaqiang Capital, Yangtze River Capital, Optics Valley Industrial Investment, Xishan State - owned Investment, Jinyu Maowu, Xinding Capital, Lingyang Investment, Caixin Capital, Zhangke Yaokun, and Chengzhu Investment. Among them, old shareholders such as CICC Capital, Huaqiang Capital, Caixin Capital, and Zhangke Yaokun continued to over - invest heavily.

Today, the Series B1 financing is unveiled. As we can see: Almost all types of top - tier investment institutions in the market are behind Jijia Vision. This includes well - known technology giants, multiple CVCs of industrial giants, multiple top national - level funds, multiple first - tier dual - currency financial institutions, and state - owned platforms from various regions. Multiple old shareholders collectively continued to over - invest. Undoubtedly, the investors have cast a crucial vote with real money. So far, Jijia Vision has raised 2.5 billion yuan in just one month.

Post - 90s Tsinghua University doctor

Assembled an AI "Dream Team"

Through Jijia Vision, an AI "Dream Team" emerges.

Huang Guan is undoubtedly the soul of the team. In 2009, he was admitted to the School of Automation at Huazhong University of Science and Technology. He graduated as the top student in his major at the undergraduate level and then entered the Institute of Automation of the Chinese Academy of Sciences for a master's degree. Later, he became a doctor at the Department of Automation of Tsinghua University. During this period, he interned at Microsoft Research Asia, becoming one of the earliest domestic scholars to conduct in - depth learning research.

In 2016, Huang Guan joined Horizon Robotics as the person in charge of visual perception. He led the creation of the world's largest face recognition dataset at that time, WebFace260M, and led the team to achieve large - scale implementation in smart cities. Later, he participated in the establishment of Jianzhi Robotics as a partner, overall responsible for the autonomous driving business, and developed the widely recognized BEV series of models with the team.

Looking back, Huang Guan's scientific research achievements are impressive: the champion of the COCO Challenge held by ECCV in 2020; the world's first in the NIST - FRVT face recognition evaluation in 2021, which is known as the "golden event in face recognition"; in the same year, BEVDet ranked first in the pure - vision 3D object detection on the authoritative autonomous driving evaluation dataset NuScenes; in 2022, BEVDet4D again topped the NuScenes list... Looking back, every step of his entrepreneurship is traceable.

It is worth noting that in his previous career, Huang Guan has cumulatively led or participated in financing of over 2 billion yuan. It can be seen that Huang Guan is a rare composite leading talent in the industry who has top - level scientific research experience in the physical AI field, mass - production engineering experience, business implementation experience, and continuous entrepreneurship experience.

Until June 2023, Huang Guan led the team to start a business again and officially established Jijia Vision. An "AI Dream Team" was quickly assembled: Co - founder and Chief Scientist Zhu Zheng graduated with a doctorate from the Institute of Automation of the Chinese Academy of Sciences in 2019 and then conducted post - doctoral research at the Department of Automation of Tsinghua University; Co - founder Sun Shaoyan once served as the director of Alibaba Cloud and the general manager of the data closed - loop product line at Horizon Robotics; Partner & Engineering Vice - President Mao Jiming once served as an architect at Baidu and Yingche and was the person in charge of Baidu Apollo's simulation technology.

It can be found that this team has experienced the development of physical AI in the past decade. They have continuously achieved excellent results in technological innovation and industrial implementation at each stage, such as CV, autonomous driving, embodied intelligence, and world models. It is also a rare full - stack team with both "world - class innovation ability in the new - generation physical AGI" and "full - stack large - scale mass - production experience in traditional physical AI".

As early as the establishment of Jijia Vision, Huang Guan emphasized that "the world model is the next most important thing after the language model." Time has proven his judgment: After the beginning of 2026, the world model has become increasingly popular. The industrial influence and commercial value represented by physical AGI are no less than those of digital AGI represented by language models.

Quietly, Jijia Vision has become a pioneer in the world model field. Centered around the world model, Jijia Vision has successively released GigaWorld - 0, GigaWorld - Policy, and GigaWorld - 1, creating a landmark moment in the industry.

Among them, GigaWorld - 0 is the world's first milestone work to verify that "data generated by the world model can effectively improve the performance of real physical robots". Its open - source code on GitHub has received more than 1,500 stars, laying the important role of the embodied world model as a data engine in high - level embodied base models.

GigaWorld - Policy breaks the "speed - performance - efficiency" impossible triangle. It innovatively adopts an "action - centered" architecture, making video generation an optional item during inference, achieving a leap - forward improvement of 10 times in inference speed and 10 times in training efficiency, while the task success rate has increased significantly by 30%.

GigaWorld - 1 even defeated the models of international technology giants and academic institutions such as Google, NVIDIA, and Alibaba in the authoritative evaluation benchmark WorldArena in the world model field, ranking first globally.

At the same time, centered around VLA, Jijia Vision has also successively released representative models such as GigaBrain - 0, GigaBrain - 0.1, and GigaBrain - 0.5M*, which have conducted important verifications on the feasibility of Scaling.

All these are impressive.

Welcome to the era of physical AGI

As you can see, a world model storm is sweeping the globe.

Not long ago, World Labs, founded by well - known scientist Fei - Fei Li, just announced the completion of a new round of financing of 1 billion US dollars (approximately 7 billion yuan). In just over a year since its establishment, the valuation of World Labs has reached an astonishing 5 billion US dollars.

As the "Godmother of AI", Fei - Fei Li previously published a long article of ten thousand words, straightforwardly pointing out and proposing that large language models are still "wordsmiths in the dark", eloquent but lacking experience, knowledgeable but not well - rooted. "Spatial intelligence will change the way we create and interact with the real and virtual worlds, and building spatial intelligence AI requires something more ambitious than large language models: the world model."

The reason why the world model is so popular is that it is regarded as the core engine of physical AGI, filling the key link of "understanding and interacting with reality" for AI, and then enabling AI to move from "generating" to "understanding, deducing, and acting".

As the key carrier to physical AGI, embodied intelligence has become the first explosive application scenario of the world model. Further, the world model affects not only the model architecture and data source of embodied intelligence; more importantly, it changes the overall system architecture including world prediction pre - training, visual action pre - training, post - training, and data system.

Relying on this, Jijia Vision has launched the full - stack self - developed physical AGI native general robot Maker H01 - a humanoid robot with a high - degree - of - freedom hardware design, with more than 20 degrees of freedom throughout the body and equipped with a multi - modal high - precision perception system, designed for dexterous operation and mobile interaction tasks in open environments.

It is understood that Maker H01 has achieved commercial implementation breakthroughs in multiple high - value scenarios such as automotive, 3C, warehousing, and home services, and is aiming for a delivery volume of a thousand units this year.

Remember in April this year, Jijia Vision joined hands with FAW Mold and Alibaba Cloud to complete the full - process solution implementation of the embodied intelligent robot in a real industrial manufacturing scenario. This is the first domestic full - process solution centered around a real automotive industrial manufacturing scenario, focusing on high - frequency operation tasks such as box unstacking, cross - regional handling, dynamic obstacle avoidance, and precise operation, including data collection and annotation, model training, application deployment, and upgrade and iteration.

Moreover, as the world model's expression in the real - space continues to strengthen, more scenarios with explosive potential will be unlocked. Among them, the directions of autonomous driving and content creation are particularly important, and Jijia Vision will naturally not miss them.

As early as September 2023, Jijia Vision first released DriveDreamer in the field of autonomous driving, which is the world's first world model for the physical world. Later, it successively released the DriveDreamer series and the ReconDreamer series. So far, Jijia Vision's new - generation driving simulator centered around the world model has served more than 30 leading automotive OEMs and autonomous driving companies at home and abroad.

As is well - known, current video generation models in the industry mostly focus on visual effects and have significant limitations in 3D consistency and physical accuracy. For this reason, Jijia Vision has reached cooperation agreements with multiple leading customers in the film, television, and game industries and is developing multiple world models for the content direction, aiming to bring about essential breakthroughs.

It is reported that Jijia Vision's Yisu Engine is a one - stop content creation engine that integrates the first domestic ultra - long - duration Sora - level video generation large model, the world's first real - time 3D world model framework, and the first 3D human body generation framework combining "generation + reconstruction".

It is worth mentioning that Jijia Vision is building the largest and highest - quality data base in the industry. This year, the high - quality visual - action data will exceed 1 million hours, and the pre - training data for the world model will exceed 10 million hours. The underlying logic is self - evident. Jijia Vision may be the first to reach the GPT - 3 moment of embodied intelligence and then sprint towards the peak of physical AGI.

Looking around, the historical wheels of the AI era are rolling forward. Although the path of technological breakthrough is always full of hardships and twists and turns, and no one can predict when the turning point will come.

But we know that it will come eventually.

This article is from the WeChat official account "Investment Community" (ID: pedaily2012), author: Liu Bo, published by 36Kr with authorization.