Humanoids
Daily
Published on

From Huawei 'Genius' to Robotics Entrepreneur: The Rise of Peng Zhihui and AgiBot

Authors
Peng Zhihui together with friends

From Huawei 'Genius' to Robotics Entrepreneur: The Rise of Peng Zhihui and AgiBot

In the rapidly evolving landscape of Chinese robotics, few figures have captured as much attention as Peng Zhihui. Known online as "Zhihui Jun" and nicknamed the "Wild Iron Man" by his followers, Peng transitioned from a high-profile role in Huawei's prestigious "Genius Youth" program to founding AgiBot (also known as Zhiyuan Robotics), a Shanghai-based startup aiming to become a leader in embodied intelligence and humanoid robots.

The 'Wild Iron Man'

Before AgiBot Peng Zhihui built a significant online following even before his corporate career took off. A graduate of the University of Electronic Science and Technology of China (UESTC) with a master's degree in information technology (2018), Peng gained popularity on platforms like Bilibili by showcasing complex DIY tech projects.

His creations, including self-balancing robots, a miniature TV, a robotic arm nicknamed "Dummy" (watch in the video above!) capable of stitching grape skins, and even a self-driving bicycle, demonstrated impressive engineering skills and attracted millions of views, establishing his reputation as a talented maker. After a stint at OPPO's AI Lab, Peng joined Huawei in 2020 through its highly selective "Genius Youth" program, designed to attract top global talent with high salaries – Peng reportedly earned up to 2 million yuan (approx. US$287,000) annually. At Huawei, he worked as an AI algorithm engineer in the computing product line, focusing on AI edge heterogeneous computing and contributing to projects like the Ascend AI platform. His viral Bilibili projects continued during his time at Huawei, sometimes incorporating the company's technology like HarmonyOS and 5G.

Launching AgiBot: Ambition in Embodied Intelligence

In December 2022, Peng announced his departure from Huawei, citing a desire to pursue his passion for robotics and tackle new challenges.

If programmers are considered gods in the digital world, then shaping robots with our own hands and infusing them with souls using AI is the ultimate romance for true geeks!

he stated, explaining his motivation. Just two months later, in February 2023, Shanghai Zhiyuan Innovation Technology Co., Ltd., operating as AgiBot, was founded. Peng, initially taking roles like Chief Innovation Officer and CTO, co-founded the company with other industry veterans, including Professor Yan Weixin from Shanghai Jiao Tong University. The company quickly established itself in Shanghai's Lingang New Area, a hub for tech innovation. AgiBot's mission is focused on the fusion of AI and robotics to create "world-class leading embodied intelligent robot products and application ecosystems."

AgiBot's Expanding Robot Family and Tech Stack

The company has rapidly developed several robot lines:

  • Yuanzheng Series (A1, A2): Full-sized bipedal humanoid robots designed for commercial applications like interactive services (showrooms, reception) and potentially industrial tasks. The A2 model stands 175cm tall and emphasizes ergonomic design for human-robot interaction.
  • Genie Series (Genie-1, A2-W): Dual-armed robots, often on wheeled bases, targeted at flexible manufacturing and logistics scenarios.
  • Lingxi Series (X1, X2): Smaller, more agile humanoids (X2 stands 1.3m tall) aimed at developers, consumers, and research. Peng Zhihui reportedly oversees development for this line. The recently demonstrated X2 showed capabilities like riding scooters and hoverboards.

Underpinning these robots is a focus on embodied AI. AgiBot has developed its own AI models, like the recently unveiled Genie Operator-1) (GO-1), a generalist foundation model using a Vision-Language-Latent-Action (ViLLA) architecture. This model aims to enable robots to learn from human videos and generalize tasks with minimal specific training. Data is crucial for this approach. AgiBot established a 2,000-square-meter data collection center in Shanghai where around 100 robots reportedly generate tens of thousands of motion data records daily. In late 2024, the company open-sourced its 'AgiBot World' dataset, claiming it contains over one million diverse manipulation data entries, surpassing existing datasets like Google's Open X-Embodiment in scale and real-world relevance for humanoid scenarios. AgiBot has also open-sourced designs for its Lingxi X1 robot and its AimRT robot middleware framework.

Rapid Scaling and Funding

AgiBot has attracted significant investment interest since its inception. It has completed multiple funding rounds (Angel through Series B), securing backing from prominent investors including Tencent, BYD, HongShan (formerly Sequoia China), Hillhouse Investment, Warburg Pincus, CDH Investments, Lanchi Ventures, and various government-backed funds. Recent reports suggest a Series B round led by Tencent, valuing the company at around 10 billion yuan (approx. US$1.4 billion), with talk of a new round potentially targeting a 15 billion yuan (approx. US$2.1 billion) valuation. The company has also moved quickly towards production. In late 2024, AgiBot announced it had begun mass production, reporting nearly 1,000 units produced by mid-December 2024 and hitting the 1,000-unit milestone by January 2025 (including both wheeled and bipedal models). For 2025, AgiBot executives have stated ambitious production targets of 3,000 to 5,000 robots, aiming to match projected output for Tesla's Optimus. A new plant is reportedly planned to boost capacity.

Peng Zhihui and AgiBot have emerged as key players in China's burgeoning robotics sector, particularly in the competitive humanoid field. Their rapid progress in product development, fundraising, and scaling production is noteworthy. However, challenges remain. The humanoid robotics industry is still nascent, with high production costs and the need to identify truly valuable real-world applications beyond demonstrations. AgiBot aims to deploy robots in industrial settings first, replacing humans in specific tasks, projecting household adoption is still roughly five years away. Achieving widespread adoption will likely require significant cost reductions, with executives citing a target price near 50,000 yuan (US$6,900) per unit. Furthermore, developing truly general-purpose intelligence requires vast amounts of high-quality, multimodal data, a challenge AgiBot acknowledges and attempts to address through its data factory and open-source initiatives. While comparisons to Tesla's Optimus generate buzz, AgiBot's strategy appears focused on iterative commercialization in targeted sectors to fund ongoing R&D, potentially a more pragmatic approach for a startup. Peng Zhihui's journey from viral tech influencer and Huawei 'genius' to robotics entrepreneur exemplifies the dynamism and ambition within China's tech ecosystem. AgiBot's trajectory will be closely watched as a barometer for the humanoid robotics industry's transition from research labs and hype cycles to tangible commercial value.