Editor’s note: Meta has tapped Shanghai-based Fourier to help build “Max,” a five-foot-tall humanoid developed in about three months, with a second version already underway. The deal puts a spotlight on a startup company that has grown from rehabilitation robotics into humanoids, ranked sixth globally in 2025 humanoid-robot shipments by Omdia, and attracted investors including SoftBank Vision Fund 2, Aramco-backed Prosperity7 Ventures and IDG Capital.
A Body for Meta’s AI
In late August, Meta demonstrated Max to employees at an internal robotics summit. The 1.5-meter, 48-kilogram humanoid, developed with Shanghai-based Fourier, ran on Nvidia’s Jetson Thor, displayed facial expressions and grasped objects with five-fingered hands.
The two companies built it in about three months, with Meta robotics product vice president Li-Chen Miller spending nearly two months at Fourier’s Shanghai factory.
The companies began work on Max 2 in September. Meta’s efforts with Fourier are for research and experimentation and are not currently intended for sale. Later, Fourier told The Paper that further information would be released through official channels.
The collaboration may suggest that Fourier brought more than a finished robot: an established hardware platform, engineers capable of rapidly adapting it and access to a mature robotics supply chain.
From Rehab Wards to Humanoids
Founded by Gu Jie in Shanghai in 2015, Fourier began with rehabilitation robots, including exoskeletons and training systems for hospitals. The business gave it years of experience in motion control, force feedback and human-machine interaction.
Its move into humanoids came gradually. Fourier began its humanoid robotics project in 2019 and achieved a key milestone in 2022, when an early prototype stood and walked untethered.
It unveiled the full-size GR-1 in 2023, followed by GR-2 in 2024. In 2025, Fourier launched the open-source N1 and GR-3, a 55-degree-of-freedom “Care-bot” built around visual, audio and tactile interaction.
That background gives Fourier a broader focus. Alongside industrial and logistics applications, it has also targeted rehabilitation, care and public-service settings where robots need to understand gestures, voices and physical contact.
Why Fourier Fit Meta
Fourier’s capabilities closely matched what Meta wanted from Max.
According to The Information, Meta wanted the robot to recognize voices from five meters away, interpret gestures from three meters, maintain eye contact and signal states such as listening or uncertainty.
And Fourier’s experience in rehabilitation and multimodal human-robot interaction offered a relevant starting point.
Its broader ecosystem also helped. In 2024, Nvidia named Fourier among humanoid developers working with Project GR00T, and researchers later demonstrated household manipulation on its GR-1.
Speed and supply chains may have mattered too. Meta said the industry faces difficulties sourcing some components domestically in the U.S., while Fourier built the customized prototype in roughly three months within China’s mature robotics ecosystem.
Capital and the Next Test
Investors noticed Fourier long before Meta did. IDG Capital backed the company early, while SoftBank Vision Fund 2 led a RMB 400 million Series D in 2022, joined by Aramco-backed Prosperity7 Ventures and Vision Plus Capital. Prosperity7 said the deal set a record for the largest single financing round in Fourier’s field at the time.
In January 2025, Fourier’s Series E financing had reached nearly RMB 800 million, with the funding aimed at advancing humanoid robots and expanding commercial applications. Later that year, Runyang Technology invested RMB 300 million in an E3 round at a pre-money valuation of RMB 8 billion.
Fourier has also begun to show scale. Omdia estimated it shipped 300 humanoid robots in 2025, ranking sixth globally.
The Meta project is a high-profile validation. What comes next will show whether Fourier can turn engineering edge into lasting commercial scale.
DragonLab cuts through the noise to find the Chinese startups, founders, and technologies that could shape what comes next.
Have a startup we should know about? Reach us at dragonlab@technode.com.
