A "no acceleration, no remote control" robot video has unexpectedly stirred Silicon Valley elites. Recently, footage from Chinese startup MindOn Tech went viral globally, showcasing remarkable technical prowess while sparking a trans-Pacific authenticity debate.

The video features a humanoid robot from Shenzhen-based MindOn Tech—founded by Tencent alumni—performing fluid tasks like watering plants, tossing trash, and playing frisco with children using Unitree's G1 hardware. The company emphasized these were autonomous actions without post-production manipulation.

U.S. skepticism emerged when Figure CEO Brett Adcock publicly dismissed the demonstration as "open-loop playback RL control," alleging human intervention during object-grasping sequences. This echoed his prior accusations against UBTECH, which he claimed used CGI in promotional videos.

Supporters countered with multi-angle footage validating the demo's authenticity. Tech analyst Mike Kalil defended MindOn's achievements, noting their integration of peer-reviewed research (ResMimic, HDMI, OmniRetarget) through imitation and reinforcement learning—training robots in simulations before real-world deployment.

The debate reveals deeper industry fissures: - U.S. leaders like Figure and Tesla pursue Apple-style vertical integration (proprietary hardware/software), yielding premium-priced products. - MindOn's "Android approach" decouples AI software from affordable mass-produced hardware (e.g., Unitree), potentially disrupting the market by shifting competition from hardware superiority to AI intelligence.

Analysts suggest successful execution of this model could democratize humanoid robotics, mirroring smartphone ecosystems where open platforms challenge integrated systems. The controversy signals an impending inflection point—China's rapid advancements in both hardware manufacturing and software intelligence are reshaping this global race at an unforeseen pace.