Skip to main content

Jia Yueting Unveils America's Cheapest Humanoid Robot, Compact Enough to Fit in a Backpack

Jia Yueting's FFAI company has released the FF Nova mini humanoid robot. The base unit is priced at $2,490 (approximately RMB 16,600), with the accompanying developer skill pack priced at $390.

This is also the first entry-level embodied AI humanoid robot in the US priced under $2,500, touted as a Physical AI laboratory that fits in a backpack.

FF Nova targets two main audiences: developers, geeks, and Physical AI KOLs; as well as educational institutions, research and teaching organizations, and students interested in Physical AI.

FF Nova is equipped with a multimodal sensor matrix, where LiDAR, wide-angle vision, depth perception, and a ring microphone array work in synergy.

With a total of 20 degrees of freedom across its body, it is a true bipedal humanoid robot capable of walking, turning, waving, and standing back up after being knocked over.

贾跃亭发布美国最便宜人形机器人 还能塞入背包

It features two movement modes: bipedal walking towards you during interaction, and autonomous navigation when mounted on a detachable chassis.

Nova also integrates large voice models, supporting natural conversation and voiceprint recognition, while opening up AI Agent access to become the physical embodiment of cloud-based intelligent agents.

Additionally, Nova natively supports the developer platform, providing a fully open SDK tailored for ROS 2 development workflows and supporting OTA continuous iteration. Nova will also integrate a complete Sim2Real reinforcement learning pipeline, accompanied by a Skills Store and developer community, allowing future creations to be distributed, reused, and monetized.

Currently, FF Nova has officially opened online applications and invite-only limited sales, initially targeting universities, research laboratories, educational institutions, independent developers, and education innovators.

FF Nova will be released in three phases: Phase one offers limited sales of 100 units to developers and educational institutions, with the first batch of 20 units delivered immediately. Phase two targets all developers and educational institutions; phase three will see the full consumer version release.