Robbyant Debuts LingBot Models for Robot Spatial Perception

A New Entrant in the Robot Foundation Model Race

Chinese embodied AI company Robbyant has introduced LingBot-Depth 2.0 and LingBot-Vision, a pair of AI models aimed at sharpening how robots perceive depth, navigate space, and interpret their visual surroundings 1. The launch adds Robbyant to a rapidly crowding field of firms racing to build the perceptual and cognitive "brains" that let humanoid and industrial robots operate reliably outside tightly controlled factory settings. While details on specific benchmarks remain limited, the framing of these releases as next-generation spatial perception tools signals a broader industry push toward robots that can understand three-dimensional environments with the kind of nuance previously reserved for specialized computer-vision research 1.

A Global Scramble for Robot Foundation Models

Robbyant's announcement lands amid a wave of similar moves worldwide. In the West, Agile Robots has partnered with Google DeepMind to embed DeepMind's robotics foundation models into its machines, a deal that also gives DeepMind access to real-world operational data to refine its own systems 2. That arrangement underscores a pattern taking shape across the industry: robotics companies gain sophisticated perception and reasoning models, while AI labs gain the physical-world data needed to train them further, a symbiotic exchange that is quickly becoming standard practice.

Japan, meanwhile, is treating the technology as a matter of national strategy. Noetra, a government-backed venture building a foundational model for physical AI and robotics, has been described by its CEO as the country's "last chance" to secure a competitive foothold in the domestic supply chain for critical robotics technology 4. That framing reflects growing anxiety that Japan, historically a robotics powerhouse, risks falling behind China and the United States in the software and AI layers now driving the industry's value.

Hardware and Ethics Enter the Conversation

The race extends beyond software into silicon and, more unsettlingly, into weapons policy. Foundation, the company behind the MK-2 Phantom humanoid, is turning to AMD Ryzen chips to power its robots, positioning AMD as a direct challenger to Nvidia's dominant grip on AI chip supply for robotics 5. That same company has also drawn scrutiny for a more contentious reason: its 2024 launch marked a notable break from an industry norm in which leading robotics firms had pledged not to develop weaponized autonomous systems, fueling debate over armed robots and the erosion of human sign-off in lethal decision-making 3.

Why It Matters

Taken together, these developments show an embodied AI sector advancing on multiple fronts at once — perception models like Robbyant's, foundation-model partnerships like DeepMind's with Agile Robots, national industrial strategy in Japan, chip competition challenging Nvidia, and unresolved ethical questions around autonomous weapons. As spatial perception and reasoning capabilities mature, the competitive and policy stakes surrounding humanoid and physical AI systems are only intensifying.