Automakers are making a strategic pivot away from building robot brains in-house, instead investing in outside artificial intelligence platforms that can power multiple types of machines. Chery Automobile, a Chinese carmaker that recently joined the Fortune Global 500, is leading a nearly $100 million funding round for Psibot, a Beijing-based robotics startup, at a valuation of $1.48 billion. This move reflects a broader industry recognition that the most defensible position in embodied AI, the field of robots that can perceive and act in the physical world, may not be the robot body itself but the software intelligence powering it.

What Makes Psibot's Approach Different From Other Robotics Companies?

Founded in 2024, Psibot targets a less glamorous but more commercially viable corner of the robotics market: logistics automation and precise object manipulation, rather than the humanoid robots that dominate headlines and trade shows. The company's core product is the Psi R1, a Vision-Language-Action (VLA) model, which is a type of artificial intelligence that takes visual observations of a robot's environment and natural language instructions as input, then directly outputs the low-level motor commands needed to move the robot's joints and end-effectors.

This architecture eliminates the need for separate perception, planning, and control systems that traditional industrial robots require. Instead of being programmed with specific instructions for each task, a VLA-based robot can generalize across novel objects, different lighting conditions, and task variations without requiring reprogramming. Psibot demonstrated this capability by having its Psi R1 robot play Mahjong autonomously for over 30 continuous minutes, a task the company classifies as "L3 level" dexterous manipulation, requiring the system to reason through long-horizon decisions rather than execute pre-programmed sequences.

Psibot's dual revenue model sets it apart from pure hardware robotics companies. Alongside its robotic hardware offerings, including a 21-degree-of-freedom dexterous hand, the company licenses the Psi R1 model to third-party robot developers and offers simulation and training data platforms. This platform-first strategy means that if the Psi R1 becomes the standard intelligence layer for logistics robots built by multiple manufacturers, the training data flywheel compounds over time regardless of which hardware form factor ultimately wins the market.

Why Are Automakers Suddenly Investing in Robotics Intelligence?

Chery's decision to lead this investment round is consistent with an aggressive pivot the Wuhu-based carmaker has been executing throughout 2026. In April, Chery announced a global strategic collaboration with Nvidia covering autonomous driving, cabin AI, and robotics. Chery's premium EXEED brand is deploying Nvidia's DRIVE Hyperion platform for Level 3 and Level 4 intelligent vehicle development, and the robotics dimension of that partnership explicitly extends to Isaac Sim simulation infrastructure.

The strategic logic is straightforward for an automaker pivoting to physical AI: Chery's existing supply chains already contain motors, sensors, batteries, and structural components that translate directly to robotics hardware. Its autonomous driving programs have generated years of perception and decision-making software that robotics intelligence systems share. By investing in an outside VLA platform like Psibot, Chery's in-house AiMOGA Robotics subsidiary can access advanced intelligence models without building them from scratch.

What Technical Challenges Still Limit Chinese Robotics Companies?

Despite rapid progress, Chinese embodied AI companies face three significant performance gaps that constrain their ambitions:

• Dexterity Bottleneck: Adaptability to diverse use cases without advanced fine-tuning remains impossible at this stage, according to independent analysts. China's humanoids lack precision and dexterity and are mostly deployed in limited tasks and site-specific trials.
• Hardware Dependency: China's robotics sector remains deeply dependent on Nvidia's hardware and software infrastructure. Major Chinese robotics companies, including UBTech, Galbot, Unitree, AgiBot, and Engine AI, rely on Nvidia's Jetson compute modules and Isaac Lab reinforcement-learning framework.
• Component Supply Chain: High-precision components for sophisticated manipulation tasks continue to come largely from European and Japanese suppliers. Germany's Schaeffler and Japan's THK and NSK together provide approximately 90 percent of the special ball screws used for precision positioning in robotics.

The hardware dependency on Nvidia represents a structural risk if US export controls extend to edge computing chips, which could disrupt the entire Chinese robotics sector's ability to deploy advanced systems.

How Does Psibot's Training Data Strategy Reduce Costs?

A critical engineering constraint shapes the entire embodied AI sector: VLA models require multi-modal training data, including vision, language, touch, and spatial information, that cannot be scraped from the internet the way text data can. Psibot claims its portable crowdsourced data collection system can reduce training data costs to roughly one-tenth of traditional teleoperation methods, though this figure has not been independently verified.

This cost reduction is significant because training data collection has historically been one of the largest expenses in developing robotics systems. If Psibot can deliver on this claim, it would give the company a substantial competitive advantage in licensing its platform to other robot manufacturers, who could deploy advanced intelligence systems at a fraction of the current cost.

What Does This Mean for the Broader Autonomous Vehicle Industry?

Psibot's $1.48 billion valuation makes it at least the 22nd Chinese embodied AI company to cross the unicorn threshold in 2026 alone, according to Bloomberg reporting. This proliferation of well-funded robotics startups suggests that the industry is consolidating around the platform-as-a-service model rather than the traditional hardware-first approach. Automakers like Chery are positioning themselves to benefit from this shift by investing early in the platforms they believe will become industry standards.

The involvement of Lens Technology, a precision glass supplier that counts Apple and Tesla as customers, in Psibot's funding round underscores how deeply robotics and autonomous systems are becoming integrated into the broader technology supply chain. As automakers invest in robotics platforms alongside their autonomous driving programs, the convergence of these two fields accelerates, creating new opportunities for companies that can bridge hardware manufacturing with advanced AI software.