PsiBot Series A: Industrial Capital Funds Embodied AI

The capital behind China's robot boom is increasingly coming from the factory floor. Shanghai-based Lingchu Intelligence, the embodied AI startup that operates as PsiBot and specializes in dexterous manipulation, has closed a Series A round of more than $100 million, according to the company. The PsiBot Series A, reported this week, counts auto parts maker Tuopu Group, Chery-backed Ruicheng Fund and precision glass processor Lens Technology among its industrial backers, with existing shareholder Zhuhai Technology Industry Group continuing its stake.

The round's capital comes mainly from the industrial side, a structure that sets it apart from the venture-led raises dominating Chinese robotics this month. Tuopu Group supplies chassis and powertrain components to electric vehicle makers, Lens Technology manufactures glass and cover lenses for smartphones and cars, and Ruicheng Fund invests under Chery Holding. Also participating were 37 Interactive Entertainment, Wuhu Investment Group and Fosun Fortune.

The PsiBot Series A proceeds are earmarked for two areas: frontier research on embodied intelligence built from human operation data, and commercial deployment in logistics and advanced manufacturing, Lingchu said. On the research side the company is iterating Psi-W0, its action-conditioned world model; on the commercial side it is pushing Psi-SynRobot from pilot validation toward volume production.

PsiBot Series A: What the money buys

Lingchu's technical stack rests on a dual-system design. Psi-R2, the operation policy, handles real-time manipulation decisions, while Psi-W0 predicts how the environment will respond to an action before the robot commits to it. That division of labor explains why world-model research has become a funding magnet across the sector: a policy that can simulate outcomes plans better than one that learns only from trial and error. Psi-R2 currently leads the MolmoSpace global leaderboard for embodied manipulation, according to Lingchu, which cites the ranking as validation of its human-data approach.

Feeding that policy stack is SynGlove Air, a glove-based capture system that records human hand movements during manipulation tasks. Lingchu says the device cuts data-collection cost to roughly one-tenth of traditional teleoperation. The figure matters because demonstration data is the sector's scarce input: guiding a robot through thousands of pick-and-place repetitions by teleoperation is slow and labor-intensive, and whoever makes that data cheap controls how fast policies can be trained. Grasping objects of varying size, shape, material and weight remains one of the field's hardest problems, which is why manipulation-focused startups such as PsiBot and PokeBot have drawn investor attention that once went to locomotion.

Glove capture removes the robot from the loop at data-collection time, letting the same operator generate demonstrations far faster than a teleoperation setup allows. If the one-tenth cost figure holds at scale, Lingchu's pipeline carries a structural cost advantage over rivals that remain dependent on teleoperation, and that advantage compounds into faster iteration between model updates and production trials.

From warehouse pilot to factory production

Psi-SynRobot, the flagship product, has completed small-scale validation in customer warehouses and is moving from model to mass production, Lingchu said. The company has also shown a new dexterous actuation unit at a Suzhou AI expo, aimed at precision manipulation in flexible manufacturing. On the supply side, Lingchu draws on the Wuhu National Robot Industry Cluster, where it connects to more than 300 upstream and downstream supporting enterprises, and it takes deployment scenarios from Tuopu and Lens Technology across automotive manufacturing, precision processing and modern logistics.

The PsiBot Series A arrives in the middle of a financing and listing wave for Chinese robotics. Unitree Robotics went public on the STAR Market in August 2026, humanoid developer Zhiyuan Innovation has opened its IPO process, and the Ministry of Industry and Information Technology has begun soliciting opinions on a humanoid robot standards system, an early step toward common technical and safety norms. Capital is following: Noin Intelligence raised $74 million for its GLOW generative model line, PokeBot closed a nine-figure pre-A round dedicated to robot manipulation, which the company treats as the field's hardest open problem, Discover Robotics collected more than $100 million for its wheeled AlphaBot 2 logistics platform, and Infiforce added nearly $150 million across its Series A rounds for embodied world models and DataGrid infrastructure.

Policy is moving in the same direction. China's 15th Five-Year Plan (2026-2030) names embodied intelligence a priority, and Li Yan, a researcher at the State Council's Development Research Center, has described the field as shifting from fixed-program production tools toward agents capable of environmental perception, task decision-making and flexible collaboration, with expected applications in industrial manufacturing, agriculture and public services. ROBOTERA founder Chen Jianyu has said his company's robots already operate regularly in logistics and industrial settings across more than 10 cities. A common standards framework would make such deployments easier to compare and integrate, which is what makes the MIIT consultation relevant to buyers.

What distinguishes the PsiBot Series A from the rest of August's field is who is writing the checks. Tuopu and Lens bring capital plus working production floors, so PsiBot's models get a place to run and a customer pipeline at the same time. That structure reflects a particular bet: the value in embodied AI sits in the brain, the model layer that decides what a robot does, while hardware-focused rivals compete on the body and the actuators. PsiBot competes in the embodied-brain track, a category drawing capital alongside hardware builders such as PokeBot and Discover Robotics and model developers such as Noin and Infiforce. The nine-figure PsiBot Series A war chest covers the two costs that typically sink model-first startups: data acquisition and deployment engineering.

For manufacturers and logistics operators evaluating embodied AI, the numbers to watch are data cost and deployment status. A one-tenth-of-teleoperation data pipeline and a product already past warehouse pilot are concrete markers that manipulation robots are approaching a usable total cost of ownership for repetitive picking, packing and assembly work. Procurement teams should watch three things: the cost per demonstration hour as SynGlove Air scales, the production ramp of Psi-SynRobot, and whether the Wuhu cluster's 300-plus suppliers turn into serviceable deployment capacity. The open question is scale, since small-batch validation is not the same as full production reliability.

Why this matters

For decision-makers, the PsiBot Series A shows industrial capital treating embodied AI as an operations technology with a production roadmap. When manufacturers invest in the same companies they might buy robots from, they shorten the path from pilot to purchase and raise the bar for rivals to show floor-level results. Whether PsiBot's warehouse validations scale will decide if the embodied brain market converts this funding round into production volume.

Related Articles

• XPeng Robotics Funding Round: Humanoid Unit Valued at $6.3B in Record Raise
• Physical Intelligence Acquisition Rumors Draw a Denial as a $1B Series C Looms
• Stord Reaches $3B Valuation to Advance Physical Intelligence in Logistics

✔Human Verified

Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.