MindOn

Summary

About

About MindOn

MindOn develops a shared intelligence layer intended to transfer across different robot bodies. Its Mind-0 work combines human-centred demonstration data, a cross-embodiment data pipeline and a whole-body action model, with lower-level controllers handling the dynamics of each physical platform. A separate compensation model is used to reduce real-world execution errors, while hierarchical coordination links high-level task reasoning to lower-level skills.

The company has demonstrated one model coordinating humanoid and fixed dual-arm robots in a logistics workflow covering retrieval, transport, sorting and packing. Those demonstrations establish capability under the shown configurations, not general deployment across every environment. MindOn says it is extending the same architecture to additional embodiments and scaling its data engine after a June 2026 Series Pre-A round. The company supplies the model and data stack; robot hardware, safety case and site integration remain the responsibility of each deployment configuration.

PRICE
Range 1Y
Return
High
Low
Updated
DATA: KORTHOS

Articles

04 JUL 2026

Robotics capital moves — 29 June 2026–5 July 2026

Funding

Investors

  • Binfu Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • Bohua Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • CCTV Rongmedia Industry Investment Fund - Participant - VC - 1 funding event - first seen 2026-03-10
  • Damai Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • Gaorong Ventures - Participant - VC - 1 funding event - first seen 2026-06-30
  • Granite Asia - Participant - Angel / Individual - 1 funding event - first seen 2026-06-30
  • Guokai Financial - Participant - Angel / Individual - 1 funding event - first seen 2026-03-10
  • Guozhong Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • Hillhouse Ventures - Lead - VC - 1 funding event - first seen 2024-11-13
  • Junshan Investment - Participant - Angel / Individual - 1 funding event - first seen 2026-03-10
  • Lanchi Ventures - Lead - VC - 1 funding event - first seen 2024-11-13
  • MSA - Participant - Unknown - 1 funding event - first seen 2026-06-30
  • Prosperity7 Ventures - Participant - VC - 1 funding event - first seen 2026-06-30
  • Pufeng Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • Taiming Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • Wofu Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • Wuxi Venture Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • Xuhui Capital - Lead - VC - 1 funding event - first seen 2026-03-10
  • Yangtze Optical Fibre affiliated fund - Participant - VC - 1 funding event - first seen 2026-03-10
  • Yanyuan Ventures - Participant - VC - 1 funding event - first seen 2026-03-10
  • YuanSheng Capital - Participant - VC - 1 funding event - first seen 2026-03-10
  • Zhuhai Technology Industry Group - Participant - Unknown - 1 funding event - first seen 2026-03-10

Team

  • Zhu Qingxu - Founder
  • Zhou Qinqin - Co-founder

Timeline

Events

  • 2026-10-12 - Demonstration - MindOn reports 142 garments per hour in extended Mind-1 folding demo

    The upload lasts 59 minutes and 22 seconds, with a displayed folding counter reaching 142. A person supplies each new garment to a fixed dual-arm station with grippers. The footage is labelled normal-speed autonomous folding; garment selection and fold-quality criteria are not standardized.

    Source
  • 2026-06-30 - Funding - MindOn completes Series Pre-A financing

    The official announcement disclosed no round amount. MindOn said the funding supports robot data-engine scaling, model capability expansion and real-world deployment in healthcare, construction, warehousing and manufacturing. The company framed its approach around a unified robot brain that starts with humanoids and transfers across other robot forms including dual-arm robots.

    Source
  • 2026-06-18 - Capability - MindOn demonstrates one model operating across humanoid and dual-arm robots

    Two humanoid robots retrieve and transport objects between shelves and workstations in the demonstrated workflow. Two dual-arm robots sort, pack and complete the parcel workflow at fixed workstations. MindOn says the model is trained from human-centric data rather than task-level robot-collected training data.

    Source
  • 2026-03-29 - Award - MindOn wins EAI-100 recognition and two embodied AI awards

    MindOn said it was selected to the EAI-100 embodied AI annual list on the strength of its dexterous-manipulation algorithms and VLA model work. The same event also recognized co-founder Chen Yuanpei in the EAI Pioneer 20 list and awarded the company's mahjong robot demo as one of the annual top ten demos. The post ties the recognition to MindOn's software-toolchain strategy, joint lab work, and data-collection / long-horizon execution positioning rather than a hardware product launch.

    Source
  • 2026-03-10 - Funding - MindOn completes RMB 2 billion across angel and Pre-A funding rounds

    MindOn said it completed angel and Pre-A financing totaling RMB 2 billion, with state-backed and industrial investors participating across both rounds. The company said the capital will be used for scaled deployment in logistics scenarios and for building a large-scale embodied-data collection solution system. The announcement also framed MindOn around end-to-end VLA models, its Psi-SynEngine embodied data engine, and commercial validation in real logistics environments.

    Source
  • 2026-01-15 - Award - MindOn recognized at the 2025 Entrepreneur 100 Future Unicorn conference

    MindOn said it was named to the 2025 Entrepreneur 100 Future Unicorn list at the 18th Entrepreneur annual conference. Founder and CEO Wang Qibin also took part in the event's roundtable discussion, where the company positioned long-horizon dexterous manipulation and data-driven robotics as its core thesis. The writeup is materially company-level and strategic rather than a product release, so this fits as an award / recognition event rather than a launch.

    Source
  • 2025-11-14 - Demonstration - MindOn goes viral with Unitree G1 household-chore autonomy demo

    MindOn released a widely circulated demo showing a Unitree G1 handling household tasks including watering plants, opening curtains, collecting a delivery, wiping surfaces, sorting items, and picking up trash. The company's central claim in the demo was "no speed up, no teleoperation," framing the result as software-driven autonomy rather than human remote operation. The event matters as a public proof point for MindOn's software-first positioning: the intelligence layer was the story, while the robot body was Unitree G1 hardware.

    Source
  • 2025-06-16 - Demonstration - MindOn showcases embodied AI stack at NVIDIA startup exhibition during Volcano Engine FORCE 2025

    MindOn said it was invited to exhibit at the NVIDIA startup showcase held alongside Volcano Engine's 2025 FORCE conference in Beijing. The company used the event to present its VLA architecture, in-house dexterous hand, reinforcement-learning manipulation algorithms, and PsiBot system lineup to a large industry audience. The writeup also notes that MindOn participated in an open-mic session focused on the three pillars of general embodied intelligence, making this a clear market-facing demonstration event.

    Source
  • 2025-04-29 - Capability - MindOn expands Psi R1 with reinforcement-learning-based generalization upgrade

    Two days after the initial Psi R1 release, MindOn published a second post highlighting generalization gains from combining VLA with reinforcement learning. The company said the updated system demonstrated millimeter-level dexterous manipulation, rule understanding, interactive decision-making, more than 30 minutes of chained task execution, and multi-agent coordination in the mahjong setting. This reads as a capability milestone on top of the Psi R1 launch rather than a separate product family.

    Source
  • 2025-04-27 - Product Launch - MindOn launches Psi R1 for long-horizon complex embodied tasks

    MindOn introduced Psi R1 as a VLA model aimed at long-horizon complex task execution in open environments, using mahjong as its flagship demonstration scenario. The company framed the release around multi-step reasoning, dynamic strategy generation, fine-grained dexterous manipulation, and compound human-robot / robot-robot / robot-environment interaction. The announcement positions Psi R1 as a step beyond simple pick-and-place by targeting long-duration embodied planning and execution in a stochastic real-world task.

    Source
  • 2025-03-03 - Product Launch - MindOn releases Psi R0.5 end-to-end VLA model with lower data requirements

    MindOn announced Psi R0.5 as an enhanced end-to-end VLA model built on a reinforcement-learning-augmented hierarchical architecture. The company said the release improved generalization, dexterity, chain-of-thought behavior, and long-horizon task capability while cutting data requirements for generalized grasping training. The announcement also tied the launch to four disclosed papers covering grasp generalization, object retrieval in stacking scenarios, environment-assisted grasping, and VLA safety alignment.

    Source
  • 2025-02-28 - Demonstration - MindOn appears at the 2025 Global Developer Conference as the only invited embodied foundation-model company

    MindOn said it was the only invited embodied-model company at the 2025 Global Developer Conference. The post framed the appearance around a full stack from core algorithms to commercialization, with emphasis on dexterous operation, embodied intelligence models, and scenario-led deployment. This is best treated as a demonstration / market-positioning event rather than a product release because the page centers on conference visibility and technical positioning.

    Source
  • 2024-12-30 - Product Launch - MindOn launches Psi R0 as an end-to-end reinforcement-learning embodied model

    MindOn introduced Psi R0 as its first end-to-end embodied model based on reinforcement learning. The company said the model supports bimanual dexterous manipulation, multi-skill chaining, and scene- and object-level generalization for long-horizon tasks. The post explicitly positioned Psi R0 against the industry's fixation on basic pick-and-place generalization by claiming longer, more complex task closure.

    Source
  • 2024-11-13 - Funding - MindOn raises angel funding led by Hillhouse and Lanchi

    MindOn announced completion of its angel round, led by Hillhouse and Lanchi, before later disclosing the larger combined angel and Pre-A total in 2026. The company framed the round around reinforcement-learning-first embodied intelligence and long-horizon manipulation rather than chassis manufacturing.

    Source

Relations

Relationships

  • NVIDIA Inception

    Nvidia Startup Showcase Participant (force 2025) - MindOn appeared in NVIDIA's startup showcase at Volcano Engine FORCE 2025.

Context

Comparable profiles

Context tags

  • Autonomy software platform - Stack Layer - MindOn develops robot intelligence software for humanoids and other robot forms.
  • Embodied AI model layer - Stack Layer - MindOn describes a general-purpose intelligence system for robots and VLA/RL model work.
  • Dexterous manipulation - Capability - MindOn public materials emphasize long-horizon dexterous manipulation and household/service task execution.
  • Embodied data capture - Stack Layer - MindOn funding materials describe scaling a robot data engine and embodied-data collection.