MindOn

MindOn

"One Model. Multiple Robots. Human Efficiency."

Embodied AI startup founded 2025 by an ex-Tencent Robotics X researcher; its Mind-0 and Mind-1 models run humanoid and dual-arm robots.

Visit Website

Company Overview

MindOn is an embodied AI company developing a single foundation model intended to control many kinds of robots. Rather than building its own robot bodies, it trains models on large-scale human-centric data and deploys them on third-party humanoid and dual-arm platforms; its earliest public demonstrations used Unitree humanoid hardware. Dealroom and KrASIA describe it as a Chinese startup, and all roles on its careers page are listed in Singapore; the company does not publish a headquarters address.

The company was started in 2025 by Zhu Qingxu, a researcher who spent about four years at Tencent Robotics X after graduating from a joint ETH Zurich and RWTH Aachen programme in 2021. Its technical approach avoids teleoperation as the main data source: it captures human demonstrations with optical motion capture and a universal manipulation interface (UMI) system, and splits the model into a motion-control layer, described as the cerebellum, and a planning and generalisation layer, described as the cortex.

Public milestones follow a clear sequence. In November 2025 it released a demonstration of a humanoid robot watering plants, opening curtains, wiping surfaces and disposing of rubbish in a home. In June 2026 it introduced Mind-0, a unified model for mobile manipulation, with a logistics demo in which two humanoid robots retrieved items from shelves and carried them to a workstation where two dual-arm robots sorted and packed them into parcels; the company said all four robots ran the same model trained only on human-centric data, without task-level robot-collected training data. On 30 September 2026 it released Mind-1, a model aimed at human-level cycle times, with demos of dual-arm robots handling parcels and folding shirts and humanoids collecting laundry, labelled fully autonomous at 1.0x speed. The company argues that teleoperation data is slower than natural human motion and that faster execution needs changes to training data and control, not just higher motor speeds. It published no quantitative comparison with human workers.

Funding came quickly. KrASIA reported close to CNY 100 million raised across three early rounds within the first four months, from Oriza Seed, MSA Capital, Innoangel Fund, Yuanbio Venture Capital, Genspark, Plum Ventures and G&O. On 30 June 2026 MindOn announced a Series Pre-A round backed by Gaorong Ventures, Granite Asia, Prosperity7 Ventures (the growth fund of Aramco Ventures) and existing shareholder MSA; the amount was not disclosed. The company said the money would scale its robot data engine, expand model capabilities and move toward deployment in healthcare, construction, warehousing and manufacturing.

Capabilities & Activities

Primary type & automation activities this supplier delivers:

Applications & Industries

Product Categories

Products & Solutions

Mind-1

Physical AI model targeting human-level cycle times for manipulation on dual-arm and humanoid robots.

Mind-0

Unified mobile-manipulation model trained on human-centric data and shared across humanoid and dual-arm robots.

Partnership & Notable clients

Associating Events

Listed in RobotToday Supplier Discovery. Verified Profile

More related news

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
    ROSCon Korea 2027: Feb 17–19 at DGIST, Daegu RobotToday  · Oct 05, 2026
  10. 10
  11. 11
  12. 12
  13. 13
    RobotToday Weekly September 21 – 25, 2026 RobotToday  · Sep 26, 2026
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20

Contact MindOn

WEBSITE

https://www.mindon.tech

EMAIL

[email protected]

Company Facts

Founded

2025

Primary Role

Software/Algorithm

Company Size

-

Primary Region

-

Annual Sales

-

Funding Stage

Early-Stage VC

Funding Total

-