Realset AI

Summary

About

About Realset AI

Realset AI captures demonstrations of skilled real-world work for training embodied agents. Its collection workflow uses expert operators, egocentric and bimanual capture, and teleoperation to record how physical tasks are performed rather than relying only on internet-scale video or simulated trajectories.

The resulting data is organized around actions and outcomes so model developers can train and evaluate systems against complete task sequences. Realset works across data collection, action-level annotation and evaluation infrastructure, positioning the company as a training-data and experience layer for physical AI rather than a robot manufacturer.

PRICE
Range 1Y
Return
High
Low
Updated
DATA: KORTHOS

Articles

27 SEPT 2026

Robotics capital moves — 21–27 September 2026

Team

  • Hunter Guo - Founder - Directs the company's expert-capture, annotation and evaluation approach for physical-AI training data.

Timeline

Events

  • 2026-09-23 - Funding - Realset AI and Flatkey announce joint USD10 million Series A

    Realset plans to expand its network of real workplaces and studio environments for expert data capture. Its embodied-data work includes egocentric and bimanual teleoperation demonstrations from homes, warehouses and light assembly lines. The company plans an open household-manipulation benchmark for Q4 2026, including evaluation of open-source vision-language-action policies.

    Source

Context

Comparable profiles

  • Vision Lab

    stack adjacency - Embodied Data Capture · Physical AI Training Data

  • LabyrinthAI

    stack adjacency - Physical AI Training Data

  • MANUS

    stack adjacency - Embodied Data Capture

  • Rokoko

    stack adjacency - Embodied Data Capture

  • StretchSense

    stack adjacency - Embodied Data Capture

  • Maniformer

    stack adjacency - Embodied Data Capture

  • Jiiov / 极豪科技

    stack adjacency - Embodied Data Capture

  • Physen AI

    stack adjacency - Embodied Data Capture

  • MiraiBit

    stack adjacency - Physical AI Training Data

  • XDOF

    stack adjacency - Embodied Data Capture

  • Noitom

    stack adjacency - Embodied Data Capture

  • Ropedia

    stack adjacency - Embodied Data Capture · Physical AI Training Data

Shared workflows

  • Lightwheel

    workflow adjacency - Robot Learning Data Collection

  • Mind Robotics

    workflow adjacency - Robot Learning Data Collection

  • Wuwen Zhike

    workflow adjacency - Robot Learning Data Collection

  • TacnIQ.ai

    workflow adjacency - Robot Learning Data Collection

  • PHYMI

    workflow adjacency - Robot Learning Data Collection

  • Reward AI

    workflow adjacency - Robot Learning Data Collection

Context tags

  • Physical AI training data - Stack Layer - https://realset.ai/ — exact-company evidence reviewed during today-new-supply-profile-qa-2026-09-28-run18.
  • Embodied data capture - Stack Layer - https://realset.ai/ — exact-company evidence reviewed during today-new-supply-profile-qa-2026-09-28-run18.
  • Robot learning data collection - Workflow - https://realset.ai/ — exact-company evidence reviewed during today-new-supply-profile-qa-2026-09-28-run18.