AgiBot 智元机器人

Humanoid-robot maker shipping fast, open GO-1 model, HK IPO pending

Other Enterprise Open Source
Researched · Published · Reviewed
RECATOOLS Score
7 / 10
Capability
8
Value for money
6
Ease of use
5
ASEAN readiness
4
API quality
5
Founded
HQ
Users
Launched
Developer

Overview

AgiBot (Shanghai; 智元机器人) builds humanoid robots powered by its GO-1 embodied-AI model, ships hardware via its Genie Studio dev platform, and open-sources the AgiBot World manipulation dataset. It shipped its 10,000th humanoid by March 2026 and targets a Hong Kong IPO in Q3 2026.

Advertisement

What you can produce with AgiBot 智元机器人

  • GO-1 embodied foundation model (1M+ trajectories, 217 tasks)
  • Genie Studio embodied-AI development platform
  • Open-sourced AgiBot World bimanual-manipulation dataset
  • Humanoid robot lineup (G1, A2, X2, Expedition A3) -- 10,000th unit shipped by March 2026
  • Genie Sim simulation platform for robot training
  • Public GitHub org (AgibotTech) with inference/training code for AgiBot X1
  • Hong Kong IPO targeted Q3 2026
Advertisement

ASEAN Perspective

AgiBot 智元机器人 in Southeast Asia

ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).

RECATOOLS Verdict

AgiBot is one of two companies (alongside Unitree) driving China's humanoid-robot shipment numbers, and it backs the hardware with real software credibility: the GO-1 foundation model, the Genie Studio development platform, and -- genuinely useful for outside researchers -- the openly released AgiBot World manipulation dataset and its active GitHub org (AgibotTech), whose repos carry over a thousand stars apiece. Hitting a 10,000th mass-produced unit by March 2026 is a legitimate industry-first pace.

Worth tempering the hype, though: independent write-ups have clocked GO-1's real-world task completion around 46% outside curated demo conditions, and some of the more impressive livestream footage reportedly leans on undisclosed teleoperation rather than full autonomy. Support outside China runs through local partners rather than a dedicated service team. Treat it as a genuine research and manufacturing leader, not yet a plug-and-play deployment for buyers outside China.

Independent AI-assisted assessment by RECATOOLS.

What people say

AgiBot (formally AGIBOT Innovation, founded in Shanghai in 2023 by ex-Huawei engineers Deng Taihua and Peng Zhihui) shipped its 10,000th mass-produced humanoid robot by March 2026 -- the fastest ramp the industry has seen, according to coverage tracking the milestone. Analyst firm Omdia credited it with roughly 39% of global humanoid shipments in 2025, more than 5,100 units, putting it alongside Unitree as one of China's two dominant humanoid makers. The company is now reportedly targeting a Hong Kong IPO in Q3 2026 at a HK$40-50 billion (roughly US$5.1-6.4 billion) valuation, with CICC, CITIC Securities and Morgan Stanley as joint sponsors, aiming to raise over US$1 billion by selling 15-25% of its shares.

The software side backs up the hardware: GO-1, its embodied foundation model, trained on more than a million robot-interaction trajectories across 217 tasks, and the company has open-sourced the AgiBot World dataset -- a large-scale bimanual-manipulation dataset that was an IROS 2025 Best Paper Award finalist -- plus code through its AgibotTech GitHub org, where core repos (agibot_x1_infer, agibot_x1_train, genie_sim) each carry well over a thousand stars.

Independent assessments are more skeptical of autonomy claims than the shipment numbers suggest. One write-up pegs GO-1's real-world task-completion rate at 46% outside controlled test setups, and notes that some widely circulated demo footage involves undisclosed teleoperation -- a human quietly steering parts of the task that gets presented as autonomous. Deployments so far cluster in controlled environments like warehouses and inspection routes, where lighting and object placement can be tightly managed; genuinely unstructured settings remain a harder problem for the whole industry, not just AgiBot. For buyers outside China, AgiBot coordinates support through local partners rather than running its own service organization, which reviewers flag as a real gap for mission-critical deployments compared to competitors like Boston Dynamics or Agility Robotics.

Summary of public user & expert reviews, compiled by RECATOOLS.

About this listing

Researched on
Published on
Last reviewed

This entry was compiled from publicly available data including AgiBot 智元机器人's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with AgiBot 智元机器人 unless explicitly stated.

Data accuracy

Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.

For the latest details, please refer to AgiBot 智元机器人 directly →

Spotted something out of date? Suggest an update →

Advertisement