AIToday
RoboticsThe Robot ReportPublished: Nov 14, 2026, 04:00 JST

AGIBOT G2 robots pass 4 of 6 in 1,730-task store trial

Send on LINE
AGIBOT G2 robots pass 4 of 6 in 1,730-task store trial

3 Key Points

  1. What happened

    AGIBOT's six G2 wheeled humanoid robots ran a 24-hour shift at a live Shanghai convenience store in Xuhui District on November 2–3. They logged 1,730 tasks, including 683 customer-service interactions and 147 coffee orders, and four passed the final assessment.

  2. Why it matters

    A majority of the robots clearing the assessment suggests the hardware can handle a full day of changing store conditions, not just a controlled demo; the two that missed the criteria show reliability is still uneven.

WHO IT HITSRetail operators evaluating store automation and robotics teams building embodied AI now have a public benchmark of how a six-robot crew performs over a full day in a live store.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

AGIBOT framed the convenience store trial as the conclusion of its AI WAVE technology showcase, which launched on October 22 and spanned releases in robotics data, simulation, AI models, hardware and operational software. The store trial was the point where those pieces came together in a customer-facing setting, rather than in a lab.

The trial also tested AGIBOT's Genie Master embodied AI agent framework, which coordinates the six robots and routes execution results back for analysis. According to AGIBOT, successful and unsuccessful attempts are retained in long-term memory and effective adjustments can be converted into training data, so the robots are meant to build on prior experience instead of treating each task in isolation. AGIBOT says it has been using operational feedback, including practice during quieter overnight periods, to improve G2 performance on precise manipulation.

The company plans to keep testing across additional roles and environments and to assess which capabilities carry over between tasks and settings. AGIBOT also said in September 2026 that its 20,000th robot had rolled off the production line.

FAQ
What tasks did the robots actually do?
They covered customer service (683 interactions), product pickups and deliveries (369), coffee orders (147), breakfast servings (126), shelf inspections and restocking (277), and table cleaning (128).
How were the robots assessed?
Evaluators scored time on duty, task completion, customer feedback, response times and handling of unexpected situations. Four passed; two did not meet all criteria.
Was anyone watching?
Yes. The trial was livestreamed to more than two million viewers across online platforms.
The Robot ReportRead Original Article
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleOpenAI disbands stand-alone robotics team