aifollow.news
返回 IT之家 科技新闻
IT之家 科技新闻· · 原发布时间 AI 评分40

小鹏预告明年二季度面向外部商业客户发布 IRON 人形机器人

AI 辅助摘要

何小鹏在小鹏集团财报电话会上表示,首批量产的 IRON 机器人计划明年一季度进入小鹏部分门店试用,二季度面向外部商业客户正式发布。最新版本还将在小鹏科技日和广州车展公开亮相;其多人识别与身份记忆等交互能力目前来自何小鹏的介绍。

正文 · AI 翻译

IT Home reported on November 16 that, according to Blue Whale News, at XPeng Group’s Q3 2026 earnings call this evening, its chairman and CEO He Xiaopeng said the latest version of XPeng IRON would make a public appearance at the upcoming XPeng Technology Day and Guangzhou Auto Show, where it would demonstrate a range of distinctive capabilities.

He Xiaopeng revealed that the first mass-produced robots will enter some XPeng stores for trial use in the first quarter of next year; in the second quarter of next year, XPeng will officially launch XPeng IRON for external commercial customers. Asked about XPeng’s competitive moat in this field, He said the most important things were to build capabilities including the hardware itself, commercial deployment at scale, manufacturing and sales, and high-quality, diverse data.

He Xiaopeng said advanced intelligence is the XPeng robot’s greatest advantage. IRON supports recognition of multiple people and memory of their identities; when interacting with people, it observes and thinks in real time and responds promptly to their needs.

He Xiaopeng also said that, at the upcoming XPeng AI Technology Day, the company plans to share its technical thinking on cars, robots and the physical AI technology paradigm, as well as some results from putting those ideas into practice.

According to IT Home, Iron is an AI humanoid robot independently developed by XPeng Motors and was first unveiled at XPeng AI Technology Day on November 6, 2024. The robot has a 1:1 human biomimetic design, stands 178 centimeters tall, weighs 70 kilograms, has 82 active degrees of freedom throughout its body, and has 22 degrees of freedom in its hands, which use what the company describes as the industry’s smallest harmonic joints. It is equipped with an in-house-developed Turing AI chip and a 720° Eagle Eye vision system, and uses an end-to-end large model and reinforcement learning algorithms to walk on two legs and manipulate complex objects.

发现内容有误?提交纠错