<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title><![CDATA[Embodied Global - Humanoid Robots & Embodied Intelligence News]]></title>
    <link>https://embodiedglobal.com</link>
    <description><![CDATA[Stay ahead with daily embodied intelligence news. 200+ articles on humanoid robots, physical AI, robotics breakthroughs, and AI startup funding — in English, Spanish, and French.]]></description>
    <language>en</language>
    <lastBuildDate>Mon, 14 Sep 2026 05:52:37 GMT</lastBuildDate>
    <atom:link href="https://embodiedglobal.com/rss.xml" rel="self" type="application/rss+xml"/>
    <managingEditor>business@allinagi.com.cn (Embodied Global Team)</managingEditor>
    <webMaster>business@allinagi.com.cn (Embodied Global Team)</webMaster>
    
    
    <item>
      <title><![CDATA[TourPhysics: Bringing Physics to World Models for Exploration and Manipulation from a Single Image]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2609.04911-tourphysics-bringing-physics-to-world-models-for-exploration</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2609.04911-tourphysics-bringing-physics-to-world-models-for-exploration</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 08 Sep 2026 02:25:46 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[E-RGB-D: Real-Time Event-Based Perception with Structured Light]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2512.18429-e-rgb-d-real-time-event-based-perception-with-structured-lig</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2512.18429-e-rgb-d-real-time-event-based-perception-with-structured-lig</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 08 Sep 2026 02:25:46 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Linguistic Trajectory Encoding for Efficient Long-Horizon Spatial Memory in Embodied Agents]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2609.04802-linguistic-trajectory-encoding-for-efficient-long-horizon-sp</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2609.04802-linguistic-trajectory-encoding-for-efficient-long-horizon-sp</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 08 Sep 2026 02:25:46 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[CF-VLA: Efficient Coarse-to-Fine Action Generation for Vision-Language-Action Policies]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2604.24622-cf-vla-efficient-coarse-to-fine-action-generation-for-vision</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2604.24622-cf-vla-efficient-coarse-to-fine-action-generation-for-vision</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 08 Sep 2026 02:25:46 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Redwood: A Frontier AI Accelerator Designed, Verified, and Deployed from Scratch in 2 Weeks by AI]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.26418-redwood-a-frontier-ai-accelerator-designed-verified-and-depl</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.26418-redwood-a-frontier-ai-accelerator-designed-verified-and-depl</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 01 Sep 2026 02:23:39 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Describe-Then-Act: Proactive Agent Steering via Distilled Language-Action World Models]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2603.23149-describe-then-act-proactive-agent-steering-via-distilled-lan</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2603.23149-describe-then-act-proactive-agent-steering-via-distilled-lan</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 01 Sep 2026 02:23:39 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[GraspHOI: Full-Body 3D Human-Object Reconstruction with Finger-Level Grasps from a Single In-the-Wild Image]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.28386-grasphoi-full-body-3d-human-object-reconstruction-with-finge</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.28386-grasphoi-full-body-3d-human-object-reconstruction-with-finge</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 01 Sep 2026 02:23:39 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[uScenes: A Multimodal RGB and 3D Sonar Dataset for Underwater Robot Perception]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.27795-uscenes-a-multimodal-rgb-and-3d-sonar-dataset-for-underwater</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.27795-uscenes-a-multimodal-rgb-and-3d-sonar-dataset-for-underwater</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 01 Sep 2026 02:23:39 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Not for Backflips: Lumos Robotics' 50-kg MOS2 'AI Worker' Goes to Work on Mitsubishi's PLC Lines]]></title>
      <link>https://embodiedglobal.com/en/article/lumos-mos2-ai-worker-mitsubishi-plc-lines</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/lumos-mos2-ai-worker-mitsubishi-plc-lines</guid>
      <description><![CDATA[Lumos Robotics launches MOS2, a 50-kg-class dual-arm mobile manipulator positioned as an 'AI Worker' — already inspecting, screwdriving and hauling on Mitsubishi Electric's PLC lines after 1,000]]></description>
      <content:encoded><![CDATA[Lumos Robotics launches MOS2, a 50-kg-class dual-arm mobile manipulator positioned as an 'AI Worker' — already inspecting, screwdriving and hauling on Mitsubishi Electric's PLC lines after 1,000]]></content:encoded>
      <pubDate>Fri, 28 Aug 2026 07:58:06 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>Lumos Robotics</category>
      <category>MOS2</category>
      <category>mobile manipulator</category>
      <category>Mitsubishi Electric</category>
      <category>industrial robotics</category>
      <category>AI worker</category>
      <enclosure url="/images/industry/lumos-mos2-factory-floor-2026.jpeg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[RISE: Adaptive Imagination for World Action Models]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.20430-rise-adaptive-imagination-for-world-action-models</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.20430-rise-adaptive-imagination-for-world-action-models</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 25 Aug 2026 02:23:37 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[WorldLines: Benchmarking and Modeling Long-Horizon Stateful Embodied Agents]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2606.18847-worldlines-benchmarking-and-modeling-long-horizon-stateful-e</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2606.18847-worldlines-benchmarking-and-modeling-long-horizon-stateful-e</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 25 Aug 2026 02:23:37 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[From Simulation to the Real-World: An In-Field 6D Pose Dataset and Baseline for Robotic Strawberry Harvesting]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2606.11381-from-simulation-to-the-real-world-an-in-field-6d-pose-datase</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2606.11381-from-simulation-to-the-real-world-an-in-field-6d-pose-datase</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 25 Aug 2026 02:23:37 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[CertVLA: Certified Defense against Physical Visual Attacks for Vision-Language-Action Models]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.20791-certvla-certified-defense-against-physical-visual-attacks-fo</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.20791-certvla-certified-defense-against-physical-visual-attacks-fo</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 25 Aug 2026 02:23:36 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Beyond the 9.32s: Three Signals From the Second World Humanoid Robot Games]]></title>
      <link>https://embodiedglobal.com/en/article/world-humanoid-robot-games-2026-autonomy-dexterity-scenario-signals</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/world-humanoid-robot-games-2026-autonomy-dexterity-scenario-signals</guid>
      <description><![CDATA[Five humanoid robots beat Bolt's 100m at Beijing's second World Humanoid Robot Games — the mark now stands at 9.32s — but the real story is in the rulebook: mandatory autonomy, a debut dexterous-hand]]></description>
      <content:encoded><![CDATA[Five humanoid robots beat Bolt's 100m at Beijing's second World Humanoid Robot Games — the mark now stands at 9.32s — but the real story is in the rulebook: mandatory autonomy, a debut dexterous-hand]]></content:encoded>
      <pubDate>Sun, 23 Aug 2026 08:09:32 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>World Humanoid Robot Games</category>
      <category>humanoid robots</category>
      <category>full autonomy</category>
      <category>dexterous manipulation</category>
      <category>scenario applications</category>
      <category>embodied AI</category>
      <enclosure url="/images/hero/eg_hero_industry.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[WRC 2026 Midweek: A ¥6.8B Grid Tender and an Open-Source Wave Sweep Beijing]]></title>
      <link>https://embodiedglobal.com/en/article/wrc-2026-midweek-state-grid-tender-opensource-wave</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/wrc-2026-midweek-state-grid-tender-opensource-wave</guid>
      <description><![CDATA[Halfway through WRC 2026: State Grid moved a ¥6.8 billion embodied-AI procurement into supplier selection, central SOEs formed a robotics alliance, and developer day unleashed open-source robots and]]></description>
      <content:encoded><![CDATA[Halfway through WRC 2026: State Grid moved a ¥6.8 billion embodied-AI procurement into supplier selection, central SOEs formed a robotics alliance, and developer day unleashed open-source robots and]]></content:encoded>
      <pubDate>Sat, 22 Aug 2026 14:46:07 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>WRC 2026</category>
      <category>World Robot Conference</category>
      <category>State Grid</category>
      <category>procurement</category>
      <category>open source</category>
      <category>humanoid robots</category>
      <enclosure url="/images/hero/eg_hero_industry.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Humanoid Robot Runs 9.39s in Beijing, Inside Bolt's Human World Record]]></title>
      <link>https://embodiedglobal.com/en/article/humanoid-robot-games-tiangong-ultra-939-100m-bolt-record</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/humanoid-robot-games-tiangong-ultra-939-100m-bolt-record</guid>
      <description><![CDATA[At the 2nd World Humanoid Robot Games in Beijing, the Tiangong Ultra robot clocked 9.39s in 100m heats and jumped 2.88m — both beyond human world records — with 2,056 robots from 16 countries]]></description>
      <content:encoded><![CDATA[At the 2nd World Humanoid Robot Games in Beijing, the Tiangong Ultra robot clocked 9.39s in 100m heats and jumped 2.88m — both beyond human world records — with 2,056 robots from 16 countries]]></content:encoded>
      <pubDate>Sat, 22 Aug 2026 14:39:49 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Technology</category>
            <category>World Humanoid Robot Games</category>
      <category>Tiangong Ultra</category>
      <category>humanoid robots</category>
      <category>robot athletics</category>
      <category>Beijing</category>
      <category>embodied AI</category>
      <enclosure url="/images/hero/eg_hero_technology.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Unitree Unveils 7-Axis Dexterous Arm at ¥9,900, One Day After Its STAR Market Debut]]></title>
      <link>https://embodiedglobal.com/en/article/unitree-7-axis-dexterous-arm-9900-yuan-launch</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/unitree-7-axis-dexterous-arm-9900-yuan-launch</guid>
      <description><![CDATA[One day after its STAR Market debut, Unitree launched a biomimetic 7-axis dexterous arm starting at ¥9,900 (about $1,390), pushing capable manipulation hardware toward grant-budget pricing for labs]]></description>
      <content:encoded><![CDATA[One day after its STAR Market debut, Unitree launched a biomimetic 7-axis dexterous arm starting at ¥9,900 (about $1,390), pushing capable manipulation hardware toward grant-budget pricing for labs]]></content:encoded>
      <pubDate>Thu, 20 Aug 2026 04:01:44 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Product</category>
            <category>Unitree</category>
      <category>robotic arm</category>
      <category>dexterous manipulation</category>
      <category>product launch</category>
      <category>WRC 2026</category>
      <category>embodied AI</category>
      <enclosure url="/images/hero/eg_hero_product.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[WRC 2026 Opens in Beijing: 311 Product Debuts as Robot Industry Pivots From Stunts to Sales]]></title>
      <link>https://embodiedglobal.com/en/article/wrc-2026-beijing-opens-311-debuts-stunts-to-sales</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/wrc-2026-beijing-opens-311-debuts-stunts-to-sales</guid>
      <description><![CDATA[The 2026 World Robot Conference opened August 19 in Beijing E-Town with 373 exhibitors, 3,000 products and 311 new launches — plus a first-ever Procurement Day, 49-SOE pavilion and free robot access for global developers, as China's robot revenue tops ¥300B.]]></description>
      <content:encoded><![CDATA[The 2026 World Robot Conference opened August 19 in Beijing E-Town with 373 exhibitors, 3,000 products and 311 new launches — plus a first-ever Procurement Day, 49-SOE pavilion and free robot access for global developers, as China's robot revenue tops ¥300B.]]></content:encoded>
      <pubDate>Thu, 20 Aug 2026 02:42:19 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>WRC 2026</category>
      <category>World Robot Conference</category>
      <category>humanoid robots</category>
      <category>China robotics</category>
      <category>embodied AI</category>
      <category>Beijing E-Town</category>
      <enclosure url="/images/hero/eg_hero_industry.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[GhostPoint: Self-Supervised Representation Learning by Hallucinating Occluded LiDAR Structure]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.14428-ghostpoint-self-supervised-representation-learning-by-halluc</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.14428-ghostpoint-self-supervised-representation-learning-by-halluc</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 18 Aug 2026 02:27:39 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Zero-Shot Skeleton-Based Action Anticipation]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.14243-zero-shot-skeleton-based-action-anticipation</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.14243-zero-shot-skeleton-based-action-anticipation</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 18 Aug 2026 02:27:39 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Rollplex: Cross-Phase GPU Spatial Sharing for Vision Language Model Post-Training]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.14498-rollplex-cross-phase-gpu-spatial-sharing-for-vision-language</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.14498-rollplex-cross-phase-gpu-spatial-sharing-for-vision-language</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 18 Aug 2026 02:27:39 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[SSP: An Event-Matched Syn2Sim2Phy Cross-Domain Evaluation Framework for Autonomous Driving VLA Models]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.14024-ssp-an-event-matched-syn2sim2phy-cross-domain-evaluation-fra</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.14024-ssp-an-event-matched-syn2sim2phy-cross-domain-evaluation-fra</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 18 Aug 2026 02:27:39 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[8,288x Oversubscribed: Unitree's ¥61B STAR Listing Nears with DeepSeek at Its Side and a 'Superman' Teaser on Listing Eve]]></title>
      <link>https://embodiedglobal.com/en/article/unitree-688836-record-ipo-brain-body-era</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/unitree-688836-record-ipo-brain-body-era</guid>
      <description><![CDATA[Unitree has closed the most oversubscribed IPO in A-share history — ¥8.1 trillion in retail orders, a 0.0181% lottery rate — with STAR Market trading (ticker 688836) expected in the Aug 18–24 window. Beyond the frenzy, DeepSeek's ¥141M strategic stake with a 36-month lock-up signals the industry's shift from API vendor deals to brain-body equity integration.]]></description>
      <content:encoded><![CDATA[Unitree has closed the most oversubscribed IPO in A-share history — ¥8.1 trillion in retail orders, a 0.0181% lottery rate — with STAR Market trading (ticker 688836) expected in the Aug 18–24 window. Beyond the frenzy, DeepSeek's ¥141M strategic stake with a 36-month lock-up signals the industry's shift from API vendor deals to brain-body equity integration.]]></content:encoded>
      <pubDate>Mon, 17 Aug 2026 10:20:38 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>Unitree Robotics</category>
      <category>IPO</category>
      <category>STAR Market</category>
      <category>DeepSeek</category>
      <category>humanoid robots</category>
      <category>embodied AI</category>
      <category>688836</category>
      <category>brain-body integration</category>
      <enclosure url="/images/hero/eg_hero_industry.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[SparseVoxelDet: Fully Sparse Voxel Networks for Efficient Event-Based Drone Detection]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2603.21638-sparsevoxeldet-fully-sparse-voxel-networks-for-efficient-eve</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2603.21638-sparsevoxeldet-fully-sparse-voxel-networks-for-efficient-eve</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 11 Aug 2026 02:23:49 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Beyond Myopic World Models: Long-Horizon End-to-End Training for Direct Future Prediction]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2608.07420-beyond-myopic-world-models-long-horizon-end-to-end-training-</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2608.07420-beyond-myopic-world-models-long-horizon-end-to-end-training-</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 11 Aug 2026 02:23:49 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[SynthRender and I-AsSET: Open-Source Framework and Dataset for Bidirectional Sim-Real Transfer in Industrial Object Perception]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2602.21141-synthrender-and-i-asset-open-source-framework-and-dataset-fo</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2602.21141-synthrender-and-i-asset-open-source-framework-and-dataset-fo</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 11 Aug 2026 02:23:49 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[{\Omega}-QVLA: Robust Quantization for Vision-Language-Action Models via Composite Rotation and Per-step Scaling]]></title>
      <link>https://embodiedglobal.com/en/article/arxiv-2605.28803-omega-qvla-robust-quantization-for-vision-language-action-mo</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/arxiv-2605.28803-omega-qvla-robust-quantization-for-vision-language-action-mo</guid>
      <description><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></description>
      <content:encoded><![CDATA[New research on arXiv explores advanced approaches in robotics and embodied AI, presenting novel methods for robot learning and autonomous manipulation.]]></content:encoded>
      <pubDate>Tue, 11 Aug 2026 02:23:49 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>research</category>
      <category>arxiv</category>
      <category>paper</category>
      <category>robotics</category>
      <category>AI</category>
      <enclosure url="/images/hero/eg_hero_research.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Unitree Prices STAR Market IPO at RMB 150.80; DeepSeek Takes RMB 141M Strategic Placement with 36-Month Lock-up]]></title>
      <link>https://embodiedglobal.com/en/article/unitree-star-market-ipo-deepseek-strategic-placement</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/unitree-star-market-ipo-deepseek-strategic-placement</guid>
      <description><![CDATA[Unitree Robotics priced its STAR Market IPO at RMB 150.80 per share, a RMB 60.99 billion market cap. DeepSeek took a RMB 141 million strategic placement with an unusual 36-month lock-up, alongside an R&D pact pairing its AI models with Unitree's robot hardware.]]></description>
      <content:encoded><![CDATA[Unitree Robotics priced its STAR Market IPO at RMB 150.80 per share, a RMB 60.99 billion market cap. DeepSeek took a RMB 141 million strategic placement with an unusual 36-month lock-up, alongside an R&D pact pairing its AI models with Unitree's robot hardware.]]></content:encoded>
      <pubDate>Sun, 09 Aug 2026 08:28:04 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Funding</category>
            <category>Unitree Robotics</category>
      <category>IPO</category>
      <category>DeepSeek</category>
      <category>STAR Market</category>
      <category>humanoid robots</category>
      <category>strategic placement</category>
      <category>embodied AI</category>
      <category>Wang Xingxing</category>
      <category>foundation models</category>
      <enclosure url="/images/hero/eg_hero_funding.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Renault Puts Wandercraft's Calvin Humanoid on the Douai Tyre Line, Targets 350 Units by End-2027]]></title>
      <link>https://embodiedglobal.com/en/article/renault-wandercraft-calvin-humanoid-douai-tyre-line-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/renault-wandercraft-calvin-humanoid-douai-tyre-line-2026</guid>
      <description><![CDATA[Renault is trialling Wandercraft's Calvin humanoid on night-shift tyre handling at its Douai plant, with about 10 units planned by end-2026 and roughly 350 across factories in France and Spain by end-2027.]]></description>
      <content:encoded><![CDATA[Renault is trialling Wandercraft's Calvin humanoid on night-shift tyre handling at its Douai plant, with about 10 units planned by end-2026 and roughly 350 across factories in France and Spain by end-2027.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 17:51:36 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>Renault</category>
      <category>Wandercraft</category>
      <category>Calvin</category>
      <category>humanoid robot</category>
      <category>factory deployment</category>
      <category>automotive</category>
      <category>France</category>
      <enclosure url="/images/news/2026-07-27/renault-calvin-douai.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[FORGE-plus: Frozen LLM Sets Force Budgets for Contact-Rich Assembly, Passes 256/256 Trials]]></title>
      <link>https://embodiedglobal.com/en/article/forge-plus-frozen-llm-force-ceiling-contact-rich-assembly-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/forge-plus-frozen-llm-force-ceiling-contact-rich-assembly-2026</guid>
      <description><![CDATA[A two-layer framework lets a frozen text-only LLM assign per-object force ceilings and pick recovery maneuvers from textual force signatures, passing all 256 evaluation episodes on fragile bottle placement and 0.4 mm-clearance gear insertion without breakage.]]></description>
      <content:encoded><![CDATA[A two-layer framework lets a frozen text-only LLM assign per-object force ceilings and pick recovery maneuvers from textual force signatures, passing all 256 evaluation episodes on fragile bottle placement and 0.4 mm-clearance gear insertion without breakage.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 17:51:36 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>FORGE-plus</category>
      <category>contact-rich assembly</category>
      <category>force control</category>
      <category>robot manipulation</category>
      <category>LLM supervisor</category>
      <category>robot learning</category>
      <enclosure url="/images/news/2026-07-27/forge-plus-assembly.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Xynova Takes Meituan-Led Series A+ to Unicorn Status With RMB 1.5B Raised in Total]]></title>
      <link>https://embodiedglobal.com/en/article/xynova-500m-yuan-series-a-plus-meituan-unicorn-dexterous-hand-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/xynova-500m-yuan-series-a-plus-meituan-unicorn-dexterous-hand-2026</guid>
      <description><![CDATA[Dexterous-hand maker Xynova closed a RMB 500 million Series A+ led by Meituan with NIO Capital and China Merchants Capital participating, taking cumulative funding to about RMB 1.5 billion and confirming unicorn status within 19 months of founding.]]></description>
      <content:encoded><![CDATA[Dexterous-hand maker Xynova closed a RMB 500 million Series A+ led by Meituan with NIO Capital and China Merchants Capital participating, taking cumulative funding to about RMB 1.5 billion and confirming unicorn status within 19 months of founding.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 17:51:36 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Deals</category>
            <category>Xynova</category>
      <category>dexterous hand</category>
      <category>funding</category>
      <category>Meituan</category>
      <category>NIO Capital</category>
      <category>unicorn</category>
      <category>embodied AI</category>
      <enclosure url="/images/news/2026-07-27/xynova-dexterous-hand.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[SoftNav: 3D Scene Tokens Turn a Frozen VLM Into a Transferable Navigator]]></title>
      <link>https://embodiedglobal.com/en/article/softnav-3d-scene-tokens-vlm-embodied-navigation-iros-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/softnav-3d-scene-tokens-vlm-embodied-navigation-iros-2026</guid>
      <description><![CDATA[SoftNav, accepted to IROS 2026, injects entity-level 3D scene tokens directly into a frozen VLM's hidden space — hitting 74.2% success on HM3D-OVON with only ~17M trainable parameters and ~1,200 samples, then walking onto a real Unitree Go2 zero-shot.]]></description>
      <content:encoded><![CDATA[SoftNav, accepted to IROS 2026, injects entity-level 3D scene tokens directly into a frozen VLM's hidden space — hitting 74.2% success on HM3D-OVON with only ~17M trainable parameters and ~1,200 samples, then walking onto a real Unitree Go2 zero-shot.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 11:50:33 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>SoftNav</category>
      <category>embodied navigation</category>
      <category>VLM</category>
      <category>3D scene understanding</category>
      <category>IROS 2026</category>
      <category>open-vocabulary navigation</category>
      <category>quadruped robot</category>
      <enclosure url="/images/news/2026-07-26/softnav-arch.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[BYD's Xiaodi Humanoid Set for Early August Debut as Company Denies Factory Trial Rumors]]></title>
      <link>https://embodiedglobal.com/en/article/byd-xiaodi-humanoid-robot-august-debut-di-space-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/byd-xiaodi-humanoid-robot-august-debut-di-space-2026</guid>
      <description><![CDATA[BYD's Di Space Zhengzhou venue teased the Xiaodi humanoid for an early-August public debut — a 1.61 m, 58.5 kg showroom robot with 31 DOF — while the company officially denied the 'Yao-Shun-Yu' prototype and mass-deployment rumors.]]></description>
      <content:encoded><![CDATA[BYD's Di Space Zhengzhou venue teased the Xiaodi humanoid for an early-August public debut — a 1.61 m, 58.5 kg showroom robot with 31 DOF — while the company officially denied the 'Yao-Shun-Yu' prototype and mass-deployment rumors.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 11:41:58 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Product</category>
            <category>BYD</category>
      <category>Xiaodi</category>
      <category>humanoid robot</category>
      <category>Di Space</category>
      <category>showroom robot</category>
      <category>service robot</category>
      <enclosure url="/images/news/2026-07-26/byd-xiaodi-humanoid.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[KineBench: IDM-Free Benchmark Puts Embodied World Models Through a Kinematic Reality Check]]></title>
      <link>https://embodiedglobal.com/en/article/kinebench-idm-free-benchmark-embodied-world-models-eccv-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/kinebench-idm-free-benchmark-embodied-world-models-eccv-2026</guid>
      <description><![CDATA[KineBench, accepted to ECCV 2026, evaluates embodied world models without brittle inverse dynamics models: cascaded vision models extract 6D end-effector poses from generated videos and replay them in a physics simulator across 20 ManiSkill3 tasks.]]></description>
      <content:encoded><![CDATA[KineBench, accepted to ECCV 2026, evaluates embodied world models without brittle inverse dynamics models: cascaded vision models extract 6D end-effector poses from generated videos and replay them in a physics simulator across 20 ManiSkill3 tasks.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 11:41:58 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>KineBench</category>
      <category>world model</category>
      <category>benchmark</category>
      <category>ECCV 2026</category>
      <category>robot manipulation</category>
      <category>kinematics</category>
      <category>simulation</category>
      <enclosure url="/images/news/2026-07-26/kinebench-pipeline.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Masked Visual Actions: One Interface Turns Video Models Into Unified Robot World Models]]></title>
      <link>https://embodiedglobal.com/en/article/masked-visual-actions-video-model-world-model-robot-manipulation-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/masked-visual-actions-video-model-world-model-robot-manipulation-2026</guid>
      <description><![CDATA[A Stanford-led team introduces Masked Visual Actions, a pixel-space control interface that lets video generation models serve as both forward and inverse dynamics models for robots — finetuned with just 15 hours of masked video examples.]]></description>
      <content:encoded><![CDATA[A Stanford-led team introduces Masked Visual Actions, a pixel-space control interface that lets video generation models serve as both forward and inverse dynamics models for robots — finetuned with just 15 hours of masked video examples.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 05:40:41 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>Masked Visual Actions</category>
      <category>world model</category>
      <category>video model</category>
      <category>robot learning</category>
      <category>manipulation</category>
      <category>research</category>
      <enclosure url="/images/news/2026-07-26/mva-overview.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Zhiyuan Innovation Confirms Hong Kong Listing Process as Shipments Top Global Humanoid Market]]></title>
      <link>https://embodiedglobal.com/en/article/zhiyuan-innovation-hkex-listing-process-global-humanoid-shipments-leader-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/zhiyuan-innovation-hkex-listing-process-global-humanoid-shipments-leader-2026</guid>
      <description><![CDATA[Zhiyuan Innovation — the company behind Agibot robots — confirmed on July 24 that it has formally started its Hong Kong listing process, after Omdia ranked it the world's No.1 humanoid shipper of 2025 with a 39% market share.]]></description>
      <content:encoded><![CDATA[Zhiyuan Innovation — the company behind Agibot robots — confirmed on July 24 that it has formally started its Hong Kong listing process, after Omdia ranked it the world's No.1 humanoid shipper of 2025 with a 39% market share.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 05:40:41 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Deals</category>
            <category>Zhiyuan Innovation</category>
      <category>Agibot</category>
      <category>IPO</category>
      <category>HKEX</category>
      <category>humanoid robot</category>
      <category>Deals</category>
      <enclosure url="/images/news/2026-07-26/zhiyuan-robots-lab.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[EngineAI's Second Factory Goes Online: First Henan-Made T800 Humanoids Roll Off the Line]]></title>
      <link>https://embodiedglobal.com/en/article/engineai-t800-zhengzhou-factory-first-batch-production-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/engineai-t800-zhengzhou-factory-first-batch-production-2026</guid>
      <description><![CDATA[EngineAI rolled the first batch of Henan-made T800 humanoid robots off its new Zhengzhou production base on July 24 — the company's second factory after Shenzhen, designed as its global manufacturing hub with a 10,000-unit annual target.]]></description>
      <content:encoded><![CDATA[EngineAI rolled the first batch of Henan-made T800 humanoid robots off its new Zhengzhou production base on July 24 — the company's second factory after Shenzhen, designed as its global manufacturing hub with a 10,000-unit annual target.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 05:40:41 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>EngineAI</category>
      <category>T800</category>
      <category>Zhengzhou</category>
      <category>mass production</category>
      <category>humanoid robot</category>
      <category>manufacturing</category>
      <enclosure url="/images/news/2026-07-26/engineai-t800-zhengzhou-factory.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[The 2:1 Duty-Cycle Problem: Inside the Four-Way Energy War to Power Humanoid Robots]]></title>
      <link>https://embodiedglobal.com/en/article/humanoid-robot-energy-architecture-war-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/humanoid-robot-energy-architecture-war-2026</guid>
      <description><![CDATA[Spec-sheet runtimes rarely survive the factory floor — the honest metric is the duty cycle, and Agility's 2:1 ratio means a third of fleet capital sits at a charger. This deep dive maps the four engineering camps routing around the humanoid battery wall: autonomous hot-swap, structural battery integration, opportunistic charging, and the solid-state staircase — plus the cycle-life economics, thermal limits, and actuator endurance that will decide winners through 2030.]]></description>
      <content:encoded><![CDATA[Spec-sheet runtimes rarely survive the factory floor — the honest metric is the duty cycle, and Agility's 2:1 ratio means a third of fleet capital sits at a charger. This deep dive maps the four engineering camps routing around the humanoid battery wall: autonomous hot-swap, structural battery integration, opportunistic charging, and the solid-state staircase — plus the cycle-life economics, thermal limits, and actuator endurance that will decide winners through 2030.]]></content:encoded>
      <pubDate>Sun, 26 Jul 2026 02:41:33 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>humanoid-robotics</category>
      <category>battery</category>
      <category>energy-density</category>
      <category>solid-state-battery</category>
      <category>hot-swap</category>
      <category>duty-cycle</category>
      <category>thermal-management</category>
      <category>actuator-reliability</category>
      <category>ubtech</category>
      <category>figure-ai</category>
      <category>tesla</category>
      <category>agility-robotics</category>
      <category>xpeng</category>
      <category>research</category>
      <enclosure url="/images/articles/2026-07-26/humanoid-energy-architecture.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Bionic Face Robot Maker ShouXing Completes Seventh Funding Round in Two Years]]></title>
      <link>https://embodiedglobal.com/en/article/shouxing-aheadform-a2-round-bionic-robot-face-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/shouxing-aheadform-a2-round-bionic-robot-face-2026</guid>
      <description><![CDATA[AheadForm (ShouXing Tech), the bionic face robot startup backed by Zhiyuan Robotics and Shunwei Capital, has closed a hundreds-of-millions RMB A2 round — its seventh funding round in two years — to build standardized production lines for expressive robotic heads.]]></description>
      <content:encoded><![CDATA[AheadForm (ShouXing Tech), the bionic face robot startup backed by Zhiyuan Robotics and Shunwei Capital, has closed a hundreds-of-millions RMB A2 round — its seventh funding round in two years — to build standardized production lines for expressive robotic heads.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 23:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Deals</category>
            <category>AheadForm</category>
      <category>ShouXing Tech</category>
      <category>bionic face</category>
      <category>funding</category>
      <category>Shunwei Capital</category>
      <category>humanoid robot</category>
      <enclosure url="/images/news/2026-07-26/shouxing-bionic-face-robot.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Embodied AI Skills Competition Opens Global Call for October Finals in Shanghai]]></title>
      <link>https://embodiedglobal.com/en/article/gdps-2026-embodied-ai-skills-competition-october-shanghai</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/gdps-2026-embodied-ai-skills-competition-october-shanghai</guid>
      <description><![CDATA[The 2026 Global Developer Pioneer Summit and International Embodied AI Skills Competition has opened its global call, with finals October 23-25 at Shanghai's Zhangjiang Science Hall, Turing Award winner Yao Qizhi chairing the expert committee.]]></description>
      <content:encoded><![CDATA[The 2026 Global Developer Pioneer Summit and International Embodied AI Skills Competition has opened its global call, with finals October 23-25 at Shanghai's Zhangjiang Science Hall, Turing Award winner Yao Qizhi chairing the expert committee.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 23:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>GDPS 2026</category>
      <category>embodied AI</category>
      <category>competition</category>
      <category>Shanghai</category>
      <category>Yao Qizhi</category>
      <category>humanoid robot</category>
      <enclosure url="/images/news/2026-07-26/gdps-embodied-ai-competition.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[What Actually Matters in Humanoid Whole-Body Motion Tracking: YAHMP Study Benchmarks Design Choices]]></title>
      <link>https://embodiedglobal.com/en/article/yahmp-humanoid-whole-body-motion-tracking-empirical-study-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/yahmp-humanoid-whole-body-motion-tracking-empirical-study-2026</guid>
      <description><![CDATA[IIT HUCEBOT researchers released YAHMP, an open-source modular framework for humanoid whole-body motion tracking, with a controlled empirical study of design choices and zero-shot deployment on a real Unitree G1.]]></description>
      <content:encoded><![CDATA[IIT HUCEBOT researchers released YAHMP, an open-source modular framework for humanoid whole-body motion tracking, with a controlled empirical study of design choices and zero-shot deployment on a real Unitree G1.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 23:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>YAHMP</category>
      <category>motion tracking</category>
      <category>humanoid robot</category>
      <category>robot learning</category>
      <category>IIT HUCEBOT</category>
      <category>whole-body control</category>
      <enclosure url="/images/news/2026-07-26/yahmp-humanoid-motion-tracking.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Robostral Navigate: 8B VLM Sets New SOTA for Monocular Robot Navigation]]></title>
      <link>https://embodiedglobal.com/en/article/robostral-navigate-8b-vlm-monocular-robot-navigation-sota</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/robostral-navigate-8b-vlm-monocular-robot-navigation-sota</guid>
      <description><![CDATA[Mistral AI's Robostral Navigate is an 8B vision-language model that predicts navigation waypoints from a single RGB camera, setting new state of the art on R2R-CE and RxR-CE benchmarks across wheeled, legged and aerial robots.]]></description>
      <content:encoded><![CDATA[Mistral AI's Robostral Navigate is an 8B vision-language model that predicts navigation waypoints from a single RGB camera, setting new state of the art on R2R-CE and RxR-CE benchmarks across wheeled, legged and aerial robots.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 17:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>Mistral AI</category>
      <category>Robostral Navigate</category>
      <category>vision-language model</category>
      <category>robot navigation</category>
      <category>VLM</category>
      <category>waypoint prediction</category>
      <enclosure url="/images/news/2026-07-25/robostral-navigate-vlm.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Hyundai Charts Physical AI Vision: From Robot Platform to City-Scale Intelligence]]></title>
      <link>https://embodiedglobal.com/en/article/hyundai-physical-ai-vision-city-scale-intelligence-sf-summit-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/hyundai-physical-ai-vision-city-scale-intelligence-sf-summit-2026</guid>
      <description><![CDATA[Hyundai Motor Group Executive Chair Euisun Chung outlined the group's Physical AI vision at the San Francisco AI Summit, announcing a Robot Reference Platform collaboration with NVIDIA and major investments including the Saemangeum AI Valley.]]></description>
      <content:encoded><![CDATA[Hyundai Motor Group Executive Chair Euisun Chung outlined the group's Physical AI vision at the San Francisco AI Summit, announcing a Robot Reference Platform collaboration with NVIDIA and major investments including the Saemangeum AI Valley.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 17:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>Hyundai Motor Group</category>
      <category>Physical AI</category>
      <category>Saemangeum AI Valley</category>
      <category>robotics strategy</category>
      <category>AI Defined Factory</category>
      <category>Euisun Chung</category>
      <enclosure url="/images/news/2026-07-25/hyundai-physical-ai-summit.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Wayve Lands $1.2B Series D to Scale Embodied AI for Autonomous Driving]]></title>
      <link>https://embodiedglobal.com/en/article/wayve-1-2b-series-d-embodied-ai-autonomous-driving-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/wayve-1-2b-series-d-embodied-ai-autonomous-driving-2026</guid>
      <description><![CDATA[UK self-driving startup Wayve has raised $1.2 billion in a Series D round led by SoftBank with participation from NVIDIA, Uber and three major automakers, accelerating its end-to-end embodied AI driving platform.]]></description>
      <content:encoded><![CDATA[UK self-driving startup Wayve has raised $1.2 billion in a Series D round led by SoftBank with participation from NVIDIA, Uber and three major automakers, accelerating its end-to-end embodied AI driving platform.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 17:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Deals</category>
            <category>Wayve</category>
      <category>Uber</category>
      <category>Series D</category>
      <category>autonomous driving</category>
      <category>embodied AI</category>
      <category>funding</category>
      <enclosure url="/images/news/2026-07-25/wayve-series-d.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Zhiyuan Robotics, ZTT Launch Embodied AI Innovation Center in Nantong]]></title>
      <link>https://embodiedglobal.com/en/article/zhiyuan-ztt-embodied-ai-innovation-center-nantong-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/zhiyuan-ztt-embodied-ai-innovation-center-nantong-2026</guid>
      <description><![CDATA[The Zhongtian–Zhiyuan Embodied AI Innovation Center has entered trial operation in Nantong's Chongchuan District. Co-built by ZTT, Zhiyuan Robotics (AGIBOT) and three other partners with a total investment of 100 million yuan, the center focuses on embodied AI data collection and annotation, model iteration, and custom industrial robot development. Robots, sensors and computing facilities are in place, with data collection and simulation training systems having completed initial joint debugging.]]></description>
      <content:encoded><![CDATA[The Zhongtian–Zhiyuan Embodied AI Innovation Center has entered trial operation in Nantong's Chongchuan District. Co-built by ZTT, Zhiyuan Robotics (AGIBOT) and three other partners with a total investment of 100 million yuan, the center focuses on embodied AI data collection and annotation, model iteration, and custom industrial robot development. Robots, sensors and computing facilities are in place, with data collection and simulation training systems having completed initial joint debugging.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 14:41:19 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>Zhiyuan Robotics</category>
      <category>Zhongtian Technology</category>
      <category>ZTT</category>
      <category>embodied AI</category>
      <category>innovation center</category>
      <category>Nantong</category>
      <category>industrial robotics</category>
      <category>data collection</category>
      <enclosure url="/images/news/2026-07-25/zhiyuan-ztt-innovation-center.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Shihe Robotics (RobotPlusPlus) Closes Hundreds-of-Millions-Yuan Series C, Targets 'Wall-Climbing Embodied Robot' IPO]]></title>
      <link>https://embodiedglobal.com/en/article/shihe-robotics-series-c-hundreds-millions-yuan-special-robots-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/shihe-robotics-series-c-hundreds-millions-yuan-special-robots-2026</guid>
      <description><![CDATA[Shihe Robotics (RobotPlusPlus) has closed a Series C round worth several hundred million yuan, co-led by Qiangang Capital and Zhongji Holdings, as the ship-hull robot leader positions itself as the first public 'wall-climbing embodied robot' stock.]]></description>
      <content:encoded><![CDATA[Shihe Robotics (RobotPlusPlus) has closed a Series C round worth several hundred million yuan, co-led by Qiangang Capital and Zhongji Holdings, as the ship-hull robot leader positions itself as the first public 'wall-climbing embodied robot' stock.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 11:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Deals</category>
            <category>Shihe Robotics</category>
      <category>RobotPlusPlus</category>
      <category>Series C</category>
      <category>funding</category>
      <category>wall-climbing robot</category>
      <category>embodied AI</category>
      <category>special-purpose robots</category>
      <enclosure url="/images/news/2026-07-25/shihe-robotplus-series-c.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[Zoomlion Puts Dozens of Self-Built Humanoids to Work on Real Production Lines in Changsha]]></title>
      <link>https://embodiedglobal.com/en/article/zoomlion-embodied-ai-robots-real-production-lines-changsha-2026</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/zoomlion-embodied-ai-robots-real-production-lines-changsha-2026</guid>
      <description><![CDATA[Construction machinery giant Zoomlion is running dozens of self-developed humanoid robots on real production lines at its Changsha Smart Industry City, validating nearly 20 scenarios across machining, logistics, assembly and quality inspection.]]></description>
      <content:encoded><![CDATA[Construction machinery giant Zoomlion is running dozens of self-developed humanoid robots on real production lines at its Changsha Smart Industry City, validating nearly 20 scenarios across machining, logistics, assembly and quality inspection.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 11:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>Zoomlion</category>
      <category>humanoid robot</category>
      <category>embodied AI</category>
      <category>manufacturing</category>
      <category>factory deployment</category>
      <category>Changsha</category>
      <category>Physical AI</category>
      <enclosure url="/images/news/2026-07-25/zoomlion-humanoid-production-line.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[URF Unifies Impedance-Admittance Control for Stable Contact-Aware Robot Manipulation]]></title>
      <link>https://embodiedglobal.com/en/article/urf-unified-impedance-admittance-control-contact-aware-manipulation</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/urf-unified-impedance-admittance-control-contact-aware-manipulation</guid>
      <description><![CDATA[Sungkyunkwan University researchers propose URF, a unified control-policy framework that predicts virtual targets, stiffness matrices and an impedance-admittance switch ratio alongside actions, reducing tool breakage and safety stops in rigid-contact manipulation.]]></description>
      <content:encoded><![CDATA[Sungkyunkwan University researchers propose URF, a unified control-policy framework that predicts virtual targets, stiffness matrices and an impedance-admittance switch ratio alongside actions, reducing tool breakage and safety stops in rigid-contact manipulation.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 11:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>URF</category>
      <category>impedance control</category>
      <category>contact-rich manipulation</category>
      <category>robot learning</category>
      <category>Sungkyunkwan University</category>
      <category>control policy</category>
      <category>arXiv</category>
      <enclosure url="/images/news/2026-07-25/urf-contact-aware-manipulation.png" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[South Korea's Navy and KAIST Test PIBOT Humanoid as Ship Helmsman in First Trial]]></title>
      <link>https://embodiedglobal.com/en/article/kaist-pibot-south-korea-navy-humanoid-helmsman-first-trial</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/kaist-pibot-south-korea-navy-humanoid-helmsman-first-trial</guid>
      <description><![CDATA[The Republic of Korea Navy and KAIST have conducted South Korea's first trial of a humanoid robot serving as a ship's helmsman. PIBOT, a 165 cm, 65 kg humanoid trained with a large language model on naval terminology and ship-handling procedures, steered a simulated 7,600-ton Aegis destroyer through a narrow channel. The four-phase program aims to integrate robots into shipboard operations as Seoul faces a shrinking pool of military recruits.]]></description>
      <content:encoded><![CDATA[The Republic of Korea Navy and KAIST have conducted South Korea's first trial of a humanoid robot serving as a ship's helmsman. PIBOT, a 165 cm, 65 kg humanoid trained with a large language model on naval terminology and ship-handling procedures, steered a simulated 7,600-ton Aegis destroyer through a narrow channel. The four-phase program aims to integrate robots into shipboard operations as Seoul faces a shrinking pool of military recruits.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 05:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Industry</category>
            <category>KAIST</category>
      <category>PIBOT</category>
      <category>Republic of Korea Navy</category>
      <category>humanoid robot</category>
      <category>military robotics</category>
      <category>defense</category>
      <category>LLM</category>
      <category>embodied AI</category>
      <category>naval operations</category>
      <enclosure url="/images/news/2026-07-25/kaist-pibot-navy.jpg" type="image/jpeg" length="0"/>
    </item>
    <item>
      <title><![CDATA[AlohaMini2: Sub-$1,000 Open-Source Mobile Manipulator Trains Long-Horizon Tasks on a Consumer 8GB GPU]]></title>
      <link>https://embodiedglobal.com/en/article/alohamini2-sub-1000-open-source-robot-8gb-gpu-training</link>
      <guid isPermaLink="true">https://embodiedglobal.com/en/article/alohamini2-sub-1000-open-source-robot-8gb-gpu-training</guid>
      <description><![CDATA[The open-source AlohaMini2 project demonstrates autonomous long-horizon mobile manipulation — including a full grocery-shopping task — trained entirely on a standard 8GB consumer GPU with just 50 human demonstration episodes. With a self-build bill of materials under $1,000 and fully 3D-printable frame, the platform removes the compute and cost barriers that have kept embodied AI research inside well-funded labs.]]></description>
      <content:encoded><![CDATA[The open-source AlohaMini2 project demonstrates autonomous long-horizon mobile manipulation — including a full grocery-shopping task — trained entirely on a standard 8GB consumer GPU with just 50 human demonstration episodes. With a self-build bill of materials under $1,000 and fully 3D-printable frame, the platform removes the compute and cost barriers that have kept embodied AI research inside well-funded labs.]]></content:encoded>
      <pubDate>Sat, 25 Jul 2026 05:47:00 GMT</pubDate>
      <author>business@allinagi.com.cn (Embodied Global Team)</author>
      <category>Research</category>
            <category>AlohaMini2</category>
      <category>open source</category>
      <category>mobile manipulation</category>
      <category>imitation learning</category>
      <category>LeRobot</category>
      <category>embodied AI</category>
      <category>robot learning</category>
      <category>low-cost robotics</category>
      <category>AM-ACT</category>
      <enclosure url="/images/news/2026-07-25/alohamini2-open-source.jpg" type="image/jpeg" length="0"/>
    </item>
    
  </channel>
</rss>