<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>Biped News</title>
  <link href="https://biped.news/" />
  <link rel="self" href="https://biped.news/feed.xml" />
  <id>https://biped.news/</id>
  <updated>2026-04-21T18:08:10.288Z</updated>
  <subtitle>News and analysis covering the emerging humanoid robotics industry</subtitle>
  <author>
    <name>Biped News</name>
  </author>
  <entry>
    <title>Tesla&apos;s Robot Hands Have 50 Actuators. Here&apos;s Why.</title>
    <link href="https://biped.news/article/tesla-optimus-gen3-hands-50-actuators" />
    <id>biped-article-tesla-optimus-gen3-hands-50-actuators</id>
    <updated>2026-03-11T13:07:59.754Z</updated>
    <published>2026-03-11T00:00:00.000Z</published>
    <summary>Deep analysis of the Tesla Optimus Gen 3 hand reveal, covering the 50-actuator architecture, evolution from Gen 2, comparison with competitor humanoid robots, Tesla&apos;s Fremont factory conversion strategy, and the unsolved engineering challenges of human-level robotic dexterity.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/tesla-optimus-gen3-hands-50-actuators-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Tesla" />
    <category term="Optimus" />
    <category term="Humanoid Robotics" />
    <category term="Gen 3" />
    <category term="Dexterity" />
    <category term="Manufacturing" />
    <category term="AI Robotics" />
  </entry>
  <entry>
    <title>How Humanoid Robots Actually Walk: The Engineering Inside</title>
    <link href="https://biped.news/article/bipedal-locomotion-how-robots-walk" />
    <id>biped-article-bipedal-locomotion-how-robots-walk</id>
    <updated>2026-03-10T01:17:13.861Z</updated>
    <published>2026-03-10T01:17:13.860Z</published>
    <summary>Walking on two legs is a controlled fall requiring real-time whole-body control at 1 kHz, 6-axis force sensing in each foot, and terrain prediction from depth cameras. This explainer covers ZMP theory, capture point control, whole-body control frameworks, how reinforcement learning is improving robot gaits, the actuator design trade-offs from quasi-direct drive to series elastic actuators, and where the leading humanoid robot companies stand on deploying reliable locomotion in real industrial environments.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/bipedal-locomotion-how-robots-walk-hero.png" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="bipedal locomotion" />
    <category term="humanoid robots" />
    <category term="whole-body control" />
    <category term="reinforcement learning" />
    <category term="Figure AI" />
    <category term="Agility Robotics" />
    <category term="Boston Dynamics" />
    <category term="Unitree" />
    <category term="actuators" />
    <category term="walking robots" />
  </entry>
  <entry>
    <title>Unitree Robotics: Company Profile</title>
    <link href="https://biped.news/article/unitree-robotics-company-profile" />
    <id>biped-article-unitree-robotics-company-profile</id>
    <updated>2026-03-11T14:53:35.137Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>Founded in 2016 by Wang Xingxing, Unitree Robotics offers the G1 and H1 humanoid robots, blending advanced technology with affordability at just $16,000.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/unitree-robotics-company-profile-hero-v2.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Unitree Robotics" />
    <category term="G1 robot" />
    <category term="H1 robot" />
    <category term="Chinese robotics" />
    <category term="affordable humanoids" />
    <category term="Wang Xingxing" />
    <category term="Hangzhou" />
  </entry>
  <entry>
    <title>UBTECH Robotics: Company Profile</title>
    <link href="https://biped.news/article/ubtech-robotics-company-profile" />
    <id>biped-article-ubtech-robotics-company-profile</id>
    <updated>2026-03-09T12:16:34.799Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>Founded in 2012 by James Zhou in Shenzhen, UBTECH Robotics is redefining innovation with Walker X, a 1.65m humanoid robot now deployed by BMW Brilliance in Shenyang.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/ubtech-robotics-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="UBTECH Robotics" />
    <category term="Walker X" />
    <category term="Shenzhen" />
    <category term="BMW" />
    <category term="FAW-Volkswagen" />
    <category term="Chinese humanoids" />
    <category term="industrial automation" />
  </entry>
  <entry>
    <title>Sanctuary AI: Company Profile</title>
    <link href="https://biped.news/article/sanctuary-ai-company-profile" />
    <id>biped-article-sanctuary-ai-company-profile</id>
    <updated>2026-03-09T12:15:11.769Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>Founded in 2018 and based in Vancouver, Sanctuary AI pioneers the Phoenix humanoid robot, leveraging Carbon AI to push the boundaries of advanced cognition.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/sanctuary-ai-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Sanctuary AI" />
    <category term="Phoenix robot" />
    <category term="Carbon AI" />
    <category term="Microsoft" />
    <category term="Vancouver" />
    <category term="AGI" />
    <category term="Geordie Rose" />
  </entry>
  <entry>
    <title>Physical Intelligence (π): Company Profile</title>
    <link href="https://biped.news/article/physical-intelligence-company-profile" />
    <id>biped-article-physical-intelligence-company-profile</id>
    <updated>2026-03-09T12:18:35.946Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>Founded in 2023 by leading AI pioneers from UC Berkeley and Stanford, Physical Intelligence (π) is revolutionizing robotics with advanced foundation models.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/physical-intelligence-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Physical Intelligence" />
    <category term="pi zero" />
    <category term="foundation model" />
    <category term="robot manipulation" />
    <category term="AI" />
    <category term="San Francisco" />
    <category term="Sergey Levine" />
  </entry>
  <entry>
    <title>Figure AI: Company Profile</title>
    <link href="https://biped.news/article/figure-ai-company-profile" />
    <id>biped-article-figure-ai-company-profile</id>
    <updated>2026-03-09T12:13:21.432Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>Figure AI, valued at $2.6B after a $675M Series B in 2024, partners with BMW and OpenAI, backed by Microsoft and NVIDIA for cutting-edge AI integration.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/figure-ai-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Figure AI" />
    <category term="Figure 02" />
    <category term="humanoid robots" />
    <category term="BMW" />
    <category term="OpenAI" />
    <category term="manufacturing automation" />
    <category term="Brett Adcock" />
  </entry>
  <entry>
    <title>Apptronik: Company Profile</title>
    <link href="https://biped.news/article/apptronik-company-profile" />
    <id>biped-article-apptronik-company-profile</id>
    <updated>2026-03-09T12:14:21.559Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>Founded in 2016 in Austin, TX, Apptronik creates Apollo, a 5&apos;8&quot;, 73kg humanoid robot designed to revolutionize logistics and manufacturing.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/apptronik-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Apptronik" />
    <category term="Apollo robot" />
    <category term="NASA" />
    <category term="Samsung" />
    <category term="Austin Texas" />
    <category term="humanoid robots" />
    <category term="GXO Logistics" />
  </entry>
  <entry>
    <title>Agility Robotics: Company Profile</title>
    <link href="https://biped.news/article/agility-robotics-company-profile" />
    <id>biped-article-agility-robotics-company-profile</id>
    <updated>2026-03-09T12:17:14.795Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>Founded in 2015 as a spin-off from Oregon State University, Agility Robotics specializes in Digit, a bipedal humanoid robot designed for warehouse efficiency.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/agility-robotics-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Agility Robotics" />
    <category term="Digit" />
    <category term="Amazon" />
    <category term="warehouse automation" />
    <category term="bipedal robots" />
    <category term="Oregon" />
    <category term="RoboFab" />
  </entry>
  <entry>
    <title>Aethon / Vecna Robotics: Company Profile</title>
    <link href="https://biped.news/article/aethon-vecna-company-profile" />
    <id>biped-article-aethon-vecna-company-profile</id>
    <updated>2026-03-09T12:17:55.733Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>Aethon, founded in 2001, revolutionizes hospital logistics with TUG, its autonomous mobile robot delivering medications and meals seamlessly.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/aethon-vecna-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Aethon" />
    <category term="Vecna Robotics" />
    <category term="TUG robot" />
    <category term="hospital automation" />
    <category term="warehouse robots" />
    <category term="AMR" />
    <category term="ST Engineering" />
  </entry>
  <entry>
    <title>1X Technologies: Company Profile</title>
    <link href="https://biped.news/article/1x-technologies-company-profile" />
    <id>biped-article-1x-technologies-company-profile</id>
    <updated>2026-03-09T12:13:32.428Z</updated>
    <published>2026-03-09T00:00:00.000Z</published>
    <summary>1X Technologies, formerly Halodi Robotics, pioneers humanoid robots for home use, backed by a $23.5M OpenAI-led Series A, with HQ in Norway and California.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/1x-technologies-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="1X Technologies" />
    <category term="NEO" />
    <category term="OpenAI" />
    <category term="Norwegian robotics" />
    <category term="humanoid robots" />
    <category term="home robots" />
    <category term="Halodi Robotics" />
  </entry>
  <entry>
    <title>Who Is Winning the Humanoid Robot Race in 2026?</title>
    <link href="https://biped.news/article/humanoid-robot-race-2026" />
    <id>biped-article-humanoid-robot-race-2026</id>
    <updated>2026-02-28T22:14:53.930Z</updated>
    <published>2026-02-28T00:00:00.000Z</published>
    <summary>Five companies are running hard toward the same finish line — commercially viable humanoid robots doing real work. Figure AI, Tesla, Boston Dynamics, Agility Robotics, and 1X are all in or near production deployment in 2026. Here&apos;s where each stands, what makes each platform distinctive, and which competitive advantages are likely to compound into market leadership.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/humanoid-robot-race-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="humanoid robots" />
    <category term="Figure AI" />
    <category term="Tesla Optimus" />
    <category term="Boston Dynamics Atlas" />
    <category term="Agility Digit" />
    <category term="1X NEO" />
    <category term="robotics industry" />
    <category term="2026" />
  </entry>
  <entry>
    <title>Tesla Optimus: The Manufacturing Giant&apos;s Humanoid Robot Strategy</title>
    <link href="https://biped.news/article/tesla-optimus-gen3-2026" />
    <id>biped-article-tesla-optimus-gen3-2026</id>
    <updated>2026-03-11T14:53:35.137Z</updated>
    <published>2026-02-26T00:00:00.000Z</published>
    <summary>Tesla Optimus Gen 2 is performing battery handling at Fremont with 1,000+ units planned for 2025. Gen 3, under new Autopilot-origin AI leadership, targets external sales in 2026. The robot isn&apos;t the most advanced on paper — but Tesla&apos;s manufacturing ecosystem might make it the most important.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/tesla-optimus-gen3-2026-hero-v2.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Tesla" />
    <category term="Tesla Optimus" />
    <category term="humanoid robots" />
    <category term="Elon Musk" />
    <category term="Dojo" />
    <category term="FSD" />
    <category term="Ashok Elluswamy" />
    <category term="manufacturing automation" />
  </entry>
  <entry>
    <title>Boston Dynamics Atlas Goes Electric: A New Era of Industrial Robotics</title>
    <link href="https://biped.news/article/boston-dynamics-atlas-electric-2026" />
    <id>biped-article-boston-dynamics-atlas-electric-2026</id>
    <updated>2026-03-11T14:53:35.137Z</updated>
    <published>2026-02-24T00:00:00.000Z</published>
    <summary>The hydraulic Atlas defined a decade of bipedal robotics research. The electric Atlas, unveiled in April 2024, aims to define the next decade of industrial deployment: 90 kg, 56 DOF, IP67 rated, self-swappable batteries, Orbit fleet software. Boston Dynamics is finally going commercial.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/boston-dynamics-atlas-electric-2026-hero-v2.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Boston Dynamics" />
    <category term="Atlas" />
    <category term="electric robot" />
    <category term="Hyundai" />
    <category term="industrial automation" />
    <category term="humanoid robots" />
    <category term="Orbit" />
  </entry>
  <entry>
    <title>Figure 02: The Robot Built for Real Work</title>
    <link href="https://biped.news/article/figure-02-real-work-2026" />
    <id>biped-article-figure-02-real-work-2026</id>
    <updated>2026-02-28T22:14:53.486Z</updated>
    <published>2026-02-22T00:00:00.000Z</published>
    <summary>Figure 02 is the most commercially advanced humanoid robot in 2026 — 60 kg, 16 DOF hands, 25 kg payload, deployed at BMW&apos;s Spartanburg plant, backed by $1.75B+ in funding and a $39B valuation. Here&apos;s the full technical and business picture.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/figure-02-real-work-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Figure AI" />
    <category term="Figure 02" />
    <category term="humanoid robots" />
    <category term="BMW" />
    <category term="Helix VLA" />
    <category term="manufacturing automation" />
    <category term="Brett Adcock" />
  </entry>
  <entry>
    <title>Why 2026 Is the Year Robots Get Real Jobs</title>
    <link href="https://biped.news/article/humanoid-robots-real-jobs-2026" />
    <id>biped-article-humanoid-robots-real-jobs-2026</id>
    <updated>2026-02-28T22:14:54.077Z</updated>
    <published>2026-02-20T00:00:00.000Z</published>
    <summary>2023 was demos. 2024 was early pilots. 2026 is when humanoid robots start appearing in enough production environments — BMW, Amazon, Tesla, Hyundai — that the question shifts from &quot;can they work?&quot; to &quot;how many, and how fast?&quot; Here&apos;s the analysis of why this year is the real inflection point.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/humanoid-robots-real-jobs-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="humanoid robots" />
    <category term="manufacturing automation" />
    <category term="labor market" />
    <category term="Figure AI" />
    <category term="Agility Digit" />
    <category term="Tesla Optimus" />
    <category term="workforce" />
    <category term="2026" />
  </entry>
  <entry>
    <title>A Robot Just Beat the Human Half-Marathon World Record</title>
    <link href="https://biped.news/article/honor-lightning-marathon-world-record-2026" />
    <id>biped-article-honor-lightning-marathon-world-record-2026</id>
    <updated>2026-04-21T09:19:27.370Z</updated>
    <published>2026-04-21T09:19:27.370Z</published>
    <summary>On April 19, 2026, Honor&apos;s Lightning humanoid robot won the Beijing E-Town Humanoid Robot Half-Marathon in 50 minutes and 26 seconds, fully autonomous, beating the human world record of 57:20 by nearly 7 minutes. Honor swept the top three positions in the autonomous category. The race featured 300+ robots from 100+ teams, with roughly 40 percent running in fully autonomous mode -- a dramatic jump from nearly zero in 2025. Lightning stands 169 cm, weighs about 45 kg, carries 55 self-developed bionic joints with 400 Nm peak torque, and uses a liquid-cooling system adapted from Honor&apos;s smartphone engineering. The result marks a public proof point for outdoor autonomous bipedal locomotion at scale.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/honor-lightning-marathon-world-record-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Honor" />
    <category term="Lightning" />
    <category term="Beijing" />
    <category term="half-marathon" />
    <category term="world record" />
    <category term="autonomous" />
    <category term="China" />
    <category term="outdoor locomotion" />
    <category term="humanoid racing" />
    <category term="MagicOS" />
  </entry>
  <entry>
    <title>Kepler Robotics: The Chinese Startup That Put a Humanoid in a GM Factory</title>
    <link href="https://biped.news/article/kepler-robotics-company-profile" />
    <id>biped-article-kepler-robotics-company-profile</id>
    <updated>2026-04-21T09:08:53.598Z</updated>
    <published>2026-04-21T09:08:53.598Z</published>
    <summary>Kepler Robotics (Shanghai, founded ~2023) builds the K2 humanoid robot: 52 DOF, 85 kg, 15 kg per-hand payload, up to 8 hours battery life, walking at 4 km/h. Deployed at SAIC-GM&apos;s Shanghai plant in 2025. Raised $20M+ including a 100M+ CNY strategic round in April 2026 led by SAIF Partners. Targets $20K-$30K per unit at scale. Runs Kepler OS, an open modular operating system for third-party app development.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/kepler-robotics-hero.png" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Kepler Robotics" />
    <category term="K2 humanoid" />
    <category term="China robotics" />
    <category term="humanoid robots" />
    <category term="SAIC-GM" />
    <category term="Kepler OS" />
    <category term="company profile" />
  </entry>
  <entry>
    <title>Tactile Sensing in Humanoid Robots: How Machines Learn to Feel</title>
    <link href="https://biped.news/article/tactile-sensing-humanoid-robots-explained" />
    <id>biped-article-tactile-sensing-humanoid-robots-explained</id>
    <updated>2026-04-21T08:58:11.556Z</updated>
    <published>2026-04-21T08:58:11.556Z</published>
    <summary>Tactile sensing is the capability that separates pick-and-place robots from machines capable of genuine dexterous manipulation. This explainer covers vision-based sensors, electronic skin, the companies building them, comparison specs, and why tactile feedback is a prerequisite for both factory scale and eventual home deployment.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/tactile-sensing-humanoid-robots-explained-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="tactile sensing" />
    <category term="electronic skin" />
    <category term="haptic" />
    <category term="GelSight" />
    <category term="Touchlab" />
    <category term="manipulation" />
    <category term="dexterous robotics" />
    <category term="humanoid robots" />
    <category term="robot hands" />
  </entry>
  <entry>
    <title>Beijing&apos;s Humanoid Half-Marathon Becomes China&apos;s Biggest Public Test of Physical AI</title>
    <link href="https://biped.news/article/beijing-humanoid-half-marathon-showcase-2026" />
    <id>biped-article-beijing-humanoid-half-marathon-showcase-2026</id>
    <updated>2026-04-19T02:08:37.927Z</updated>
    <published>2026-04-19T02:00:00.000Z</published>
    <summary>China&apos;s second humanoid half-marathon has grown into one of the clearest public benchmarks for embodied AI. Reuters reported more than 300 robots from over 70 teams, a tougher course, and a sharp rise in autonomous entries versus last year. The showcase also lands amid data showing China accounted for more than 80 percent of global humanoid installations in 2025. That combination makes the event a meaningful test of whether deployment scale is starting to translate into durable robot capability.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/beijing-humanoid-half-marathon-showcase-2026-1.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Beijing" />
    <category term="Humanoid Robots" />
    <category term="Physical AI" />
    <category term="China" />
    <category term="Unitree" />
    <category term="UBTech" />
    <category term="AgiBot" />
    <category term="robotics policy" />
    <category term="robot endurance" />
  </entry>
  <entry>
    <title>China Wins the First Big Rulemaking Battle in Humanoid Robotics</title>
    <link href="https://biped.news/article/china-iso-humanoid-robot-datasets-2026" />
    <id>biped-article-china-iso-humanoid-robot-datasets-2026</id>
    <updated>2026-04-18T02:07:44.243Z</updated>
    <published>2026-04-18T02:00:00.000Z</published>
    <summary>China said on April 17 that ISO approved the first international standard project focused on humanoid robot datasets and set up the first robotics standards working group convened by a Chinese expert. On the surface, that sounds dry compared with a robot launch or funding round. In practice, it goes straight to the bottleneck behind physical AI, high-quality training data that can be used across labs, factories, and products. For a country that already dominates global humanoid installations, turning its domestic dataset work into an ISO process could give Chinese firms more influence over how the industry&apos;s core data plumbing gets defined.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/china-iso-humanoid-robot-datasets-2026-1.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="China" />
    <category term="ISO" />
    <category term="humanoid robots" />
    <category term="embodied AI" />
    <category term="robotics standards" />
    <category term="datasets" />
    <category term="policy" />
    <category term="UBTECH" />
    <category term="Unitree" />
  </entry>
  <entry>
    <title>The Power Problem: How Humanoid Robots Manage Energy From Battery to Joint</title>
    <link href="https://biped.news/article/humanoid-robot-power-energy-architecture-explained" />
    <id>biped-article-humanoid-robot-power-energy-architecture-explained</id>
    <updated>2026-04-16T16:41:53.864Z</updated>
    <published>2026-04-17T04:00:00.000Z</published>
    <summary>Power architecture is humanoid robotics&apos; most underappreciated constraint. This explainer covers battery chemistry, power distribution, the cost of transport metric, the compute tax from onboard AI, and the solid-state battery roadmap that could finally enable full-shift industrial deployment.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/humanoid-robot-power-energy-architecture-explained-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="power systems" />
    <category term="battery" />
    <category term="energy architecture" />
    <category term="actuators" />
    <category term="hardware" />
    <category term="solid-state battery" />
    <category term="efficiency" />
    <category term="cost of transport" />
  </entry>
  <entry>
    <title>Skild AI Buys Zebra&apos;s Robotics Arm to Bring Physical AI Into Live Warehouses</title>
    <link href="https://biped.news/article/skild-zebra-warehouse-physical-ai-2026" />
    <id>biped-article-skild-zebra-warehouse-physical-ai-2026</id>
    <updated>2026-04-16T09:03:12.803Z</updated>
    <published>2026-04-17T02:00:00.000Z</published>
    <summary>Skild AI said on April 15 that it acquired Zebra Technologies&apos; robotics arm to bring its omni-bodied Skild Brain into warehouses. Zebra contributes a mature fulfillment stack with autonomous mobile robots, wearable-linked orchestration, and live customer deployments, while Skild brings a general robot model that aims to work across humanoids, quadrupeds, arms, and mobile manipulators. The combination is not a humanoid launch, but it may still be one of the more important physical AI deals of the month because it connects model training, enterprise workflows, and recurring warehouse data in a single pipeline.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/skild-zebra-warehouse-physical-ai-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Skild AI" />
    <category term="Zebra Technologies" />
    <category term="Fetch Robotics" />
    <category term="Symmetry Fulfillment" />
    <category term="warehouse automation" />
    <category term="physical AI" />
    <category term="humanoid robots" />
    <category term="logistics" />
    <category term="robotics software" />
    <category term="embodied AI" />
  </entry>
  <entry>
    <title>Siemens Puts Humanoid&apos;s HMND 01 on the Factory Floor, Hits 60 Tote Moves an Hour</title>
    <link href="https://biped.news/article/siemens-humanoid-hmnd01-factory-physical-ai-2026" />
    <id>biped-article-siemens-humanoid-hmnd01-factory-physical-ai-2026</id>
    <updated>2026-04-16T16:52:55.783Z</updated>
    <published>2026-04-17T02:00:00.000Z</published>
    <summary>Siemens and UK startup Humanoid used a live proof of concept in Erlangen to show that a humanoid-class robot can handle repetitive intralogistics work inside an operating electronics plant. The HMND 01 Alpha wheeled platform met reported targets for throughput, uptime, and autonomous pick-and-place success while integrating with the broader software, controls, and fleet systems that factories actually use. That combination of measurable performance, industrial integration, and a coming Hannover Messe showcase makes this one of the clearest physical AI deployment signals of April.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/siemens-humanoid-hmnd01-factory-physical-ai-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Siemens" />
    <category term="Humanoid" />
    <category term="HMND 01" />
    <category term="NVIDIA" />
    <category term="physical AI" />
    <category term="factory automation" />
    <category term="industrial robotics" />
    <category term="humanoid robots" />
  </entry>
  <entry>
    <title>Physical Intelligence&apos;s π0.7 Shows the First Real Glimpse of a General Robot Brain</title>
    <link href="https://biped.news/article/physical-intelligence-pi07-robot-brain-2026" />
    <id>biped-article-physical-intelligence-pi07-robot-brain-2026</id>
    <updated>2026-04-17T02:09:53.578Z</updated>
    <published>2026-04-17T02:00:00.000Z</published>
    <summary>Physical Intelligence unveiled π0.7 on April 16, describing it as a steerable robot foundation model with early compositional generalization. In company demos and TechCrunch reporting, the model handled coached air-fryer use, matched or beat specialist policies on several dexterous tasks, and transferred laundry-folding behavior to a new dual-arm robot with no task-specific data on that machine. The announcement matters because the bottleneck in robotics is no longer only hardware, it is whether a single model can learn reusable physical skills instead of forcing teams to retrain for each new object, tool, and workflow.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/physical-intelligence-pi07-robot-brain-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Physical Intelligence" />
    <category term="pi0.7" />
    <category term="pi-zero" />
    <category term="robot foundation model" />
    <category term="embodied AI" />
    <category term="physical AI" />
    <category term="manipulation" />
    <category term="Sergey Levine" />
    <category term="robotics software" />
  </entry>
  <entry>
    <title>RoboTera: Tsinghua&apos;s Humanoid Bet Is Now a $1.4 Billion Unicorn</title>
    <link href="https://biped.news/article/robotera-company-profile" />
    <id>biped-article-robotera-company-profile</id>
    <updated>2026-04-16T15:17:51.000Z</updated>
    <published>2026-04-16T21:00:00.000Z</published>
    <summary>RoboTera is a Beijing-based humanoid robot company founded in 2023 as a Tsinghua University spinout led by CEO Chen Xiaoping. The company raised over $140M across three rounds with Sequoia China leading the January 2025 Series C at a $1.4B valuation. XBot-L stands 185 cm, weighs 80 kg, and features 61 DoF with proprietary quasi-direct-drive actuators and 7-DoF tactile hands rated at 10 kg payload per hand. The company is conducting factory trial deployments in Chinese automotive and electronics manufacturing, competing against Unitree, Agibot, and UBTECH.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/robotera-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="RoboTera" />
    <category term="XBot-L" />
    <category term="humanoid robot" />
    <category term="China robotics" />
    <category term="Tsinghua" />
    <category term="Sequoia China" />
    <category term="companies" />
  </entry>
  <entry>
    <title>EngineAI: The Shenzhen Startup Trying to Turn Humanoids Into Real Products Fast</title>
    <link href="https://biped.news/article/engineai-company-profile" />
    <id>biped-article-engineai-company-profile</id>
    <updated>2026-04-16T16:42:54.149Z</updated>
    <published>2026-04-16T19:00:00.000Z</published>
    <summary>EngineAI was founded in Shenzhen in October 2023 by Zhao Tongyang and has moved with unusual speed through the humanoid robotics market. In less than three years it introduced multiple robot platforms, including the compact PM01, the full-size SE01, and the newer T800, while emphasizing self-developed joints, open developer access, and highly visible locomotion demos. Publicly reported funding rounds in 2025 and 2026 pushed the company past unicorn valuation territory, and a deployment partnership with Duolun Technology suggests that EngineAI is trying to move beyond lab demos into commercial rollout. The central question now is whether its fast product cadence can turn into durable manufacturing and customer traction.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/engineai-company-profile-hero.png" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="EngineAI" />
    <category term="humanoid robots" />
    <category term="PM01" />
    <category term="SE01" />
    <category term="T800" />
    <category term="physical AI" />
    <category term="Shenzhen robotics" />
    <category term="embodied AI" />
  </entry>
  <entry>
    <title>Physical AI Goes Two Ways at Once: UniX AI Panther in Homes, Accenture Backs GRID for Factories</title>
    <link href="https://biped.news/article/physical-ai-home-factory-april-2026" />
    <id>biped-article-physical-ai-home-factory-april-2026</id>
    <updated>2026-04-16T08:55:25.950Z</updated>
    <published>2026-04-15T14:00:00.000Z</published>
    <summary>On April 15, 2026, UniX AI announced commercial home deliveries of its Panther humanoid robot in Suzhou, China, with 34-DOF arms, 8-16 hour battery, and tasks including cooking, cleaning, and laundry via imitation learning. The same day, Accenture announced a strategic investment in General Robotics to scale the GRID platform, an enterprise orchestration layer for multi-vendor robot fleets built on NVIDIA Omniverse and supporting 40+ pre-trained AI skills with sub-15-minute deployment.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/physical-ai-home-factory-april-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="UniX AI" />
    <category term="Panther" />
    <category term="Accenture" />
    <category term="General Robotics" />
    <category term="GRID" />
    <category term="home robot" />
    <category term="physical AI" />
    <category term="enterprise robotics" />
    <category term="elder care" />
    <category term="NVIDIA" />
  </entry>
  <entry>
    <title>AGIBOT G2 Achieves 99.9% Success Rate in World&apos;s First Embodied AI Electronics Factory Deployment</title>
    <link href="https://biped.news/article/agibot-g2-electronics-manufacturing-2026" />
    <id>biped-article-agibot-g2-electronics-manufacturing-2026</id>
    <updated>2026-04-16T08:51:48.969Z</updated>
    <published>2026-04-15T06:00:00.000Z</published>
    <summary>AGIBOT and Longcheer Technology announced on April 15, 2026 that AGIBOT G2 humanoid robots are operating on live tablet production lines at Longcheer&apos;s Nanchang facility, handling MMIT testing station tasks with 99.9% success at 310 UPH. Full-line integration took 36 hours; the pilot-to-production timeline was 4 months. AGIBOT targets 100 G2 units at Longcheer by Q3 2026. The deployment is described as the world&apos;s first large-scale embodied AI deployment in consumer electronics precision manufacturing.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/agibot-g2-electronics-manufacturing-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="AGIBOT" />
    <category term="G2" />
    <category term="Longcheer" />
    <category term="electronics manufacturing" />
    <category term="embodied AI" />
    <category term="China robotics" />
    <category term="factory deployment" />
    <category term="precision manufacturing" />
  </entry>
  <entry>
    <title>Unitree R1 Goes Global: A $4,900 Humanoid Robot Now on AliExpress</title>
    <link href="https://biped.news/article/unitree-r1-global-sale-2026" />
    <id>biped-article-unitree-r1-global-sale-2026</id>
    <updated>2026-04-16T08:48:02.621Z</updated>
    <published>2026-04-13T12:00:00.000Z</published>
    <summary>Unitree Robotics listed its R1 humanoid robot for global sale via AliExpress and its official shop in April 2026, with the base R1 AIR starting at $4,900 and the standard R1 at $5,900. The compact 123 cm robot has 20-26 DOF (depending on variant), a multimodal AI stack, 1-hour quick-swap battery, and WiFi 6. The R1 EDU adds optional NVIDIA Jetson Orin for AI research use. Deliveries begin May-June 2026 with international shipping adding $300-$1,200.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/unitree-r1-global-sale-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Unitree" />
    <category term="R1" />
    <category term="AliExpress" />
    <category term="consumer robot" />
    <category term="humanoid robot" />
    <category term="China robotics" />
    <category term="NVIDIA Jetson" />
    <category term="affordable humanoid" />
  </entry>
  <entry>
    <title>China&apos;s Wall-Climbing Embodied AI Robot Enters Service at Chemical Tank Sites</title>
    <link href="https://biped.news/article/china-wall-climbing-robot-hazardous-2026" />
    <id>biped-article-china-wall-climbing-robot-hazardous-2026</id>
    <updated>2026-04-16T08:44:06.615Z</updated>
    <published>2026-04-12T12:00:00.000Z</published>
    <summary>A wall-climbing embodied AI robot with 15 degrees of freedom across two arms entered service at a chemical storage tank construction site in Tangshan, Hebei Province in April 2026. The system uses electromagnetic adhesion, operates on tethered power for continuous 24/7 use, and is controlled via VR goggles with an AI stack trained on 100,000 hours of data. China&apos;s state media describes it as the country&apos;s first embodied intelligent robot deployed to replace humans in hazardous settings.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/china-wall-climbing-robot-hazardous-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="China robotics" />
    <category term="embodied AI" />
    <category term="industrial robot" />
    <category term="wall climbing" />
    <category term="hazardous work" />
    <category term="dual arm" />
    <category term="welding robot" />
    <category term="chemical plant" />
  </entry>
  <entry>
    <title>Beijing Humanoid Half-Marathon 2026: 100+ Robots Complete Full-Course Test Run</title>
    <link href="https://biped.news/article/beijing-robot-marathon-2026-test" />
    <id>biped-article-beijing-robot-marathon-2026-test</id>
    <updated>2026-04-16T08:41:10.200Z</updated>
    <published>2026-04-11T12:00:00.000Z</published>
    <summary>Beijing&apos;s E-Town district hosted a full-scale overnight test of the 2026 humanoid robot half-marathon on April 11-12, with 100+ robots from 70+ teams completing the 21.0975 km course. The defending champion Tiangong Ultra confirmed its 2:40:42 time. The April 19 race includes international teams for the first time, with Technical University of Munich entering with a Chinese-made platform.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/beijing-robot-marathon-2026-test-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Beijing" />
    <category term="half-marathon" />
    <category term="Tiangong Ultra" />
    <category term="bipedal locomotion" />
    <category term="autonomous navigation" />
    <category term="China robotics" />
    <category term="endurance test" />
  </entry>
  <entry>
    <title>Kia Extends Boston Dynamics Atlas Rollout to a Second Georgia Plant</title>
    <link href="https://biped.news/article/kia-atlas-georgia-factory-2026" />
    <id>biped-article-kia-atlas-georgia-factory-2026</id>
    <updated>2026-04-10T02:10:31.605Z</updated>
    <published>2026-04-10T04:00:00.000Z</published>
    <summary>Kia&apos;s April 9 investor presentation tied humanoid robotics directly to its manufacturing roadmap, with Atlas deployment scheduled for HMGMA in 2028 and Kia AutoLand Georgia in 2029. That matters because Boston Dynamics is no longer selling Atlas as a distant concept. It is being threaded into the production plans of multiple plants inside one of the world&apos;s largest automotive groups. The signal for the robotics industry is clear: the next competitive edge is not the best demo video, it is repeatable factory integration across a parent company&apos;s network.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/kia-atlas-georgia-factory-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Kia" />
    <category term="Boston Dynamics" />
    <category term="Atlas" />
    <category term="Hyundai Motor Group" />
    <category term="Georgia" />
    <category term="HMGMA" />
    <category term="AutoLand Georgia" />
    <category term="Humanoid Robots" />
    <category term="Manufacturing" />
    <category term="Physical AI" />
  </entry>
  <entry>
    <title>UBTECH Robots Enter Auto Supply Chain via Honda Trading</title>
    <link href="https://biped.news/article/ubtech-honda-trading-automotive-supply-chain-2026" />
    <id>biped-article-ubtech-honda-trading-automotive-supply-chain-2026</id>
    <updated>2026-04-09T02:08:44.482Z</updated>
    <published>2026-04-09T02:08:44.482Z</published>
    <summary>UBTECH and Honda Trading (China) will explore humanoid robot and autonomous logistics deployments across manufacturing and warehousing scenarios. The timing is notable after UBTECH reported 1,079 full-size humanoid deliveries in 2025 and set a 5,000-unit goal for 2026. For the wider market, the partnership is a sign that commercialization is shifting from isolated demos toward channel access, systems integration, and repeatable supply-chain deployments.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/ubtech-honda-trading-automotive-supply-chain-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="UBTECH" />
    <category term="Honda Trading" />
    <category term="Walker S" />
    <category term="Humanoid Robots" />
    <category term="Automotive Supply Chain" />
    <category term="Warehouse Automation" />
    <category term="Embodied AI" />
    <category term="Industry News" />
    <category term="China" />
  </entry>
  <entry>
    <title>UBTECH Is Paying $18M for a Robotics Scientist. Really.</title>
    <link href="https://biped.news/article/ubtech-18-million-chief-scientist-2026" />
    <id>biped-article-ubtech-18-million-chief-scientist-2026</id>
    <updated>2026-04-08T11:32:10.768Z</updated>
    <published>2026-04-08T12:00:00.000Z</published>
    <summary>UBTECH has opened one of the most aggressive recruiting campaigns yet seen in humanoid robotics, offering up to 124 million yuan, roughly $18 million, for a chief scientist of embodied intelligence. The move comes as the company says 2025 humanoid orders topped 1.4 billion yuan, production capacity is expected to exceed 10,000 units in 2026, and Airbus has begun early concept testing of Walker S2 robots. The article argues that this is not just a flashy salary story. It is evidence that talent, software infrastructure, and manufacturing systems are becoming the decisive battlegrounds in commercial humanoid robotics.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/ubtech-18-million-chief-scientist-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="UBTECH" />
    <category term="Walker S2" />
    <category term="Humanoid Robots" />
    <category term="Embodied AI" />
    <category term="Industry News" />
    <category term="Siemens" />
    <category term="Airbus" />
    <category term="China" />
  </entry>
  <entry>
    <title>The AI Architecture That Makes Humanoid Robots Work</title>
    <link href="https://biped.news/article/vision-language-action-models-explained" />
    <id>biped-article-vision-language-action-models-explained</id>
    <updated>2026-04-07T13:48:24.483Z</updated>
    <published>2026-04-07T13:48:24.483Z</published>
    <summary>VLAs merge large language model pre-training with robot control policy learning, enabling humanoid robots to generalize to new tasks and objects from minimal demonstration data. This explainer covers VLA architecture, OpenVLA and pi-zero as reference implementations, the competitive landscape among foundation model providers, commercial deployment status, and the remaining engineering challenges before VLAs reach unstructured environments.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/vision-language-action-models-explained-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="VLA" />
    <category term="vision-language-action" />
    <category term="embodied AI" />
    <category term="foundation models" />
    <category term="OpenVLA" />
    <category term="Physical Intelligence" />
    <category term="NVIDIA GR00T" />
    <category term="robot AI" />
    <category term="humanoid robots" />
    <category term="robot control" />
  </entry>
  <entry>
    <title>Toyota Deployed 7 Humanoid Robots on Its Assembly Line</title>
    <link href="https://biped.news/article/toyota-agility-digit-woodstock-2026" />
    <id>biped-article-toyota-agility-digit-woodstock-2026</id>
    <updated>2026-04-07T13:17:13.431Z</updated>
    <published>2026-04-07T13:17:13.431Z</published>
    <summary>Toyota Motor Manufacturing Canada has deployed seven Agility Digit humanoid robots at its Woodstock, Ontario plant under a commercial Robots-as-a-Service agreement, marking the first commercial humanoid deployment at a Toyota facility. The robots handle tote loading and parts transport on the RAV4 production line, addressing ergonomic strain in one of the highest-injury-risk task categories in automotive manufacturing. The RaaS structure, managed via Agility&apos;s Arc fleet platform, converts capital expenditure to operating costs and includes software upgrades and maintenance, making it a replicable model for the broader auto industry.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/toyota-agility-digit-woodstock-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Agility" />
    <category term="Digit" />
    <category term="Toyota" />
    <category term="TMMC" />
    <category term="RaaS" />
    <category term="automotive" />
    <category term="manufacturing" />
    <category term="humanoid robots" />
    <category term="Woodstock" />
    <category term="Ontario" />
  </entry>
  <entry>
    <title>How Robots Move Like Humans: Whole-Body Control</title>
    <link href="https://biped.news/article/whole-body-control-how-robots-move-like-humans" />
    <id>biped-article-whole-body-control-how-robots-move-like-humans</id>
    <updated>2026-04-07T12:56:09.663Z</updated>
    <published>2026-04-07T12:56:09.663Z</published>
    <summary>A technical explainer on whole-body control (WBC) in humanoid robotics: how QP-based optimization coordinates multiple simultaneous tasks, the role of task prioritization and contact constraints, how it compares to older approaches like ZMP control and newer RL-based policies, and how companies including Figure AI, Agility Robotics, Boston Dynamics, and Unitree are deploying it in commercial hardware.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/whole-body-control-how-robots-move-like-humans-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="whole-body control" />
    <category term="WBC" />
    <category term="humanoid robots" />
    <category term="locomotion" />
    <category term="motion control" />
    <category term="quadratic programming" />
    <category term="reinforcement learning" />
    <category term="Figure AI" />
    <category term="Boston Dynamics" />
    <category term="Agility Robotics" />
    <category term="Unitree" />
  </entry>
  <entry>
    <title>China Ships 1 Humanoid Robot Every 30 Minutes Now</title>
    <link href="https://biped.news/article/china-humanoid-production-surge-2026" />
    <id>biped-article-china-humanoid-production-surge-2026</id>
    <updated>2026-04-01T02:09:51.234Z</updated>
    <published>2026-04-01T02:09:51.234Z</published>
    <summary>AGIBOT reached 10,000 humanoid robots shipped on March 28, 2026, shipping the last 5,000 units in just three months. Days later, a factory in Foshan, Guangdong, came online producing one robot every 30 minutes, with a 10,000-unit annual capacity. The US market leaders — Figure AI, Agility, Tesla Optimus — had shipped roughly 150 units each at comparable milestones. The production gap reflects a fundamentally different supply chain maturity, national policy backing, and unit economics: AGIBOT&apos;s A3 humanoid is priced at $45,000, compared to $100,000-plus for Western premium alternatives.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/china-humanoid-production-surge-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="AGIBOT" />
    <category term="China" />
    <category term="humanoid robots" />
    <category term="manufacturing" />
    <category term="production" />
    <category term="Leju Robotics" />
    <category term="Foshan" />
    <category term="Unitree" />
    <category term="Physical AI" />
    <category term="Industry News" />
  </entry>
  <entry>
    <title>Mobileye&apos;s $900M Bet: Robot That Learns From One Video</title>
    <link href="https://biped.news/article/mentee-robotics-mobileye-acquisition" />
    <id>biped-article-mentee-robotics-mobileye-acquisition</id>
    <updated>2026-04-01T01:26:05.233Z</updated>
    <published>2026-04-01T01:26:05.233Z</published>
    <summary>Mentee Robotics, founded in 2021 by Amnon Shashua, Lior Wolf, and Shai Shalev-Shwartz in Herzliya, Israel, developed MenteeBot, an AI-first humanoid robot capable of learning new tasks from a single video demonstration. The company raised $40 million before Mobileye agreed to acquire it for $900 million in January 2026. MenteeBot V3 stands 175 cm, weighs 70 kg, carries 25 kg payloads, and runs dual NVIDIA OrinX GPUs with camera-only sensing.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/mentee-robotics-mobileye-acquisition-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Mentee Robotics" />
    <category term="MenteeBot" />
    <category term="Mobileye" />
    <category term="humanoid robot" />
    <category term="Physical AI" />
    <category term="Sim2Real" />
    <category term="company acquisition" />
    <category term="companies" />
  </entry>
  <entry>
    <title>Tesla Is Deploying 50,000 Optimus Robots This Year</title>
    <link href="https://biped.news/article/tesla-optimus-factory-deployment-2026" />
    <id>biped-article-tesla-optimus-factory-deployment-2026</id>
    <updated>2026-03-31T02:06:19.128Z</updated>
    <published>2026-03-31T02:06:19.128Z</published>
    <summary>Tesla&apos;s 2026 Optimus strategy centers on factory-first deployment: build the robots, deploy them internally, collect operational data at scale, and improve Gen 3 before any commercial sale. Over 1,200 Gen 2 units are running in Tesla facilities as of January 2026, with year-end targets of 50,000+ units. The $20 billion capex plan funds the whole program. Gen 3, with 22-DOF hands and 50 total actuators, enters low-volume production in summer 2026. Labor economists and union advocates are increasingly focused on the $2/hour effective labor cost and what it means for manufacturing employment at scale.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/tesla-optimus-factory-deployment-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Tesla" />
    <category term="Optimus" />
    <category term="humanoid robots" />
    <category term="factory automation" />
    <category term="physical AI" />
    <category term="labor" />
    <category term="Gen 3" />
    <category term="manufacturing" />
    <category term="deployment" />
  </entry>
  <entry>
    <title>Why Every Humanoid Robot Uses the Same Motor Now</title>
    <link href="https://biped.news/article/humanoid-robot-actuators-explained" />
    <id>biped-article-humanoid-robot-actuators-explained</id>
    <updated>2026-03-31T01:24:53.406Z</updated>
    <published>2026-03-31T01:24:53.406Z</published>
    <summary>Quasi-direct drive (QDD) actuators, combining high-torque BLDC motors with low-ratio cycloidal gearboxes and integrated force sensing, have become the standard joint architecture for commercial humanoid robots in 2024-2026. This explainer covers how they work, why they beat hydraulic and high-ratio geared alternatives, who makes them, and what the cost trajectory means for humanoid robot pricing over the next three years.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/humanoid-robot-actuators-explained-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="actuators" />
    <category term="quasi-direct drive" />
    <category term="QDD" />
    <category term="electric motors" />
    <category term="force control" />
    <category term="Figure AI" />
    <category term="Tesla Optimus" />
    <category term="Boston Dynamics" />
    <category term="humanoid robots" />
    <category term="cycloidal gearbox" />
  </entry>
  <entry>
    <title>Amazon Bought a Robotics Startup 59 Days After Launch</title>
    <link href="https://biped.news/article/amazon-acquires-fauna-robotics-2026" />
    <id>biped-article-amazon-acquires-fauna-robotics-2026</id>
    <updated>2026-03-30T13:34:21.736Z</updated>
    <published>2026-03-30T13:34:21.736Z</published>
    <summary>Amazon confirmed the acquisition of Fauna Robotics on March 24, 2026, just 59 days after the startup shipped its first humanoid robot, Sprout, to developers. Sprout stands 107 cm tall, weighs 22.7 kg, and runs on NVIDIA Jetson AGX Orin hardware with 29 degrees of freedom. The roughly 50-person Fauna team, including co-founders Rob Cochran (ex-Meta/CTRL-labs) and Josh Merel (ex-Google DeepMind), joins Amazon&apos;s Personal Robotics Group in New York City. The deal was Amazon&apos;s second robotics acquisition in March 2026, following the Rivr stair-climbing delivery robot purchase five days earlier. Terms were not disclosed; Fauna had raised approximately 0 million from Kleiner Perkins.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/amazon-acquires-fauna-robotics-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Amazon" />
    <category term="Fauna Robotics" />
    <category term="Sprout" />
    <category term="consumer robotics" />
    <category term="humanoid robots" />
    <category term="acquisition" />
    <category term="Google DeepMind" />
    <category term="NVIDIA Jetson" />
    <category term="Personal Robotics" />
    <category term="Industry News" />
  </entry>
  <entry>
    <title>Bipartisan Bill Would Ban Chinese Robots From U.S. Gov</title>
    <link href="https://biped.news/article/american-security-robotics-act-2026" />
    <id>biped-article-american-security-robotics-act-2026</id>
    <updated>2026-03-27T02:06:21.010Z</updated>
    <published>2026-03-27T02:06:21.010Z</published>
    <summary>The American Security Robotics Act, introduced March 26, 2026, would ban federal agencies from buying or operating Chinese-made humanoid robots and unmanned ground vehicles. Sponsored by Senators Cotton and Schumer with a House companion from Rep. Stefanik, the bipartisan bill targets companies like Unitree, UBTECH, and AgiBot while exempting U.S. manufacturers. Analysts expect it to follow the DJI drone ban playbook: federal restriction first, private-sector pressure to follow. The most likely vehicle for passage is the FY2027 National Defense Authorization Act.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/american-security-robotics-act-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Policy" />
    <category term="American Security Robotics Act" />
    <category term="Unitree" />
    <category term="UBTECH" />
    <category term="AgiBot" />
    <category term="China" />
    <category term="Congress" />
    <category term="Tom Cotton" />
    <category term="Chuck Schumer" />
    <category term="National Security" />
    <category term="Federal Procurement" />
    <category term="Humanoid Robots" />
  </entry>
  <entry>
    <title>A Humanoid Robot Just Walked Into the White House</title>
    <link href="https://biped.news/article/figure-03-white-house-melania-2026" />
    <id>biped-article-figure-03-white-house-melania-2026</id>
    <updated>2026-03-26T02:06:14.820Z</updated>
    <published>2026-03-26T02:06:14.820Z</published>
    <summary>Figure AI&apos;s Figure 03 made history on March 25, 2026, becoming the first humanoid robot to enter the White House, appearing at Melania Trump&apos;s &apos;Fostering the Future Together&apos; global education summit. The 135-pound robot walked autonomously beside the First Lady, delivered a prepared speech in 11 languages, and exited the East Room without incident. The appearance carries significant implications for Figure AI&apos;s valuation trajectory, the U.S.-China robotics competition narrative, and the emerging consumer humanoid market — where Figure 03 is targeting a ~$25,000 price point for a late-2026 launch.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/figure-03-white-house-melania-2026-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Figure AI" />
    <category term="Figure 03" />
    <category term="White House" />
    <category term="Melania Trump" />
    <category term="Helix" />
    <category term="humanoid robots" />
    <category term="physical AI" />
    <category term="AI policy" />
    <category term="consumer robotics" />
    <category term="Brett Adcock" />
  </entry>
  <entry>
    <title>How Humanoid Robots Learn: Sim, Imitation, Real Data</title>
    <link href="https://biped.news/article/how-humanoid-robots-are-trained-sim-to-real-imitation-learning" />
    <id>biped-article-how-humanoid-robots-are-trained-sim-to-real-imitation-learning</id>
    <updated>2026-03-24T01:27:30.036Z</updated>
    <published>2026-03-24T01:27:30.036Z</published>
    <summary>This explainer breaks down the AI training pipeline for humanoid robots: why simulation is essential, how domain randomization closes the sim-to-real gap, what imitation learning and behavioral cloning involve, how companies like Figure AI, Physical Intelligence, and Tesla collect training data, and what the combined sim-plus-real training playbook looks like for commercial deployments in 2025-2026.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/how-humanoid-robots-are-trained-sim-to-real-imitation-learning-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="sim-to-real" />
    <category term="imitation learning" />
    <category term="robot training" />
    <category term="reinforcement learning" />
    <category term="Physical Intelligence" />
    <category term="Figure AI" />
    <category term="Tesla Optimus" />
    <category term="AI training" />
    <category term="humanoid robots" />
    <category term="domain randomization" />
  </entry>
  <entry>
    <title>Sunday Hit $1.15B Building a Robot That Does Your Chores</title>
    <link href="https://biped.news/article/sunday-company-profile" />
    <id>biped-article-sunday-company-profile</id>
    <updated>2026-03-24T00:00:00.000Z</updated>
    <published>2026-03-24T00:00:00.000Z</published>
    <summary>Sunday, founded in April 2024 by Berkeley and Columbia robotics researchers, is building Memo — a wheeled humanoid robot designed to handle real household tasks. The ACT-1 foundation model, trained on human activity data from 2,000+ Skill Capture Gloves, enables Memo to fold laundry, load dishwashers, and make espresso. A $165M Series B at $1.15B valuation positions Sunday as the leading home robotics contender heading into 2026.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/sunday-company-profile-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Sunday" />
    <category term="Memo" />
    <category term="home robot" />
    <category term="humanoid robot" />
    <category term="ACT-1" />
    <category term="Diffusion Policy" />
    <category term="household AI" />
    <category term="robotics" />
    <category term="companies" />
  </entry>
  <entry>
    <title>Congress Held Its First Hearing on Chinese Robots</title>
    <link href="https://biped.news/article/us-congress-chinese-humanoid-robots-security-hearing" />
    <id>biped-article-us-congress-chinese-humanoid-robots-security-hearing</id>
    <updated>2026-03-23T02:08:10.290Z</updated>
    <published>2026-03-23T02:08:10.290Z</published>
    <summary>The House Committee on Homeland Security convened its first hearing on the national security implications of Chinese humanoid robot manufacturers on March 18, 2026. Executives from Scale AI and Boston Dynamics testified that Chinese firms like Unitree Robotics now dominate global shipments at roughly 90% market share, with production targets exceeding 100,000 units for 2026. Witnesses recommended expanded AI chip export controls, federal procurement bans, and formal investigations into Chinese robotics companies.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/us-congress-chinese-humanoid-robots-security-hearing-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Congress" />
    <category term="China" />
    <category term="National Security" />
    <category term="Unitree" />
    <category term="Scale AI" />
    <category term="Boston Dynamics" />
    <category term="Export Controls" />
    <category term="Policy" />
    <category term="Humanoid Robots" />
    <category term="Trade War" />
  </entry>
  <entry>
    <title>Amazon Bought the Robot That Climbs Stairs to Deliver</title>
    <link href="https://biped.news/article/amazon-acquires-rivr-doorstep-delivery-robot" />
    <id>biped-article-amazon-acquires-rivr-doorstep-delivery-robot</id>
    <updated>2026-03-22T02:05:49.868Z</updated>
    <published>2026-03-22T02:05:49.868Z</published>
    <summary>Amazon announced its acquisition of Rivr, a Zurich-based robotics startup building hybrid wheeled-legged quadruped robots for last-mile doorstep delivery. Rivr&apos;s robots can roll at 15 km/h on flat ground and climb stairs autonomously, solving the &apos;last 100 yards&apos; problem that defeated Amazon&apos;s earlier Scout program. The company, originally called Swiss-Mile, spun out of ETH Zurich and had raised $25 million before the acquisition, with Amazon&apos;s own Industrial Innovation Fund among its investors.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/amazon-acquires-rivr-doorstep-delivery-robot-hero.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Amazon" />
    <category term="Rivr" />
    <category term="Swiss-Mile" />
    <category term="delivery robots" />
    <category term="last-mile delivery" />
    <category term="quadruped robots" />
    <category term="autonomous delivery" />
    <category term="acquisition" />
    <category term="ETH Zurich" />
    <category term="logistics" />
  </entry>
  <entry>
    <title>Unitree Files $610M IPO as First Humanoid Robot Stock</title>
    <link href="https://biped.news/article/unitree-ipo-shanghai-610m-2026" />
    <id>biped-article-unitree-ipo-shanghai-610m-2026</id>
    <updated>2026-03-21T02:06:11.955Z</updated>
    <published>2026-03-21T02:06:11.955Z</published>
    <summary>Unitree Robotics filed for a 4.2 billion yuan ($610 million) IPO on the Shanghai Stock Exchange&apos;s STAR Market on March 20, 2026. The Hangzhou-based company shipped over 5,500 humanoid robots in 2025, generating 335% revenue growth year-over-year. Half the IPO proceeds will fund AI model development, while a quarter will expand manufacturing capacity. Unitree plans to ship 20,000 humanoid units in 2026, positioning itself as the world&apos;s highest-volume humanoid robot maker.</summary>
    <link rel="enclosure" type="image/jpeg" href="https://storage.googleapis.com/biped-f5a14-media/article-images/unitree-ipo-shanghai-610m-2026-1.jpg" />
    <author>
      <name>Biped News</name>
    </author>
    <category term="Unitree Robotics" />
    <category term="IPO" />
    <category term="Shanghai Stock Exchange" />
    <category term="STAR Market" />
    <category term="humanoid robots" />
    <category term="G1" />
    <category term="H1" />
    <category term="China" />
    <category term="robotics funding" />
    <category term="Physical AI" />
  </entry>
</feed>