Skip to content
Nov 10Tue
  1. 雷科技52

    Musk says Optimus humanoid robot will be rental-only, not for sale, with initial cost of about $30,000

    Speaking at the Baron Investment Conference, Musk said Tesla's Optimus humanoid robot will likely be offered through leasing rather than direct sale. He disclosed an initial per-unit cost of about $30,000, dropping to about $20,000 after mass production; the Fremont factory has begun early mass production, with the line built for a scale of 1 million units per year, and external deliveries not expected until late 2027 at the earliest. The article also analyzes that the leasing model can generate ongoing rental and software maintenance revenue and let Tesla control robot operating data.

  2. 机器人全球资讯 · 微信公众号42

    Wang Xingxing: Unitree to focus on robot brain technology, seeks global cooperation

    Unitree Technology founder Wang Xingxing said in an interview at the World Internet Conference Wuzhen Summit that Unitree will focus on conquering robot brain technology in the future and seek cooperation globally, and is optimistic about physical AI technology. Unitree has previously defined three major strategic directions: improving large-scale robot mass-production capability to meet demand at the tens-of-millions and even hundreds-of-millions scale; advancing the deployment of physical AI small models suited to its own hardware through both in-house development and partnerships; and building out ultra-large-scale computing clusters.

  3. OFweek Robotics42

    Unitree Robotics shares rebound for second straight session: Apple uses G1 to validate HALO paper, Kunlun Manufacturing partnership still in public notice period

    On November 9, Unitree Robotics shares rose more than 15% intraday, with market capitalization briefly topping 300 billion yuan, before closing up 9.04% at 716 yuan. An Apple research team used a modified Unitree G1 to validate the HALO paper, completing 360 single-task tests across six real-world environments without prior mapping, with an average success rate of 91%. Unitree and Kunlun Manufacturing plan to establish a new joint venture to conduct industrial embodied intelligent robot solution business; as of the time of verification, the plan was still in its public notice period.

Nov 9Mon
  1. OFweek Robotics48

    The robotics bubble is bursting: IPOs are closed and backdoor listings are blocked

    The IPO channel for the humanoid robotics industry is rapidly narrowing: the A-share and Hong Kong listing windows are essentially closed, and the backdoor listing route is also difficult to pursue. After Zhiyuan Robotics acquired control of Shangwei New Materials, Shangwei New Materials announced in October 2025 that Zhiyuan has no backdoor listing plan within the next 3 years. The closure of exit channels has sharply increased financing difficulty; over the past two months, quoted prices for secondary sales of old shares in the embodied intelligence sector have generally been only 65% to 80% of the previous round's valuation. In the first half of 2026, China's embodied intelligence-themed investment and financing reached 172 deals totaling 109.174 billion yuan, but financing activity fell rapidly after September.

  2. OFweek Robotics60

    China-US Embodied Brain Showdown: Bodies Have Won, Brains Are Still Racing

    The Chaoyong AI editorial team compares China-US paths in embodied intelligence, arguing that China leads in complete-machine cost, supply chain, and shipment scale, while US companies enter the physical world through general-purpose models. The article cites cases including Qianxun Intelligence's Moz1 maintaining a plug-in success rate above 99% on a CATL PACK production line, the HOST framework from Zhiyuan (Variable) achieving an average success rate of 62% across 50 desktop tasks, and Physical Intelligence negotiating a funding round of over $1 billion. It proposes that cross-embodiment and cross-task success rates, customer repurchases, and whether model revenue can be independent of hardware delivery are the key criteria for judging the value of brain companies.

Nov 6Fri
  1. 人形大讲堂 · 微信公众号38

    Apple's J595 desktop robot prototype revealed: 9-inch screen plus robotic arm, target price around $1,000

    The Information reports that Apple's prototype codenamed J595 is a 9-inch screen plus a robotic arm; the arm moves the screen to face whoever is speaking in the room, the unit stays on the desktop, and the target price is around $1,000. Apple has neither announced nor confirmed that the machine exists. Most of the hardware was ready about two years ago, but software has held up its launch; the team has kept cutting features over the past year, narrowing the first-generation concept to FaceTime calls and turning on lights. People familiar with the matter say Apple has about 10 people who truly understand robot R&D, and the software team has about 300 people.

Nov 5Thu
  1. 亿欧网 · 微信公众号62

    Wang Xingxing vs. Zhi Hui Jun: Unitree adopts general models while AgiBot develops in-house models for production lines

    The article compares Unitree and AgiBot's divergent approaches to embodied intelligence: Unitree integrates GPT-6 Astra into the G1 as an "external brain," with Wang Xingxing judging that general models will be able to handle planning after another one or two generations; AgiBot only lets Astra serve as System2, while System1 and the 100 to 1000Hz force-control closed loop are developed in-house, and GO-3 Preview tops three leaderboards: RoboDojo, RoboTwin, and RoboCosieum.

  2. 钛媒体38

    GPT-6 Astra shakes up the embodied AI brain industry: risks and opportunities for Unitree, AgiBot, Qianxun and other vendors

    GPT-6 Astra, a general-purpose large model, drove a several-hundred-dollar robotic arm to complete tasks such as drawing the Golden Gate Bridge without specialized training. Unitree Technology used it to autonomously identify a palette, change paints, and clean brushes. SenseTime's evaluation says it can handle robot manipulation and navigation tasks that specialized models such as VLA and WAM perform.

  3. OFweek Robotics42

    Unitree Technology Bets on 2027: The Humanoid Robot Tipping Point and the First Round of Consolidation

    At the Wuzhen Summit, Unitree Technology founder Wang Xingxing proposed that 2027 is likely to become a tipping point for the robotics industry, defined as robots completing 80% of tasks via voice commands in 80% of unfamiliar scenarios. Unitree's 2025 revenue was 1.699 billion yuan, with 5,215 humanoid robots shipped, but in the first half of 2026 its non-recurring net profit fell 19.34% year on year; of the 4.2 billion yuan raised, 2.022 billion yuan is directed to intelligent robot model R&D.

Nov 4Wed
  1. 36Kr Embodied Intelligence38

    After TypeSafeAI's decision model Jev goes viral, why does the physical world need on-device streaming multimodal decision-making?

    TypeSafeAI released the decision model Jev, which outputs only yes/no, option numbers, or probability scores. Within weeks of launch, users exceeded one million, about one-third of Fortune 500 companies began using it, and its valuation rose from about $200 million to $7.5 billion in 24 days.

  2. Union Square Ventures · X38

    USV on the path to scaling robot intelligence

    Deploy, learn, improve, repeat. We applied this idea to robotics, in partnership with @tutorintel, and then to manufacturing, in partnership with @IsembardGroup. We invited two CEOs to share their experiences on rapidly scaling intelligence in the physical world.

Nov 3Tue
  1. OFweek Robotics60

    From storytelling to crunching the numbers: will the robot company IPO feast end?

    Unitree Technology listed on the Shanghai Stock Exchange on August 19 at 150.8 yuan per share, with its market value briefly exceeding 440 billion yuan before falling below 170 billion yuan, then recovering to 267.824 billion yuan at the November 2 close. The article argues that this fluctuation marks an industry turning point, with companies such as AgiBot and Weiyi Intelligent Manufacturing successively delaying their listings; primary-market financing continues, but valuation approaches are shifting toward performance and orders. The author notes that robot IPOs have not cooled off; rather, the industry has become more rational, and only companies that can achieve sustainable gross margins in a single scenario are qualified to enter the next round.

  2. 雷科技42

    Embodied AI Brains in Fierce Competition: 9 Leading Companies Pursue 4 Technical Routes, Brain-Inspired Architecture Emerges

    Zhizhi Pingfang released NeuroVLA, a brain-inspired embodied intelligence system designed with a three-layer structure of cortex, cerebellum, and spinal cord; the spinal cord layer is driven by a spiking neural network, consumes 0.4 watts, and completes a reflex within 20 milliseconds after a collision. Qianxun Intelligence's Spirit v1.6 ranked first overall in the RoboArena evaluation, surpassing NVIDIA Cosmos3 and Physical Intelligence's π0.5. Leading companies have formed four routes—world-model-native, VLA and world model fusion, brain-inspired architecture, and unified model—but there is still no unified classification standard.

Nov 2Mon
  1. 机器人daily · 微信公众号42

    Agibot Proposes Worken as a Measure of Effective Intelligent Work Hours: From Horsepower and Kilowatt-Hours to Tokens, Humanoid Robots Gain a New Ruler

    At its A.I.DAY event on October 29, Agibot proposed Worken (effective intelligent work hours) as a unit of measurement. The formula is the number of active robots × autonomous working duration × output efficiency ratio. It counts only robots that are online, able to execute tasks, and meet standards; it excludes teleoperation and charging/battery-swap periods from continuous autonomous working time; and it uses the qualified output of a standard worker per unit time as one equivalent.

  2. OFweek Robotics58

    Trading money for time: embodied AI companies step into investing

    Embodied AI companies are investing heavily in the robotics supply chain as industrial capital. Unitree Robotics and Sequoia China established the "Shanyu Innovation Program," and Zhiyuan Robotics appeared in 47 investment and financing events from August 2024 to October 2026, most of them directed upstream of its own complete machines. The article argues that three overlapping uncertainties—unconverged technology routes, an unformed supply chain, and unvalidated scenarios—are leading top companies to use capital to lock in supply chains, technology options, and industry positions in advance, as the industry moves from free competition toward an oligopoly rehearsal.

Nov 1Sun
  1. 钛媒体82

    The First Public Ledger for Humanoid Robots: An In-Depth Analysis on the Eve of Agility Robotics' Listing

    Agility Robotics will list on Nasdaq on November 5 via SPAC under the ticker AGLT, with a pre-money valuation of $2.5 billion and confirmed 2025 revenue of only $1.782 million, about 63.5% of which comes from related parties.

    Why it matters: Using Agility's listing filing data as a sample, it breaks down humanoid robot unit economics and RaaS contract structures, and provides transferable metric definitions.

Oct 31Sat
  1. 雷科技30

    Tencent Global Digital Ecosystem Summit: WorkBuddy Enterprise Edition deployed in over 50 industries; the challenge for agents entering enterprises lies in cost and workflows

    At the 2026 Tencent Global Digital Ecosystem Summit, Tencent disclosed that WorkBuddy has completed 9 billion task requests cumulatively since launching in March this year; its Enterprise Edition has been deployed in over 50 industries, covering 50% of China's Fortune 500 companies and 60% of China's Top 100 companies.

  2. The Humanoid Hub40

    Zhiyuan's Peng Zhihui proposes Worken metric for robot workload

    Peng Zhihui of Zhiyuan Robotics proposed using Worken (Work + Token) to measure a robot's actual workload, with the formula: number of active robots × autonomous working hours × output efficiency. It counts only time spent genuinely executing tasks and running autonomously, with output compared against a human baseline. The metric draws on cloud computing's usage-based metering approach and is intended to shift the industry's competition from shipment volumes, units off the line, and fleet size toward work actually completed.

Oct 30Fri
  1. 36Kr Embodied Intelligence20

    The Second Half for AI Startups: Where Is the Money Going?

    In a 36Kr conversation, Ji Wei of Meridian Capital and Lui Chun-yung of Invest Hong Kong judged that AI investment is shifting from competition over model parameters to application deployment and the "data flywheel." Coding, with clear feedback, is currently the largest AI application by revenue, with Cursor's annualized revenue rising from over $1 billion to about $2.6 billion within a year. Embodied intelligence is seen as the most active but also the earliest stage; Ji Wei gave a conservative expectation of a "ChatGPT moment" in 5-10 years, while Lui Chun-yung noted that screwdriving still requires engineers as a fallback and that logistics handling has not yet met stability standards, with current deployment mostly in limited scenarios such as patrol inspection.

  2. 36Kr Embodied Intelligence67

    Neolix completes nearly $400 million Series E, CEO Yu Enyuan discusses logistics vehicle TCO and overseas expansion

    Neolix announced on October 20 that it completed a Series E round of nearly $400 million, co-led by Boyu Capital and a state-level fund, with participation from Alibaba's AEF NextGen Hong Kong fund, GAC Capital, GGV Capital, Temasek's Vertex Growth and others. CEO Yu Enyuan said the original plan was to raise $200 million, and the round was ultimately oversubscribed.

  3. OFweek Robotics42

    Embodied AI's "Two Worlds of Fire and Ice": How Can Galaxea Support a 20 Billion Yuan Valuation?

    After completing two funding rounds in the first half of this year, Galaxea's valuation surpassed 20 billion yuan, with cumulative funding of nearly 5 billion yuan. However, the company's "thousand-unit" order figures have changed multiple times and have not been publicly verified. It adheres to full-stack in-house development under a "one brain, multiple forms" approach and has formed a product matrix of three complete wheeled dual-arm robot models. In June it released and open-sourced the G0.5 model, and it signed a three-year strategic cooperation with JD.com to jointly build robot forward warehouses. As industry IPO reviews tighten and related-party transactions at data collection centers come under scrutiny, embodied AI is shifting from technology demonstrations to a knockout round of delivery fulfillment.

Oct 29Thu
  1. 未来机器人网 · 微信公众号35

    Jensen Huang, Musk, and Wang Xingxing Speak in Unison: Physical AI Wave Nears as Humanoid Robots Hit an Industry Turning Point

    Nvidia's Jensen Huang, Tesla's Musk, and Unitree's Wang Xingxing have recently made aligned public judgments: physical AI is on the eve of an explosion, and humanoid robots are its most important deployment vehicle. Huang said that, relying on the Cosmos world model and the Isaac simulation platform, AI factory deployment cycles have been compressed to 14 days; Musk plans for Optimus to be supplied externally by the end of 2027 after resolving issues such as electromechanical reliability, and says early production has already begun at the Fremont factory.

  2. 机器之心 · 微信公众号42

    What kind of data does spatial intelligence need? Interview with Yingjie Technology and CityU Associate Professor Ma Chen

    Ma Chen, associate professor at City University of Hong Kong, has joined Yingjie Technology to lead pre-research on the data direction. The company has completed two consecutive seed funding rounds over the past few months, totaling tens of millions of yuan. Yingjie is building a data pipeline spanning multimodal collection, spatiotemporal alignment, spatial structuring, and dynamic tracking and prediction validation, while advancing spatial intelligence benchmarks and post-training environments, feeding model failure modes back into data collection and training. Ma Chen proposes comparing different supervision methods under the same foundation model and the same budget, focusing on performance in unseen scenarios and task success rates.

  3. 钛媒体48

    Unitree shares rebound over 70% in 12 days as GPT-6 Astra fuels embodied intelligence narrative

    Unitree Technology's share price has rebounded by more than 70% over 12 days, with total market value recovering to over 280 billion yuan. JPMorgan raised its target price from 300 yuan per share to 995 yuan per share. Its first post-IPO financial report, released after market close on October 28, showed Q3 revenue of 665 million yuan, up 69.93% year on year, but non-GAAP net profit of 105 million yuan, down 17.42% year on year, and R&D investment of 127 million yuan, up 249.58% year on year.

  4. OFweek Robotics60

    After a Capital Surge and Deployment Milestones, AGIBOT's Real Test as an Embodied AI Chain Leader Is Just Beginning

    The article reviews AGIBOT's moves in recent months: it publicly introduced a nine-person core partner lineup, launched a Hong Kong IPO, and after completing a new funding round reached a valuation of 50 billion yuan. It also took a controlling stake in SWANCOR Advanced Materials for about 2.1 billion yuan, while spinning off subsidiaries including Mifeng, Critical Point, Qingtianzu, AGIBOT KuoTuo, Zhiding, and Yuansheng Wotu, which was established in October.

  5. OFweek Robotics58

    Unitree Technology's Q3 revenue rises but profit falls; JPMorgan still raises target price from 300 yuan to 995 yuan

    Unitree Technology's third-quarter operating revenue was 665 million yuan, up 69.93% year on year, while net profit attributable to shareholders was 113 million yuan, down 17.37% year on year, making its first financial report after listing show higher revenue but lower profit. On October 25, JPMorgan raised its target price for Unitree from 300 yuan to 995 yuan, upgraded the rating directly from underweight to overweight, and raised its humanoid robot sales forecasts for 2027 to 2030. The article also notes that cases such as YunShenChu's STAR Market review and Huandong Robot's termination of listing show that listing reviews for robotics companies are tightening, and several second- and third-tier manufacturers have experienced unpaid wages or bankruptcy.

  6. 钛媒体58

    From Genie 4 to ByteDance's world model: Zhang Yiming's "generate the next second" plan

    Google released the world model Genie 4 on October 28; users can use natural language to change the weather, move the camera, and introduce characters in real time, and the model includes physical simulations such as aerodynamics and collisions, and adds sound generation for the first time. The article also reviews ByteDance's work on world models since 2024, including two routes, VLA and 3D simulation, as well as Seedance 2.5, Seed3D, and SeedRealtime, and Zhang Yiming personally coordinating resources with the aim of connecting AI models, cloud computing, content platforms, and Pico hardware.

  7. Robotics & Automation News34

    Faraday Future discusses shift from Chinese robot sourcing to US manufacturing

    Faraday Future plans to launch a robot factory in the US by the end of 2026, with the first US-assembled robot scheduled to roll off the line in February 2027, and to launch three to five new or next-generation robot products in the US each year. The company says its "Built in USA" strategy aims to strengthen compliance, supply chain resilience, quality control, and customer support, and it plans to move from "US assembly" to "US manufacturing" by 2028. Global Executive Chairman Jerry Wang said the company has not disclosed the share of US-manufactured content or the specific components that will still need to be imported in 2028.

Oct 28Wed
  1. 电子工程专辑 · 微信公众号38

    Musk responds to Counterpoint humanoid robot DRAM forecast: robot memory demand will surpass all applications on Earth

    On October 27, Musk responded on X to Counterpoint's report on humanoid robot DRAM demand, calling it an "absurd analysis" and judging that robots' demand for memory will surpass every other application on Earth. Counterpoint forecasts that per-unit DRAM capacity in humanoid robots will rise from the current 19GB to 39GB in 2030, with total demand growing from about 8,300TB in 2026 to about 172,000TB in 2030.

  2. 未来机器人网 · 微信公众号32

    Musk Says Humanoid Robot Memory Demand Will Surpass All Scenarios, Global Memory Sector Moves

    On October 27, Musk responded on X to Counterpoint's forecast for humanoid robot memory demand, saying robot RAM demand will surpass every other application scenario on Earth, driving Western Digital up 6.95% and Micron Technology up 2.77% in US stocks, while the A-share humanoid robot sector rose over 4% on October 28.

  3. 车东西71

    Exclusive: Neolix's Yu Enyuan on raising double its target and aiming for 40,000 autonomous vehicles by year-end

    Neolix founder and CEO Yu Enyuan said at a media briefing that the company's latest funding round originally targeted USD 200 million but ultimately raised double that amount. The company aims to achieve overall profitability in 2027 and plans to gradually shift from selling vehicles to providing transport capacity services. He said gross margin exceeds 60% on a per-vehicle revenue basis, and the year-end target for autonomous vehicle scale is about 40,000 units. The new-generation product's overall operating efficiency is more than 3 times that of previous products, with design service life extended from 5 years to 8 years and daily effective service time increased from about 12 hours to 18 hours.

  4. 雷科技42

    Interview with SparkLink Alliance's Zeng Guosong: SparkLink 3.0 Sees AI as an Opportunity to Overtake, Replacing Wired Robot Joints with Wireless

    Zeng Guosong, Secretary-General of the International SparkLink Alliance, said after the release of the SparkLink 3.0 standard that SparkLink 3.0 will develop around AI, with the goal of becoming the technical infrastructure of the AI intelligence era. The alliance has established an embodied intelligence working group and plans to use SparkLink modules to reduce cabling in robot joints. Its standard specifies 99.999% reliability and is divided into four communication modes: SLE, SLB, SLP, and SLZ. The alliance currently has nearly 1,260 member units, more than 100 of which are international members, and in April next year it will add an industry promotion group for the smart energy sector.

Oct 27Tue
  1. 硅星人Pro · 微信公众号22

    Manycore, Meshy, and Hyper3D Discuss AI 3D: Why Chinese Companies Are Leading

    Manycore Tech, Meshy, and Hyper3D discussed why Chinese AI 3D companies are leading at the ACC 2026 AI Creator Carnival. Manycore has launched the spatial generation model SpatialGen and is exploring the 3D generation model Aholo Lux3D, whose materials support up to nine channels; Hyper3D launched Agentic Mode, moving from Production Ready toward Agent Ready. The three companies believe that domestic high-quality 3D data, the production pipeline ecosystem, and talent accumulation are the main advantages.

  2. OFweek Robotics38

    Humanoid robot sector begins to stampede: old shares sold at 50% discount

    Old shares in the humanoid robot sector are being offloaded at a discount: a domestic unicorn working on world models, founded only a few years ago, raised tens of billions in the first half of this year and is valued at tens of billions, and is now urgently selling old shares at 50-60% of the original price. Another embodied intelligence company's valuation rose from 3 billion yuan to 22 billion yuan, and investors transferring old shares at 12 billion yuan (a discount of about 45%) still have not found buyers. In September, regulators began paying more attention to corporate valuations and revenue quality, and private-market valuations of some projects have already been cut by 30% to 50%.