与「EmbodiedAI」相关的搜索结果

POV: You just witnessed small robot N2 unleash huge energy 🤯 🏃 Running 🤸 forward flips ✨ Yingge Dance 💃 Cha-Cha-Cha A little robot with a champion’s spirit. 🏅 #Noetix# #N2# #HumanoidRobots# #RobotTechnology# #RobotMotion# #EmbodiedAI#

The future of humanoids isn't just mechanical—it's neural. A humanoid robot is an embodied AI system. Motors, sensors, and actuators provide the body, but neural intelligence provides the mind. Neural models power vision, speech, language, navigation, manipulation, planning, memory, decision-making, and continuous learning. As robotics evolves, nearly every cognitive subsystem is becoming neural-first. That's why represents more than a niche—it's a foundational concept for the next generation of intelligent robotics. #HumanoidRobots# #EmbodiedAI# #PhysicalAI# #NeuralNetworks# #ArtificialIntelligence# #Robotics# #MachineLearning# #DeepLearning# #RobotLearning# #ReinforcementLearning# #ComputerVision# #GenerativeAI# #AIResearch# #Automation# #FutureOfAI#

A new approach to acoustic simulation from PKU could support future work in embodied intelligence! The research team led by Li Sheng and Wang Guoping from the School of Computer Science, Peking University, has won the Best Paper Award at the prestigious CASAXR 2026 (39th International Conference on Computer Animation, Social Agents, and Extended Reality) held in Geneva, Switzerland. Targeting key bottlenecks in traditional acoustic simulation—including numerical errors and poor adaptability to complex dynamic boundaries—the team proposed the novel SonicRadiation framework. This ghost-cell-free hybrid boundary-grid coupling method achieves simultaneous improvements in simulation accuracy and efficiency. It lays a solid physical foundation for high-fidelity acoustic simulation, embodied intelligence training, and accurate sound-physics coupling modeling for world models, empowering AI agents with authentic acoustic perception and interactive capabilities in complex dynamic scenarios. #PKUResearch# #EmbodiedAI# @CampusBiotech

Wait… is Oli the crowd favorite at the first AI Fashion Competition? 🤖✨ Oli joined the grand finale of the first global "Al + Fashion" Innovative Application Competition, experiencing #2026FashTech# up close. The runway was stunning, but the crowd’s excitement? It's even harder to miss. #LimXDynamics# #HumanoidRobot# #LimXOli# #EmbodiedAI#

Can robot foundation models scale? Xiaomi-Robotics-1 explores this question with over 100,000 hours of real-world manipulation data. Pre-trained on large-scale real-world trajectories and post-trained with cross-embodiment robot data, Xiaomi-Robotics-1 demonstrates consistent scaling across data and model size, strong generalization in unseen environments, and efficient adaptation to new tasks. 🔗 #Robotics# #EmbodiedAI# #FoundationModels# #RobotLearning# #XiaomiAI# #XiaomiRobotics#

Our robot intern just reached another milestone. Earlier this year, Xiaomi's humanoid robot began its internship at Xiaomi EV Factory. After one quarter, its dual-side task success rate at the self-piercing nut workstation reached 98%—just 1 percentage point shy of the success rate achieved by experienced human operators. Now it's taking on two new factory roles: • Center console side-cover sorting • Returnable box folding Both have already achieved over 90% task success. Every new task is another step toward real-world deployment. #HumanoidRobots# #Manufacturing# #EmbodiedAI# #Robotics# #Xiaomi#

XPENG’s robotics business has completed a 900M+ Series A funding round at a post-money valuation exceeding $6.3B, setting a new fundraising record for China’s embodied AI industry. The round was led by IDG Capital, joined by Gaorong Ventures, and strategically backed by Tencent and Alibaba, with XPENG retaining controlling ownership. The capital will accelerate full-stack hardware/software R&D, physical AI training, smart manufacturing infrastructure, and global market expansion. Backed by 12 years of in-house tech innovation—from physical foundation models to Turing AI chips and robust AI infrastructure—XPENG is uniquely positioned to lead the scale-up and commercial rollout of high-end humanoid robots. XPENG IRON is slated for mass production in late 2026, launching first in XPENG retail stores and campuses before rolling out globally with customer deliveries in 2027. $XPEV

.@base 生态现在整体很热,Meme 轮番上， 这类标的也上了 Coinbase，资金和注意力都在往里涌。交易、Agent、上币预期成了最热闹的主线 但我更想往上游看一层 Base → Virtuals → Robotics → Data 这条逻辑其实很干净。 Base 在强化交易和 Agent 入口，@virtuals_io 也已经把 Robotics / Embodied AI 放进核心方向。 而 AI 一旦从数字世界进入物理世界，真正难的问题就变了。 LLM 可以直接吃互联网数据。 机器人需要学习人在真实环境里怎么抓、怎么放、怎么使用工具、怎么失败，以及失败以后怎么修正。 这些数据很难靠爬虫获得，纯仿真也很难覆盖真实世界里大量长尾情况。 所以 Physical AI 接下来一个很重要的瓶颈，就是： 真实世界 Human Demonstration Data 怎么规模化生产。 Caspius 卡的就是这一层。 目前公开能看到： 1000+ 小时 Human Demonstration Data 500+ Active Contributors Phase 1 第一人称数据采集已经运行 下一阶段是价值密度更高的 Teleoperation Data 规模还很早，但至少数据生产闭环已经开始运转。 Miner NFT 这一层我也觉得被低估了。 它更接近 DePIN 里的“矿机”： 决定数据产能 → 升级消耗 $CAS → 支持委托 → 按贡献产生排放 NFT、Token 和真实数据生产因此被连接到了一起。 所以相比追最热闹的机器人名字，我更想找那些机器人真正大规模进入现实世界以后，所有玩家都绕不开的环节。 @caspius_ai 值得继续跟踪的原因就在这里： 它站在一条正在升温的产业链上，也卡着一个还没有被充分解决的数据问题。 剩下的，就交给数据增长、Teleoperation，以及真实买方💪

Robot walked into the White House. It talked. It left. And the company that built it wants you to believe nobody was steering. $39,000,000,000. That's what investors just bet on you believing that. REAL March 25, 2026. Figure 03 — built by Figure AI — walked the East Room carpet next to the First Lady of the United States. It scanned the room. It spoke first, before any human did. Introduced itself. Greeted the room in 11 languages. Walked out. AP photographers were three feet away. This is not CGI. This is not a demo reel. Six months earlier, Figure AI closed $1 billion+ in new funding. Backers: Nvidia. LG. Valuation: $39 billion. RUMOR The company's own CEO called it "fully autonomous." Zero human control. Zero scripted lines fed in real time. I spent an evening trying to find one confirmation of that from anyone who doesn't work for Figure AI. Found zero. The White House's own materials don't even confirm which robot model showed up that day. Every "100% autonomous" quote traces back to one source — the company selling the robot. That's not proof. That's marketing wearing a lab coat. WHY THIS SHOULD SCARE AND EXCITE YOU AT THE SAME TIME Nobody's saying the quiet part — governments don't adopt robots because they're proven. They adopt them because a rival government already has one in a photo. Figure AI didn't need a federal contract. It needed five minutes in the East Room. That's the entire playbook for embodied AI right now: visibility first, verification later. Honestly — I don't know if that robot could survive one unscripted question from a reporter. Nobody's tested that on camera. Not once. That silence is louder than the speech it gave. Every government on earth wants their own robot in the room now. Whether it's actually thinking — that's the question nobody in that room asked out loud.

China’s robotics startups are starting to invest in other companies while they’re still only at the Series A stage themselves. And another pattern is starting to appear: companies that were invested in are now turning around and investing in others. AgiBot → PsiBot → Mifeng Technology → KONVERY 29 embodied AI companies have already made 125 external investment deals. I think this says a lot about how fast the sector is moving in China: Companies don’t want to wait years to build every capability in-house. They’re using capital to lock in technologies, teams, and supply-chain positions early.