The Google DeepMind angle is the tell. Robot Park data helps advance #GeminiRobotics, while Gemini Robotics is built to generalize across embodiments, including Apptronik's Apollo. The loop is: tasks -> data -> model -> better robots -> more tasks.

Hi folks! Any roboticists at #ICML2026 working with agents? Let's connect, get to know each other, and grab a coffee☕️ or matcha🍵! I'm really keen to chat about self-evolving harness and post-training for agents in robotics and automation. See you in Seoul! 🇰🇷 #GeminiRobotics #NvidiaGR00T #VLM #MetaHarness

Check out this great tutorial of #GeminiRobotics for developers. Looking forward to seeing what you build! 🦾🧠🔥 @GoogleDeepMind Robotics Thank you @ptruiz_dev ! youtu.be/kCXQce1dtNE?si=A-Qm…

Physical AI is no longer sci-fic it’s here and it’s raising big questions about safety & governance. Unlike normal AI that just lives on screens, Physical AI powers real robots and machines that actually move and work in the physical world. In 2024 alone, 542,000 industrial robots were installed worldwide. The market is set to explode from $82 billion in 2025 to nearly $1 trillion by 2033! Google DeepMind just dropped Gemini Robotics — it can fold paper, pack bags, handle random objects, and even reason “is this safe?” before acting. But governing these physical robots is way harder than regular AI. We need strong safety rules, human oversight, and clear boundaries. Right now, only about 1 in 3 companies have solid AI governance in place. Are we moving too fast with autonomous robots, or should we slow down and get the safety frameworks right first? Drop your thoughts 👇 --- Full article: artificialintelligence-news.… #PhysicalAI #Robotics #AIGovernance #GeminiRobotics #AISafety

RoboCasa365 is amazing, interesting to find this leaderboard, where we can have ready-to-use ckpt. And results are not surprising low lol Excited to see #DreamZero, #GeminiRobotics, #Generalist appearing here, we need more scalable sim-based benchmark

Google DeepMind just released Gemini Robotics-ER 1.6, a major upgrade that gives robots far better spatial and physical reasoning. The model now excels at understanding objects, obstacles, dynamics, and multi-view scenes, enabling more complex real-world task planning and higher success rates in dynamic environments. This brings embodied AI closer to practical deployment in warehouses, factories, and homes. What robot task should this upgrade tackle first? Drop it below. Follow @rise_raise_ai for more AI systems. #GeminiRobotics #EmbodiedAI #DeepMind

Very excited to bring #GeminiRobotics intelligence 🤖 🧠 to Spot with the @BostonDynamics team! We are just getting started! 🔥 spectrum.ieee.org/boston-dyn… @GoogleDeepMind Robotics

『四足歩行ロボットにGemini AIを統合』 現場の状況を理解・判断し、“歩けるAI”として機能。 犬との散歩でも臨機応変に遊んであげる。 bostondynamics.com/blog/tool… #PhysicalAI #EmbodiedAI #IndustrialAI #VLM #GeminiRobotics #quadrupedal #robot #Spot #BostonDynamics

Excited to announce #GeminiRobotics ER 1.6 our upgraded embodied reasoning model with: ⏱️Advanced instrument reading 👉improved spatial pointing ✅enhanced success detection (including multiview) 🛑more robust safety features Read more: deepmind.google/blog/gemini-…

The real money in Gemini Robotics-ER 1.6 isn’t the model — it’s the businesses that deploy it first. • Integrate the upgraded reasoning layer into warehouse robots, home assistants, or industrial arms for 2–3x productivity gains • Build or license vertical solutions (logistics, eldercare, manufacturing) and charge $10k–$100k per deployment • Package custom fine-tuned agents on top of ER 1.6 as high-margin SaaS or managed services • Early movers are already locking in contracts before competitors even finish pilots Embodied AI just became investable and monetizable at enterprise scale. The labs built the upgrade — smart operators are building the businesses. How would you monetize this robotics breakthrough in your industry? Tell us below. Follow @rise_raise_ai for more AI systems. #GeminiRobotics #EmbodiedAI #AIRobotics

Google DeepMind just dropped Gemini Robotics-ER 1.6 — the upgrade that finally lets robots truly reason about the physical world. • Significantly improved visual and spatial understanding lets robots plan multi-step tasks with human-like intuition • Robots can now interpret object properties, obstacles, dynamics, and spatial relationships before acting • Task completion rates jump in real-world environments — warehouses, homes, factories, and beyond This isn’t another lab demo. It’s the moment embodied AI becomes practically useful at scale. What real-world robot task do you want solved first with Gemini Robotics-ER 1.6? Drop it below. Follow @rise_raise_ai for more AI systems. #GeminiRobotics #EmbodiedAI #DeepMind

Thrilled to partner with Agile Robots! 🤝 By bringing #GeminiRobotics foundation models to their incredible industrial platform, we’re enabling robots to truly reason about and adapt to the physical world 🦾🧠. A key step toward building the intelligent, autonomous systems of the future!🚀

#Google développe une #IA appelée #GeminiRobotics et travaille avec Apptronik pour développer “la prochaine génération de #robots humanoïdes 👉Cette startup vient de lever près d’un milliard de $ pour accélérer la commercialisation de son robot Apollo presse-citron.net/menace-opt…

🇬🇧伦敦Google除了是 #DeepMind 总部外，还是Android, Chrome和YouTube全球核心工程基地，包括全球最大的知识库Search和Scholar。围绕下一代视频模型、多模态AI和World Model演进，有几条非常明确的信号： 1️⃣Next-gen video model视频模型（#Veo 级别）不是单纯卷画质，而是走向实时交互。这是一个比“AI 生成视频”大得多的赛道，Veo 已经被用于 NBA 等大型商业营销项目，进入真实生产环境 2️⃣World Model才是真正的下一代AI底座，从自动驾驶到 #GeminiRobotics ，从仿真训练到复杂决策，AI正在学习“理解真实世界” 3️⃣多模态进入“可控时代”，开始接管专业工作流，真正的瓶颈不在模型，PoC 很容易，规模化很难，拼的是数据、流程、组织能力 4️⃣计算资源支撑：依托谷歌庞大的算力和数据中心，它可以进行更大规模的实验，加上深度集成生态原生集成了Gmail、Docs、Android 和 Search。这种AI就在你手边的无缝感 5️⃣未来不是 #AI 会不会取代人，而是不会用 AI 的人一定会被淘汰 具有科学研究底蕴的 #谷歌 是全球唯一一家在AI领域获得诺贝尔奖（2024年化学奖）的企业实验室，这是目前AI技术在基础科学领域获得的最高认可，我们正在从“生成内容的 AI”，走向“理解世界、参与决策、执行任务的 AI”

It is indeed fun to play with #Apollo with #GeminiRobotics . You can trick him and change commands and move objects and change them. I realize we do call him a “he” @FryRsquared, and we feel a little bit bad when we trick him but he stays helpful :) youtube.com/shorts/CzcM4Dx3a… via @YouTube

World models are helping us evaluate #GeminiRobotics generalist policies more effectively, including auto-red-teaming to safely find and address vulnerabilities in the models. Learn more at: veo-robotics.github.io @GoogleDeepMind Robotics

A peek into our lab @GoogleDeepMind Robotics and #GeminiRobotics with @FryRsquared. 🦾❤️

『Google DeepMind ロボット研究所の見学ツアー』 彼らが どのようにして文脈を理解し、自ら物事を判断できるエージェントを構築しているのかが垣間見える。 youtu.be/UALxgn1MnZo #robotics #ArtificialInteligence #humanoidrobot #officetour #GeminiRobotics #DeepMind #GoogleDeepMind