# What is true on Models?

> Markdown twin of https://roboterminal.app/topics/models · https://roboterminal.app/topics/models.md

## On this layer

- org — [NVIDIA](https://roboterminal.app/companies/nvidia) (35)
- org — [Google](https://roboterminal.app/companies/google) (22)
- org — [Figure](https://roboterminal.app/companies/figure) (19)
- org — OpenAI (17)
- org — World Labs (15)
- org — [Skild AI](https://roboterminal.app/companies/skild) (13)
- org — [BMW](https://roboterminal.app/companies/bmw) (12)
- org — [Hugging Face](https://roboterminal.app/companies/hugging-face) (11)
- org — [Unitree](https://roboterminal.app/companies/unitree) (10)
- org — a16z (9)
- org — [Tesla](https://roboterminal.app/companies/tesla) (9)
- org — Meta (6)
- org — axis robotics (5)
- regulator — [Fcc](https://roboterminal.app/companies/fcc) (5)
- org — Agibot (4)
- org — [Agility Robotics](https://roboterminal.app/companies/agility) (4)
- org — Alibaba (4)
- org — Anthropic (4)
- org — Gimlet Labs (4)
- org — [pi](https://roboterminal.app/companies/pi) (4)
- org — agentic video understanding (3)
- org — [Apptronik](https://roboterminal.app/companies/apptronik) (3)
- org — Dyna Robotics (3)
- org — Microsoft (3)
- org — Oracle (3)
- org — Samsung (3)
- org — The University of Hong Kong (3)
- org — Thinking Machines Lab (3)
- org — Zhipu (3)
- org — [1x](https://roboterminal.app/companies/1x) (2)
- org — a16z New Media (2)
- org — Aglaé Lab (2)
- org — AI chip inference cluster (2)
- org — Alpha Intelligence Capital (2)
- org — AMI Labs (2)
- org — Apple (2)
- org — Applied Intuition (2)
- org — Artémis (2)
- org — Association Familiale Mulliez (2)
- org — Atlas (2)
- org — Booster Robotics (2)
- org — [Boston Dynamics](https://roboterminal.app/companies/boston-dynamics) (2)
- org — Bpifrance Digital Venture (2)
- org — Din Tai Fung (2)
- org — Fable (2)
- org — Gaorong Ventures (2)
- org — GPT (2)
- org — Groupe Industriel Marcel Dassault (2)
- org — Hyundai (2)
- org — IDG Capital (2)
- org — Lightwheel (2)
- org — LimX Dynamics (2)
- org — Nabla (2)
- org — napkin production (2)
- org — New Legacy Ventures (2)
- org — Noble Machines (2)
- org — Pixomondo (2)
- org — Pollen Robotics (2)
- org — Publicis Groupe (2)
- org — Qwen (2)
- org — Qwen3.8-Max-0902 (2)
- org — Sanctuary AI (2)
- org — SBVA (2)
- org — Scale AI (2)
- org — Schaeffler (2)
- org — Sea (2)
- org — Temasek (2)
- org — Tencent (2)
- org — tokens processed (2)
- org — Toyota Ventures (2)
- org — UnitreeRobotics (2)
- org — Vercel (2)
- org — ZEBOX Ventures (2)
- org — Zhejiang University (2)
- org — 3D reconstruction (1)
- org — ACE Robotics (1)
- org — adaptation time for new robot embodiments (1)
- org — ADLINK (1)
- org — AgiBot World platform comprises over 1 million trajectories. (1)
- org — Agibot's performance in the World Humanoid Robot Games. (1)
- org — AGIBOTofficial (1)
- org — AI spending by a large consulting firm. (1)
- org — Aimalysheva (1)
- org — Amazon (1)
- org — AMD (1)
- org — AMD Ventures (1)
- org — AMI Labs funding (1)
- org — Amount raised by AMI Labs in funding. (1)
- org — app downloads (1)
- org — Astribot (1)
- org — Atlas model for spatial intelligence. (1)
- org — Autel (1)
- org — average per-step success (1)
- org — average performance improvement of policies using the dataset. (1)
- org — AWS (1)
- org — badges for contributions in the AXIS campaign. (1)
- plant — bmw-spartanburg-plant (1)
- org — Bosch (1)
- org — Bronze medals won by Agibot. (1)
- org — BYD (1)
- org — Caterpillar (1)
- org — Caterpillar's operational data. (1)
- org — Cerebras (1)
- org — Cerebras WSEs needed for reasonable concurrency of 256 requests. (1)
- org — Cerebras WSEs needed for running a large model. (1)
- org — chemistry (1)
- org — Cloudflare (1)
- org — Columbia University (1)
- org — Cosmicbrainai (1)
- org — cost of infrastructure (1)
- org — Cover (1)
- org — CRWV (1)
- org — Current hours of available data for training robots. (1)
- org — Current throughput of valid trajectories collected per hour. (1)
- org — Cycle time for completing a plug insertion. (1)
- org — data efficiency comparison (1)
- org — Deep Robotics (1)
- org — DeepgramAI (1)
- org — DeepSeek (1)
- org — Delta in cost due to supply chain differences (1)
- org — Demonstrations collected for the tasks. (1)
- org — deployments of Spot (1)
- org — development timeline of Moby. (1)
- org — development timeline reduction. (1)
- org — Dexmal (1)
- org — Direct wafer paths reduction in latency. (1)
- org — DiscoLoopAI (1)
- org — DJI (1)
- org — DJI Flip (1)
- org — DoorDash (1)
- org — DROID dataset robot interaction data (1)
- org — Dusty Robotics (1)
- org — Dyna (1)
- org — E-Commerce Bench (1)
- org — Embodied AI model (1)
- org — episodes in RoboLab evaluation (1)
- org — estimated revenue in 2026 (1)
- regulator — europe (1)
- org — examples for adapting to new robotic embodiments. (1)
- plant — factory-in-china (1)
- org — Fetch Robotics (1)
- org — Fictiv (1)
- org — FieldAI (1)
- org — fiftyyears (1)
- org — filtered simulation episodes (1)
- org — Flash (1)
- plant — fremont (1)
- org — Fudan University (1)
- org — G1 humanoid grasp and lift task (1)
- org — Galbot (1)
- org — Gamma Robotics (1)
- org — Geely Auto (1)
- org — GEN-1.5 can learn new tasks in a few seconds. (1)
- org — GEN-1.5 model capabilities (1)
- org — General Catalyst (1)
- org — general purpose robot foundation model (1)
- org — generalist (1)
- org — GGUF conversion and quantization (1)
- plant — giga-texas (1)
- org — global market forecast for humanoid robots by 2027 (1)
- org — Gold medals won by Agibot. (1)
- org — Grok (1)
- org — hardware design (1)
- org — Hark (1)
- org — Harvard University (1)
- plant — herzogenaurach (1)
- org — Horizon Robotics (1)
- org — Humanoid robots procurement (1)
- org — humanoid robots shipped globally in 2026 (1)
- org — humanoid robots shipped in 2025 (1)
- org — humanoids built (1)
- org — Hydra-0 training (1)
- org — Hyper3D (1)
- org — [Intel](https://roboterminal.app/companies/intel) (1)
- org — Irregular (1)
- regulator — justice-department (1)
- org — LG (1)
- org — LightOrigins (1)
- org — LM Studio (1)
- org — Lotus Cars (1)
- org — Manycore Tech (1)
- org — MarathonMP (1)
- org — market size for Autonomous AI (1)
- org — McAfee (1)
- org — memory manufacturer capacity (1)
- org — Meta and Muse Spark 1.3 performance upgrade (1)
- regulator — meti (1)
- regulator — ministry-of-science-and-ict (1)
- org — MISUMI (1)
- org — MIT (1)
- org — Moby deployment in a semiconductor facility. (1)
- org — Model merge process (1)
- org — model size (1)
- org — Morgan Stanley (1)
- org — Mostik (1)
- org — Munari (1)
- org — Muse (1)
- org — NBIS (1)
- org — NBT average for RandER. (1)
- org — NBT average with 1% Memory Anchors removed. (1)
- org — NBT average with 10% Memory Anchors removed. (1)
- org — NBT average with 5% Memory Anchors removed. (1)
- regulator — nedo (1)
- org — Nemotron 3.5 Lightning on Jetson AGX Thor (1)
- org — NEURA (1)
- org — NotionHQ (1)
- org — Nscale (1)
- org — Number of downloads of the Index platform. (1)
- org — Number of global contributors and collected trajectories on the platform. (1)
- org — number of objects in AgiBot World. (1)
- org — Ollama (1)
- org — outsetcap (1)
- org — Perceptron (1)
- org — Perceptron AI (1)
- org — perceptroninc (1)
- org — performance comparison (1)
- org — performance improvement of the proposed policy over prior arts. (1)
- org — Performance metric for model. (1)
- org — performance of S1 compared to VLAs (1)
- org — performance of S1 on novel tasks (1)
- org — Perplexity (1)
- org — Personal investment by XPENG founder and co-president (1)
- org — physical AI data delivered (1)
- org — PiPER (1)
- plant — plant-spartanburg (1)
- org — Pollen (1)
- org — post-training duration (1)
- org — production requirement for napkins (1)
- org — Promise (1)
- org — prototype (1)
- org — Qualcomm (1)
- org — Qwen3.5-4B on Jetson Orin Nano (1)
- org — rhoda (1)
- org — robot deployment (1)
- org — robot interaction data from the DROID dataset (1)
- org — robot learning economics (1)
- org — robot programming (1)
- org — robot training (1)
- org — ROBOTIS (1)
- org — Runway (1)
- org — S1 compared to current VLA models (1)
- org — S1 model capabilities (1)
- org — S1 robot foundation model (1)
- org — S1's accuracy compared to conventional VLA models. (1)
- org — S1's accuracy with one example compared to current VLA models (1)
- org — S1's performance on manipulation tasks (1)
- org — SAGE-10K dataset (1)
- org — sanctuary (1)
- org — Scale AI data delivery (1)
- org — Schaeffler’s agreement with Humanoid (1)
- org — Seeed (1)
- org — Seeed Studio (1)
- org — Shanghai AI Lab (1)
- org — Shanghai Innovation Institute (1)
- org — Silver medals won by Agibot. (1)
- org — Solomon (1)
- regulator — south-korea-ministry-of-planning-and-budget (1)
- org — SpaceX (1)
- org — SpaceXAI (1)
- plant — spartanburg (1)
- org — speedup in generative performance (1)
- org — Stanford University (1)
- org — status of physical AI companies (1)
- org — success rate of the post-trained policy. (1)
- org — successful grasps by the G1 humanoid on hardware. (1)
- org — target action-loss (1)
- org — Target valid hours of data per month from the Mobile Ego-Centric App. (1)
- org — tariffs on Chinese drones (1)
- org — tasks available in RoboLab. (1)
- org — technia (1)
- org — TeleAI (1)
- org — teleoperation hours (1)
- org — teleoperation reduction (1)
- org — the project stand not just as a piece of marketing, but as a vision statement: a world built by the very technology that made it possible. (1)
- org — Total number of videos collected by Index. (1)
- org — total payment (1)
- org — total production of humanoid robots (1)
- org — Total robotic systems procurement (1)
- org — Toyota Research Institute (1)
- org — Training Data (1)
- org — training hours for LDA-1B model (1)
- org — training needed for current VLA models to match S1's accuracy (1)
- org — training video hours (1)
- org — TranscEngram (1)
- org — TRON 2 (1)
- regulator — u.s.-government (1)
- org — UC Berkeley (1)
- org — validated run uses 64 nodes of 4× GB200 for 60K iterations, roughly 68 hours (17.4K GB200-hours). (1)
- org — video generation (1)
- org — video uploads (1)
- org — wafer (1)
- org — Wafer's funding announcement. (1)
- org — weekly active users for data collection (1)
- org — Wing_VC (1)
- org — Wonik Robotics (1)
- org — xAI (1)
- org — XPENG (1)
- org — XPeng Motors (1)
- org — XPENG's robotics funding (1)
- org — ycombinator (1)
- org — ZiNovaLabs (1)
- org — 北京人形机器人创新中心有限公司 (1)
- regulator — 国家发展改革委 (1)
- org — 宇树科技 (1)

## Who captures

- NVIDIA (23)
- Figure (14)
- OpenAI (13)
- World Labs (12)
- @smsehy (12)
- Google (11)
- Skild AI (10)
- @stretchcloud (6)
- a16z (5)
- axis robotics (5)
- @ricci_nov (4)
- Meta (4)
- @chris_j_paxton (3)
- Zhipu (3)
- Tesla (3)
- Anthropic (3)
- Unitree (3)
- The University of Hong Kong (3)
- Physical Intelligence (π) (3)
- Dyna Robotics (3)
- Joseph Amigo, Rooholla Khorrambakht, Nicolas Mansard, Ludovic Righetti (3)
- Microsoft (2)
- Qwen (2)
- Hugging Face (2)
- @RoboPapers (2)
- @pstAsiatech (2)
- @Scobleizer (2)
- Agibot (2)
- @tchsignal (2)
- Fable (2)
- Alibaba (2)
- Agility Robotics (2)
- UnitreeRobotics (2)
- @0xconglomerate (2)
- @frontrunvc (2)
- BMW (2)
- @gupta_abhinav_ (2)
- AMI Labs (2)
- Galbot (1)
- Apple (1)
- ROBOTIS (1)
- @martin_casado (1)
- TeleAI (1)
- LimX Dynamics (1)
- @deepakpathak (1)
- Thinking Machines Lab (1)
- @SemiAnalysis_ (1)
- Astribot (1)
- Vercel (1)
- Munari (1)
- @AppliedInt (1)
- @rimtoln (1)
- Zhejiang University (1)
- Dusty Robotics (1)
- @MelvinInvests (1)
- @elonmusk (1)
- Horizon Robotics (1)
- SpaceXAI (1)
- @OperationsPLS (1)
- General Catalyst (1)
- @ruima (1)
- Perplexity (1)
- Mostik (1)
- Caterpillar (1)
- Morgan Stanley (1)
- ZiNovaLabs (1)
- Noble Machines (1)
- sanctuary (1)
- wafer (1)
- https://x.com/ChongZzZhang (1)
- Hyper3D (1)
- Perceptron AI (1)
- Schaeffler (1)
- @NVIDIARobotics (1)
- technia (1)
- Runway (1)
- @whitee_rhinoo (1)
- DJI (1)
- Sanctuary AI (1)
- XPENG (1)
- @techniahqrobot (1)
- Hyundai (1)
- @realgalleryx (1)
- @du_maximilian (1)
- Cerebras (1)
- Dyna (1)
- Perceptron (1)
- Cosmicbrainai (1)
- @dredgefactory (1)
- rhoda (1)
- @SkildAI (1)
- BYD (1)
- NEURA (1)
- 北京人形机器人创新中心有限公司 (1)
- @EdmondIsARobot (1)
- XPeng Motors (1)
- https://x.com/RoboPapers (1)
- @TheHumanoidHub (1)
- https://x.com/Majumdar_Ani (1)
- https://x.com/Sylviaposts (1)
- Boston Dynamics (1)
- Stanford University (1)
- Promise (1)
- Kevin Black (1)
- Edmond (1)

## Sourced numbers

- **Gimlet Labs:** $300M
  - _"We're thrilled to lead a $300M investment in Gimlet Labs."_
  - Source: [@a16z on X](https://x.com/a16z/status/2095905665516736760)
- **Gimlet Labs:** $300M
  - _"We're thrilled to lead a $300M investment in Gimlet Labs."_
  - Source: [@a16z on X](https://x.com/a16z/status/2095904215969477103)
- **Thinking Machines Lab:** $1billion
  - _"Thinking Machines Lab is in talks to raise $1 billion at a $40 billion pre-money valuation, according to The Information."_
  - Source: [@stretchcloud on X](https://x.com/stretchcloud/status/2095624626084716955)
- **Thinking Machines Lab:** $1billion
  - _"Thinking Machines Lab is in talks to raise $1 billion at a $40 billion pre-money valuation, according to The Information."_
  - Source: [@stretchcloud on X](https://x.com/stretchcloud/status/2095624626084716955)
- **Meta and Muse Spark 1.3 performance upgrade:** $0
  - _"Mark Zuckerberg calling it "frontier performance almost too cheap to meter.""_
  - Source: [@TheRundownAI on X](https://x.com/TheRundownAI/status/2095234538267877768)
- **3D reconstruction:** 25.3 units
  - _"Atlas claims mean AbsRel error of 25.3, better than specialist tools built for that task."_
  - Source: [@stretchcloud on X](https://x.com/stretchcloud/status/2095225248774029545)
- **Digit's manipulation capabilities in varied environments.:** 1 units
  - _"Digit uses to take on mixed loads and changing layouts on a real warehouse floor."_
  - Source: [Agility on X: "Every object a different size, weight, and grip. Rearranging a full room's worth of stuff is the same generalized manipulation Digit uses to take on mixed loads and changing layouts on a real warehouse floor.

#Agility #DigitRobot" / X](https://x.com/HumanoidInvest/status/2095166627708887090)
- **Hydra-0 training:** 2202 hrs
  - _"It was trained on about 2,202 hours of multi-embodiment video spanning human hands, handheld grippers, single-arm robots and bimanual systems."_
  - Source: [@techniahqrobot on X](https://x.com/techniahqrobot/status/2095145553482891753)
- **Caterpillar's operational data.:** 0 units
  - _"Caterpillar has something AI startups can’t easily manufacture: millions of machines, job sites, failures, and operating hours."_
  - Source: [Wall St Engine on X: "CATERPILLAR PARTNERS WITH FIELDAI ON PHYSICAL AI

$CAT is working with FieldAI to bring AI-powered autonomy and robotics into jobsites and factories, using Caterpillar’s operational data, FieldAI’s robot foundation models and NVIDIA technologies.

Early applications include au… / X](https://x.com/CKCapitalxx/status/2095141910465618065)
- **valuation of Unitree:** $1,200
  - _"companies like Unitree reach valuations exceeding 1,200x earnings"_
  - Source: [@10xthinker on X](https://x.com/10xthinker/status/2095127748188741887)
- **Optimus Gen 2 cost estimation:** $45,500
  - _"Morgan Stanley estimates building an Optimus Gen 2 costs $45,500 using China's supply chain, versus $131,800 without it."_
  - Source: [@smsehy on X](https://x.com/smsehy/status/2095119584890085731)
- **Optimus Gen 2 cost estimation:** $45,500
  - _"Morgan Stanley estimates building an Optimus Gen 2 costs $45,500 using China's supply chain, versus $131,800 without it."_
  - Source: [@smsehy on X](https://x.com/smsehy/status/2095119584890085731)
- **Delta in cost due to supply chain differences:** $86,300
  - _"That $86,300 delta is not simply raw neodymium pricing."_
  - Source: [@smsehy on X](https://x.com/smsehy/status/2095119584890085731)
- **Qwen3.8-Max-0902:** 2,400 units
  - _"2.4T parameters."_
  - Source: [Chubby♨️ on X: "wtf Qwen 3.8 has been updated to version 0902 and is now almost at the same level as Fable 5 in benchmarks.

The interesting two aspects:

1) For such significant jumps, it seems that you don't even change the version number anymore (no Qwen 3.9), rather the development happens so q… / X](https://x.com/stretchcloud/status/2095220210999664739)
- **Qwen3.8-Max-0902:** 1,000,000 units
  - _"1M context tokens."_
  - Source: [Chubby♨️ on X: "wtf Qwen 3.8 has been updated to version 0902 and is now almost at the same level as Fable 5 in benchmarks.

The interesting two aspects:

1) For such significant jumps, it seems that you don't even change the version number anymore (no Qwen 3.9), rather the development happens so q… / X](https://x.com/stretchcloud/status/2095220210999664739)
- **badges for contributions in the AXIS campaign.:** 14 units
  - _"First, 5 new badges have been added, bringing the total to 14."_
  - Source: [Omai Leidi (3/3) on X: "Happy Wednesday frens👋

Everyone is watching the race to build better robots. 
I’m starting to think the real race is the data teaching those robots how to act.

I’ve been digging into @axisrobotics, and it changed how I look at the space.

At first glance, AXIS looks like… / X](https://x.com/OperationsPLS/status/2095167309207982333)
- **TRON 2:** 1 units
  - _"TRON 2 serves as a modular and extensible embodied robotic platform for multi-tool, multi-step tasks across large workspaces."_
  - Source: [LimX Dynamics on X: "Together with ZINOVA's Tool Intelligence, TRON 2 takes on increasingly complex construction workflows.

@ZiNovaLabs builds on TRON 2 to explore an innovative robotic configuration for construction, demonstrating key tasks in a scaled-down tilt-up construction workflow, includi… / X](https://x.com/chris_j_paxton/status/2095114381818253768)
- **Moby deployment in a semiconductor facility.:** 3 units
  - _"bring two Moby3 units to a semiconductor facility for material-handling workflows."_
  - Source: [Noble Machines Accelerates Humanoid Robot Development 3X | NVIDIA](https://x.com/NVIDIARobotics/status/2094923298589024698)
- **development timeline of Moby.:** $3
  - _"Accelerate Moby's development timeline by nearly 3x, from an expected four years to 18 months."_
  - Source: [Noble Machines Accelerates Humanoid Robot Development 3X | NVIDIA](https://x.com/NVIDIARobotics/status/2094923298589024698)
- **development timeline reduction.:** 18 hrs
  - _"from an initial estimated timeline of four years with 50 people to 18 months with 15 people."_
  - Source: [Noble Machines Accelerates Humanoid Robot Development 3X | NVIDIA](https://x.com/NVIDIARobotics/status/2094923298589024698)
- **performance comparison:** 2 units
  - _"2x faster than Opus 5 while using half the tokens"_
  - Source: [Yuchen Jin on X: "Even if most AI labs are benchmark-maxxing now, Fable 5.1 still looks like an insane jump.

From Anthropic’s blog:
- Found a 1-in-a-million crash Millennium’s team couldn’t explain for 4–5 years
- 2x faster than Opus 5 while using half the tokens

Big if true." / X](https://x.com/pstAsiatech/status/2095219413431529817)
- **Gemini models' efficiency in video analysis:** 88 units
  - _"They can now analyze videos with better accuracy while using up to 88% fewer tokens."_
  - Source: [Google DeepMind on X: "We’re bringing agentic video understanding to our latest Gemini models.

They can now analyze videos with better accuracy while using up to 88% fewer tokens. 🧵" / X](https://x.com/GoogleDeepMind/status/2094840179676660097)
- **Atlas model for spatial intelligence.:** 1 units
  - _"Atlas is our scalable foundation model that can perceive, generate, reason, and interact with virtual and physical worlds."_
  - Source: [World Labs on X: "All of this is powered by one unified architecture: a multimodal autoregressive diffusion transformer, pretrained from scratch.

This foundation blends the best of modern LLMs and video models, benefiting from the architectural, algorithmic, and systems advances from both areas." / X](https://x.com/theworldlabs/status/2094839768115036581)
- **Atlas:** 1 units
  - _"Atlas is built to scale: its performance improves with increased training compute, and we expect this trend to hold as we continue scaling."_
  - Source: [Atlas: A World Model for Spatial Intelligence | World Labs](https://x.com/theworldlabs/status/2094839769616589051)
- **Wafer's funding announcement.:** $40M
  - _"We’re excited to announce that we’ve raised a $40M Series A!"_
  - Source: [wafer on X: "We’re excited to announce that we’ve raised a $40M Series A!

co-led by @MarathonMP and @chemistry, with participation from @Wing_VC, @AMD Ventures, @outsetcap, @fiftyyears, and @ycombinator, and our existing investors doubling down on @wafer_ai. we are also joined by an incredible lis… / X](https://x.com/frontrunvc/status/2094841271244247488)
- **hardware design:** 1 units
  - _"The work demonstrates 1) hardware design, with a passive hook and a head-mounted solid-state lidar; 2) learning pipeline with teacher student learning; 3) sim2real designs for perception and actuation, including lidar noise patterns, temperature modelling and voltage modelling."_
  - Source: [C. Zhang on X: "New paper release: Learning Agile Perceptive Traversal of Sparse 3D Structures for Humanoids

Paper https://t.co/LlrDsajHqx
Videos https://t.co/jXLtG68oUg

We study how to do learning and sim2real for 3d traversal, such as monkey bars and overhanging obstacles." / X](https://x.com/chris_j_paxton/status/2094815833516060984)
- **estimated revenue in 2026:** $2,900,000,000
  - _"FactSet-estimated revenue ramp from RMB 2.9bn in 2026E to RMB 4.9bn in 2027E."_
  - Source: [@OptionKing666 on X](https://x.com/OptionKing666/status/2094638592253370646)
- **Google DeepMind's RT-2 model:** 2 units
  - _"it achieved nearly double the performance on novel tasks compared to its predecessor, RT-1."_
  - Source: [smartfitguide on X: "Google DeepMind's RT-2 model treats robot actions as text tokens. By co-fine-tuning pre-trained Vision-Language Models with robotic data, it achieved nearly double the performance on novel tasks compared to its predecessor, RT-1. #AI" / X](https://x.com/bareani21645/status/2094613978739593316)
- **AI chip inference cluster:** 100,000 units
  - _"It already has 100k domestic AI chip inference cluster & plan to increase significantly over next 6 months."_
  - Source: [tphuang on X: "GLM-6 disclosed - more parameter & lower activation ratio & inference cost & improved post training expected. Model to make self training decisions

Expecting significant increased domestic compute addition in next 3 to 6 months. Using self owned, leased & purchased c… / X](https://x.com/tphuang/status/2094601178684264502)
- **tokens processed:** 62,000,000 units
  - _"It did 62T tokens in 6 days w/ Ox Alpha"_
  - Source: [tphuang on X: "GLM-6 disclosed - more parameter & lower activation ratio & inference cost & improved post training expected. Model to make self training decisions

Expecting significant increased domestic compute addition in next 3 to 6 months. Using self owned, leased & purchased c… / X](https://x.com/tphuang/status/2094601178684264502)
- **AI chip inference cluster:** 100,000 units
  - _"It already has 100k domestic AI chip inference cluster & plan to increase significantly over next 6 months."_
  - Source: [tphuang on X: "How it views its own models in terms of the kind of tasks them can do. GLM-5.3 in terms of long time horizon work is still short of 1 wk, need at least 1 month to be autonomous.

Flash is the 1st one to use domestic chips & have low cost + intelligence to support long tasks." / X](https://x.com/tphuang/status/2094601170102747138)
- **tokens processed:** 62,000 units
  - _"It did 62T tokens in 6 days w/ Ox Alpha"_
  - Source: [tphuang on X: "How it views its own models in terms of the kind of tasks them can do. GLM-5.3 in terms of long time horizon work is still short of 1 wk, need at least 1 month to be autonomous.

Flash is the 1st one to use domestic chips & have low cost + intelligence to support long tasks." / X](https://x.com/tphuang/status/2094601170102747138)
- **market size for Autonomous AI:** $30,000,000,000
  - _"last step having 30T mkt size according to Show more"_
  - Source: [tphuang on X: "How it views its own models in terms of the kind of tasks them can do. GLM-5.3 in terms of long time horizon work is still short of 1 wk, need at least 1 month to be autonomous.

Flash is the 1st one to use domestic chips & have low cost + intelligence to support long tasks." / X](https://x.com/tphuang/status/2094601170102747138)
- **Funding for Skild AI:** $1.7billion
  - _"Skild has raised nearly $1.7 billion since its founding in 2023 to develop a general-purpose robot brain."_
  - Source: [The Robot Report: Skild AI unveils S1 robot foundation model | AI Understanding](https://x.com/aiuorg/status/2094602992565522608)
- **agentic video understanding:** $66
  - _"reduces costs by up to 66%"_
  - Source: [Introducing Agentic Video in Gemini](https://x.com/GoogleDeepMind/status/2094840182457422260)
- **agentic video understanding:** 88 units
  - _"cuts token consumption by up to 88%"_
  - Source: [Introducing Agentic Video in Gemini](https://x.com/GoogleDeepMind/status/2094840182457422260)
- **agentic video understanding:** 7 units
  - _"boosts quality by up to 7%"_
  - Source: [Introducing Agentic Video in Gemini](https://x.com/GoogleDeepMind/status/2094840182457422260)
- **Embodied AI model:** 36,000,000,000 units
  - _"A 36B parameter model just blurred one of robotics’ biggest boundaries, the gap between seeing a scene and acting on it."_
  - Source: [@techniahqrobot on X](https://x.com/techniahqrobot/status/2094550793743708256)
- **Schaeffler’s agreement with Humanoid:** 1,000,000 units
  - _"covering a seven-digit number of units."_
  - Source: [The hardest problem in physical AI may be the magnet, not the model](https://thenextweb.com/news/physical-ai-hardest-problem-magnet-not-model-crma-schaeffler)
- **humanoids shipped by Figure:** 1,000 units
  - _"has already shipped 1,000+ humanoids."_
  - Source: [@Astra1Byte on X](https://x.com/Astra1Byte/status/2094493801410662483)
- **@johnkoetsier 2094485667409760625:** $10B
  - _"$10B"_
  - Source: [NVIDIA's robot business is already worth $10 billion/year, and China is its best customer | HumanoidHub](https://x.com/johnkoetsier/status/2094485667409760625)
- **Qwen3.5-4B on Jetson Orin Nano:** 100 units
  - _"The example uses 100 samples and 30 steps so you can complete it quickly."_
  - Source: [Optimize Models with Unsloth | Jetson AI Lab](https://x.com/NVIDIARobotics/status/2094480283135316182)
- **Nemotron 3.5 Lightning on Jetson AGX Thor:** 44.8 hrs
  - _"On Jetson AGX Thor, the three training steps completed in 44.8 seconds and saved a 438 MB adapter."_
  - Source: [Optimize Models with Unsloth | Jetson AI Lab](https://x.com/NVIDIARobotics/status/2094480283135316182)
- **Model merge process:** 14 units
  - _"The tested merge produced 14 model shards totaling 65.8 GB."_
  - Source: [Optimize Models with Unsloth | Jetson AI Lab](https://x.com/NVIDIARobotics/status/2094480283135316182)
- **GGUF conversion and quantization:** $63.2
  - _"The tested export produced a 63.2 GB BF16 GGUF and a 24.5 GB Q4\K\M GGUF."_
  - Source: [Optimize Models with Unsloth | Jetson AI Lab](https://x.com/NVIDIARobotics/status/2094480283135316182)
- **Agibot's performance in the World Humanoid Robot Games.:** 46 units
  - _"Agibot finished #1 in both golds and total medals, with 18 gold, 16 silver and 12 bronze, 46 medals altogether."_
  - Source: [@ruima on X](https://x.com/ruima/status/2094468785591964066)
- **Gold medals won by Agibot.:** 18 units
  - _"Agibot finished #1 in both golds and total medals, with 18 gold, 16 silver and 12 bronze, 46 medals altogether."_
  - Source: [@ruima on X](https://x.com/ruima/status/2094468785591964066)
- **Silver medals won by Agibot.:** 16 units
  - _"Agibot finished #1 in both golds and total medals, with 18 gold, 16 silver and 12 bronze, 46 medals altogether."_
  - Source: [@ruima on X](https://x.com/ruima/status/2094468785591964066)
- **Bronze medals won by Agibot.:** 12 units
  - _"Agibot finished #1 in both golds and total medals, with 18 gold, 16 silver and 12 bronze, 46 medals altogether."_
  - Source: [@ruima on X](https://x.com/ruima/status/2094468785591964066)
- **humanoid robots shipped in 2025:** 5,500 units
  - _"unitree shipped 5,500 humanoid robots in 2025 – more than tesla, figure ai and agility robotics combined."_
  - Source: [@thehypedotnews on X](https://x.com/thehypedotnews/status/2094464209895424198)
- **total production of humanoid robots:** 6,500 units
  - _"total production exceeded 6,500 units."_
  - Source: [@thehypedotnews on X](https://x.com/thehypedotnews/status/2094464209895424198)
- **global market forecast for humanoid robots by 2027:** 100,000 units
  - _"counterpoint forecasts the global market hitting 100,000 units by 2027 – 6x the 2025 numbers."_
  - Source: [@thehypedotnews on X](https://x.com/thehypedotnews/status/2094464209895424198)
- **@0xconglomerate 2094460816947081633:** $399
  - _"$399"_
  - Source: [@0xconglomerate on X](https://x.com/0xconglomerate/status/2094460816947081633)
- **@0xconglomerate 2094460816947081633:** $1
  - _"$1 "_
  - Source: [@0xconglomerate on X](https://x.com/0xconglomerate/status/2094460816947081633)
- **@ycombinator fall cohort companies:** 5 units
  - _"5 @ycombinator fall cohort companies across Compute, Energy, Hardware and DevTools."_
  - Source: [@frontrunvc on X](https://x.com/frontrunvc/status/2094448223775494221)
- **tariffs on Chinese drones:** $100
  - _"The #US will impose up to 100% #tariffs on #Chinese #drones over 25kg, thermal models, and key parts starting Sept 3, with lighter drones facing a 25% rate."_
  - Source: [@whitee_rhinoo on X](https://x.com/whitee_rhinoo/status/2094444376655024384)
- **Demonstrations collected for the tasks.:** 100 units
  - _"For each task, we collect 100 teleoperated demonstrations at 30 FPS and fully fine-tune the pretrained π0.5 policy."_
  - Source: [GitHub - hku-sail/StreamPI: StreamPI: Streaming Multimodal Temporal Modeling for Vision-Language-Action Models · GitHub](https://x.com/DexRobotics/status/2094411289380765788)
- **humanoids:** 1,000 units
  - _"has built 1,000+ humanoids"_
  - Source: [Biti8 on X: "Four years ago, @Figure_robot didn’t even have a robot.

Today, it’s valued at $39B, has built 1,000+ humanoids, deployed robots at BMW, built its own AI models, manufacturing infrastructure and a global data collection network.

What impresses me most is the speed.

Figure isn’t build… / X](https://x.com/adcock_brett/status/2094462666798723307)
- **DJI Flip:** $351.2
  - _"The budget DJI you can still buy stateside."_
  - Source: [Pentagon blacklist vs. FCC drone bans: What drone pilots…](https://x.com/djirumor/status/2094361382112419970)
- **Number of humanoids built by Figure.:** 1,000 units
  - _"has built more than 1,000 humanoids"_
  - Source: [How Figure Became the Biggest Name in Robotics | XMAQUINA DAO](https://x.com/xmaquina/status/2094334389979373824)
- **Number of BMW X3s produced with Figure's help.:** 30,000 units
  - _"contributed to the production of 30,000 BMW X3s."_
  - Source: [How Figure Became the Biggest Name in Robotics | XMAQUINA DAO](https://x.com/xmaquina/status/2094334389979373824)
- **Number of downloads of the Index platform.:** 264,000 units
  - _"Index had already reached 264,000 downloads across 108 countries"_
  - Source: [How Figure Became the Biggest Name in Robotics | XMAQUINA DAO](https://x.com/xmaquina/status/2094334389979373824)
- **Total number of videos collected by Index.:** 16,000,000 units
  - _"collected 16 million videos."_
  - Source: [How Figure Became the Biggest Name in Robotics | XMAQUINA DAO](https://x.com/xmaquina/status/2094334389979373824)
- **humanoid robots shipped globally in 2026:** 22,000 units
  - _"more than 22,000 humanoid robots shipped globally in the first half of 2026"_
  - Source: [@tchsignal on X](https://x.com/tchsignal/status/2094323502098911410)
- **Performance metric for model.:** 99.5 units
  - _"99.5 Model Performance"_
  - Source: [Physical AI Platform | Industrial Automation Solutions | Sanctuary AI](https://x.com/heetezition/status/2094280429264420968)
- **Cycle time for completing a plug insertion.:** 2.54 units
  - _"2.54 sec Cycle Time, Per Plug"_
  - Source: [Physical AI Platform | Industrial Automation Solutions | Sanctuary AI](https://x.com/heetezition/status/2094280429264420968)
- **gemini-robotics-er-2-preview paid price input per 1M tokens:** $2.00
  - _"ER 2 Preview paid $2.00 input / $10.00 output per 1M tokens"_
  - Source: [Gemini Robotics ER 1.6 shutdown date Aug 31, 2026; replacement ER 2](https://ai.google.dev/gemini-api/docs/deprecations)
- **gemini-robotics-er-2-preview paid price output per 1M tokens:** $2.00
  - _"ER 2 Preview paid $2.00 input / $10.00 output per 1M tokens"_
  - Source: [Gemini Robotics ER 1.6 shutdown date Aug 31, 2026; replacement ER 2](https://ai.google.dev/gemini-api/docs/deprecations)
- **AI spending by a large consulting firm.:** $100
  - _"a large consulting firm is increasing their AI spend by 100x at their London office for their back office teams."_
  - Source: [@ParadisLabs on X](https://x.com/ParadisLabs/status/2094203645877493923)
- **NVIDIA earnings related to AI.:** $25
  - _"AI Clouds, Industrial & Enterprise revenue was up 25% sequentially compared to 13% for the hyperscalers."_
  - Source: [@ParadisLabs on X](https://x.com/ParadisLabs/status/2094203645877493923)
- **teleoperation hours:** 100000 hrs
  - _"Skild says the equivalent training data took to collect."_
  - Source: [@Robot_AIsignals on X](https://x.com/Robot_AIsignals/status/2094201246987034914)
- **average per-step success:** 66 units
  - _"Skild reports 66% average per-step success on unseen tasks against 9% for a language-prompted baseline, once pre-training reaches 100,000 hours (company claim, internal evaluation)."_
  - Source: [@Robot_AIsignals on X](https://x.com/Robot_AIsignals/status/2094201246987034914)
- **XPENG's robotics funding:** $900million
  - _"Its robotics business just raised more than $900 million, reaching a post-money valuation above $6.3 billion."_
  - Source: [@techniahqrobot on X](https://x.com/techniahqrobot/status/2093739174746611828)
- **Personal investment by XPENG founder and co-president:** $100million
  - _"XPENG founder He Xiaopeng and co-president Brian Gu also invested roughly $100 million personally."_
  - Source: [@techniahqrobot on X](https://x.com/techniahqrobot/status/2093739174746611828)
- **Humanoid robots procurement:** 1,080 units
  - _"The government plans to buy 250 domestically produced humanoid units in 2027, expanding to 1,080 units through 2030."_
  - Source: [South Korea Commits KRW 2.3 Trillion to Build Full-Stack Humanoid Robotics Ecosystem by 2030 | Humanoids Daily](https://x.com/humanoidsdaily/status/2093648555726590278)
- **Total robotic systems procurement:** 5,000 units
  - _"total state purchases will reach 1,700 units in 2027 and approximately 5,000 units by 2030."_
  - Source: [South Korea Commits KRW 2.3 Trillion to Build Full-Stack Humanoid Robotics Ecosystem by 2030 | Humanoids Daily](https://x.com/humanoidsdaily/status/2093648555726590278)
- **model size:** 0.5 units
  - _"모델은 π0.5다."_
  - Source: [@realgalleryx on X](https://x.com/realgalleryx/status/2093635683323891741)
- **Gemini Robotics 2 (VLA):** 1 units
  - _"three models: Gemini Robotics 2 (VLA) converts vision+language into motor control first DeepMind VLA to control a full humanoid feet-to-fingertips under one checkpoint."_
  - Source: [@csentropy on X](https://x.com/csentropy/status/2093431495277043724)
- **Gemini Robotics ER 2 (VLM):** 1 units
  - _"Gemini Robotics ER 2 (VLM) the embodied-reasoning brain; publicly available."_
  - Source: [@csentropy on X](https://x.com/csentropy/status/2093431495277043724)
- **Gemini Robotics On-Device 2 (VLA):** 1 units
  - _"Gemini Robotics On-Device 2 (VLA) runs locally, adapts to new embodiments in a few hours."_
  - Source: [@csentropy on X](https://x.com/csentropy/status/2093431495277043724)
- **NBT average for RandER.:** 0.197 units
  - _"RandER: 0.197 ± 0.020 (SEM)"_
  - Source: [MemoryAnchors](https://x.com/du_maximilian/status/2093390863821131915)
- **NBT average with 1% Memory Anchors removed.:** 0.171 units
  - _"−1%: 0.171 ± 0.018 (SEM)"_
  - Source: [MemoryAnchors](https://x.com/du_maximilian/status/2093390863821131915)
- **NBT average with 5% Memory Anchors removed.:** 0.284 units
  - _"−5%: 0.284 ± 0.027 (SEM)"_
  - Source: [MemoryAnchors](https://x.com/du_maximilian/status/2093390863821131915)
- **NBT average with 10% Memory Anchors removed.:** 0 units
  - _"−10%: 0"_
  - Source: [MemoryAnchors](https://x.com/du_maximilian/status/2093390863821131915)
- **@0xconglomerate 2093373586803478854:** $399
  - _"$399"_
  - Source: [@0xconglomerate on X](https://x.com/0xconglomerate/status/2093373586803478854)
- **@0xconglomerate 2093373586803478854:** $1
  - _"$1 "_
  - Source: [@0xconglomerate on X](https://x.com/0xconglomerate/status/2093373586803478854)
- **Current hours of available data for training robots.:** 2000 hrs
  - _"Today, the industry has roughly 2,000 hours, the best public dataset, Open X-Embodiment."_
  - Source: [Cicada Market Making on X: "https://t.co/26k9SK56mf" / X](https://x.com/axisrobotics/status/2093684446465798629)
- **Number of global contributors and collected trajectories on the platform.:** 150,000 units
  - _"As of late August 2026, the platform has surpassed 150,000 global contributors and collected over 3.7 million trajectories across 4,000+ published tasks."_
  - Source: [Cicada Market Making on X: "https://t.co/26k9SK56mf" / X](https://x.com/axisrobotics/status/2093684446465798629)
- **Current throughput of valid trajectories collected per hour.:** 10,000 units
  - _"Throughput already reaches 10,000 valid trajectories per hour today;"_
  - Source: [Cicada Market Making on X: "https://t.co/26k9SK56mf" / X](https://x.com/axisrobotics/status/2093684446465798629)
- **Target valid hours of data per month from the Mobile Ego-Centric App.:** 10,000 units
  - _"Launching in September 2026 with a target of 10,000+ valid hours of data per month;"_
  - Source: [Cicada Market Making on X: "https://t.co/26k9SK56mf" / X](https://x.com/axisrobotics/status/2093684446465798629)
- **memory manufacturer capacity:** 3 units
  - _"Hearing that a leading memory manufacturer would need three entire years of capacity just to meet current demand is mind-blowing."_
  - Source: [Why Top Founders Are Racing Into AI Infrastructure - YouTube](https://x.com/a16z/status/2093340813174747432)
- **Cerebras WSEs needed for running a large model.:** $20M
  - _"that’s over $20M of CAPEX and 1MW of power consumption before you can get a forward pass on a frontier model."_
  - Source: [Cerebras's Next Generation CS-4: Fast Just Got Faster](https://x.com/SemiAnalysis_/status/2093126657682907547)
- **Cerebras WSEs needed for reasonable concurrency of 256 requests.:** 40 units
  - _"its around 40 systems."_
  - Source: [Cerebras's Next Generation CS-4: Fast Just Got Faster](https://x.com/SemiAnalysis_/status/2093126657682907547)
- **Direct wafer paths reduction in latency.:** 2 hrs
  - _"This makes it easier for the CS-4 to interface with other systems for disaggregated inference setups."_
  - Source: [Cerebras's Next Generation CS-4: Fast Just Got Faster](https://x.com/SemiAnalysis_/status/2093126657682907547)
- **future revenue from Optimus:** 80 units
  - _"80% of the company's value will come from its Optimus robot in the future."_
  - Source: [Optimus Just Entered Production at Fremont. Here's What Changes for Tesla Investors. | The Motley Fool](https://x.com/VanEmmerickKris/status/2094940107484586346)
- **status of physical AI companies:** 0 units
  - _"Most Physical AI companies are still doing lab demos."_
  - Source: [@Rewkang on X](https://x.com/Rewkang/status/2093056014967087389)
- **napkin production:** 1,500 units
  - _"To keep one Din Tai Fung dining room supplied, a robot has to produce 1,500 table-ready napkins in a day’s shift."_
  - Source: [Dyna Robotics on X: "After so many demos, models, pilots, robots are still struggling to land real deployments with real customers. Until now.

Today we’re excited to share that our robots have successfully crossed the ROI threshold, and Din Tai Fung, one of the highest revenue per location restaur… / X](https://x.com/TheHumanoidHub/status/2093387536962076683)
- **napkin production:** 480 units
  - _"A year ago, ours managed 480 with Dyna-1."_
  - Source: [Dyna Robotics on X: "After so many demos, models, pilots, robots are still struggling to land real deployments with real customers. Until now.

Today we’re excited to share that our robots have successfully crossed the ROI threshold, and Din Tai Fung, one of the highest revenue per location restaur… / X](https://x.com/TheHumanoidHub/status/2093387536962076683)
- **production requirement for napkins:** 1,500 units
  - _"To keep one Din Tai Fung dining room supplied, a robot has to produce 1,500 table-ready napkins in a day’s shift."_
  - Source: [Dyna Robotics on X: "After so many demos, models, pilots, robots are still struggling to land real deployments with real customers. Until now.

Today we’re excited to share that our robots have successfully crossed the ROI threshold, and Din Tai Fung, one of the highest revenue per location restaur… / X](https://x.com/Rewkang/status/2093048258537087479)
- **teleoperation reduction:** 5884 hrs
  - _"cut the teleoperation needed to hit their action-loss target from 5,884 hours to 28."_
  - Source: [@TheHumanoidHub on X](https://x.com/TheHumanoidHub/status/2092974952769048876)
- **target action-loss:** 28 hrs
  - _"cut the teleoperation needed to hit their action-loss target from 5,884 hours to 28."_
  - Source: [@TheHumanoidHub on X](https://x.com/TheHumanoidHub/status/2092974952769048876)
- **2026 BMW M2 Base Coupe:** $77,650
  - _"MSRP$77,650Dealer Service Fee$900ETR Fee$199Transparent Price"_
  - Source: [New BMW Cars & SUVs | BMW Dealer Serving Atlanta GA](https://x.com/BMWofAtlanta/status/2092960092014391750)
- **2026 BMW X5 xDrive40i SUV:** $84,095
  - _"MSRP$84,095Dealer Service Fee$900ETR Fee$199Transparent Price"_
  - Source: [New BMW Cars & SUVs | BMW Dealer Serving Atlanta GA](https://x.com/BMWofAtlanta/status/2092960092014391750)
- **2026 BMW 750e xDrive Sedan:** $122,290
  - _"MSRP$122,290Dealer Service Fee$900ETR Fee$199Transparent Price"_
  - Source: [New BMW Cars & SUVs | BMW Dealer Serving Atlanta GA](https://x.com/BMWofAtlanta/status/2092960092014391750)
- **2027 BMW X7 xDrive40i SUV:** $104,110
  - _"MSRP$104,110Dealer Service Fee$900ETR Fee$199Transparent Price"_
  - Source: [New BMW Cars & SUVs | BMW Dealer Serving Atlanta GA](https://x.com/BMWofAtlanta/status/2092960092014391750)
- **2027 BMW X7 xDrive40i SUV:** $99,960
  - _"MSRP$99,960Dealer Service Fee$900ETR Fee$199Transparent Price"_
  - Source: [New BMW Cars & SUVs | BMW Dealer Serving Atlanta GA](https://x.com/BMWofAtlanta/status/2092960092014391750)
- **filtered simulation episodes:** 180.55 hrs
  - _"42,046 episodes and 3,249,835 frames remained, equivalent to 180.55 hours at 5 Hz."_
  - Source: [Axis Robotics × Booster: From Digital Twins to a Robot Data Engine](https://x.com/axisrobotics/status/2092884880430661971)
- **SAGE-10K dataset:** 10,000 units
  - _"The SAGE-10K dataset contains 10,000 generated indoor scenes across 50 room types."_
  - Source: [How to Train a Cross-Embodiment Robot Navigation Policy with AI Agents | NVIDIA Technical Blog](https://x.com/NVIDIARobotics/status/2093381258243895657)
- **GEN-1.5 model capabilities:** 1 units
  - _"Introducing GEN-1.5, a one-shot learner. It can learn new tasks in a few seconds."_
  - Source: [Rhys on X: "Cool to see a bunch of new robotics foundation model releases:

π0.7 from @physical_int 
GEN-1.5 from @generalistAI 
S1 from @SkildAI 

Feels like the GPT-3 moment for robotics. Robot Models are Few-Shot Learners!" / X](https://x.com/RhysLindmark/status/2092684832199639136)
- **S1 model capabilities:** 1 units
  - _"Introducing S1, our new foundation model that learns from one example."_
  - Source: [Rhys on X: "Cool to see a bunch of new robotics foundation model releases:

π0.7 from @physical_int 
GEN-1.5 from @generalistAI 
S1 from @SkildAI 

Feels like the GPT-3 moment for robotics. Robot Models are Few-Shot Learners!" / X](https://x.com/RhysLindmark/status/2092684832199639136)
- **weekly active users for data collection:** 43,000 units
  - _"We're now over 43,000 weekly active users collecting data to train Helix, our AI model for F.03 robots"_
  - Source: [@adcock_brett on X](https://x.com/adcock_brett/status/2092628102262477268)
- **robot training:** 0 hrs
  - _"the robot couldn't learn from video alone. failed over and over."_
  - Source: [@antopatrex1 on X](https://x.com/antopatrex1/status/2092616347792580997)
- **robot deployment:** 0 units
  - _"you need to deploy robots for robots to get better."_
  - Source: [@antopatrex1 on X](https://x.com/antopatrex1/status/2092616347792580997)
- **Unitree G1 pricing:** $13.5K
  - _"Unitree G1 Price from $13.5K."_
  - Source: [@UnitreeStore_SH on X](https://x.com/UnitreeStore_SH/status/2092555273936503147)
- **Gemini Robotics ER:** 2 units
  - _"I believe Gemini Robotics ER 2 is a clear example here."_
  - Source: [@zhodonx on X](https://x.com/zhodonx/status/2092555239282843992)
- **Unitree H2:** $29,900
  - _"A 29900$ MACHINE IS DANCING ON A PAVEMENT AND NOT ONE PERSON WALKING PAST LOOKS UP."_
  - Source: [@dredgefactory on X](https://x.com/dredgefactory/status/2092395639384539542)
- **entry models of Unitree H2:** $.
  - _"Entry models go for 13500$."_
  - Source: [@dredgefactory on X](https://x.com/dredgefactory/status/2092395639384539542)
- **data efficiency comparison:** 380 units
  - _"Skild S1 is 380 times more data efficient than standard VLA models!"_
  - Source: [@gupta_abhinav_ on X](https://x.com/gupta_abhinav_/status/2092331051041612145)
- **robot learning economics:** $1
  - _"you pay the enormous data bill once during foundation-model training, then amortize it across thousands of new tasks through prompting."_
  - Source: [@rohanpaul_ai on X](https://x.com/rohanpaul_ai/status/2092312459382235407)
- **S1's performance on manipulation tasks:** $380
  - _"one video demonstration provided roughly the benefit of 380 post-training examples!"_
  - Source: [@DeryaTR_ on X](https://x.com/DeryaTR_/status/2092312053272707504)
- **general purpose robot foundation model:** 10 units
  - _"a really impressive, ten minute long one from skild."_
  - Source: [@chris_j_paxton on X](https://x.com/chris_j_paxton/status/2092304651995529392)
- **@humanoidsdaily 2092317847645032776:** $1B
  - _"$1B"_
  - Source: [Brett Adcock on X: "Introducing Index

Today we're coming out of stealth with Index, the largest & most diverse robot dataset in the world

→ 30min of video uploads/sec
→ 16M video uploads
→ Paid $15M to date
→ 264k downloads

We're committed to spending $1B the next 12 months on data & com… / X](https://x.com/humanoidsdaily/status/2092317847645032776)
- **total payment:** $15M
  - _"→ Paid $15M to date"_
  - Source: [Brett Adcock on X: "Introducing Index

Today we're coming out of stealth with Index, the largest & most diverse robot dataset in the world

→ 30min of video uploads/sec
→ 16M video uploads
→ Paid $15M to date
→ 264k downloads

We're committed to spending $1B the next 12 months on data & com… / X](https://x.com/HumanoidInvest/status/2092397968041066740)
- **video uploads:** 16,000,000 units
  - _"uploaded 16 million videos"_
  - Source: [Figure on X: "Introducing Index

Today we're coming out of stealth with the largest and most diverse robot dataset to scale general purpose robots

To date we've crossed 264,000 app downloads and uploaded 16 million videos" / X](https://x.com/Figure_robot/status/2092303621392376314)
- **S1 robot foundation model:** 1 units
  - _"Skild has released S1, a robot foundation model where you specify the task with a video demonstration instead of a language instruction."_
  - Source: [@TheHumanoidHub on X](https://x.com/TheHumanoidHub/status/2092301463989203250)
- **post-training duration:** 50 hrs
  - _"We found that to match the accuracy that S1 can achieve with just one example of prompting, current VLA models would need to be post-trained with 50-100 hours of data collection followed by fine-tuning!"_
  - Source: [Skild AI on X: "We found that to match the accuracy that S1 can achieve with just one example of prompting, current VLA models would need to be post-trained with 50-100 hours of data collection followed by fine-tuning!" / X](https://x.com/SkildAI/status/2092300859317330378)
- **performance of S1 compared to VLAs:** 0 units
  - _"For known tasks, S1 can match the performance of language-prompted VLAs."_
  - Source: [@SkildAI on X](https://x.com/SkildAI/status/2092300854359625814)
- **performance of S1 on novel tasks:** 0 units
  - _"For novel tasks, S1 exponentially outperforms any existing language-prompted VLAs as we scale the pre-training."_
  - Source: [@SkildAI on X](https://x.com/SkildAI/status/2092300854359625814)
- **S1 compared to current VLA models:** 100 hrs
  - _"current VLA models would need to be post-trained with 50-100 hours of data collection followed by fine-tuning!"_
  - Source: [Skild AI on X: "S1 learns new tasks like a language model.

You prompt it with a video demonstration, and it outputs robot actions to complete the task in any environment and in any embodiment." / X](https://x.com/SkildAI/status/2092300850492481818)
- **S1's accuracy compared to conventional VLA models.:** 100 hrs
  - _"current VLA models would need to be post-trained with 50-100 hours of data collection followed by fine-tuning!"_
  - Source: [Skild AI on X: "The first time S1 flipped a pancake, we assumed pancake flipping must have been in its pre-training data.

We searched our whole pre-training data and found no examples of flipping.

S1 inferred the out-of-distribution task from one video prompt." / X](https://x.com/SkildAI/status/2092300848034635790)
- **S1's accuracy with one example compared to current VLA models:** 50 hrs
  - _"current VLA models would need to be post-trained with 50-100 hours of data collection followed by fine-tuning!"_
  - Source: [Skild AI on X: "Introducing S1, our new foundation model that learns from one example.

It can be taught 10-minute long tasks that it has never seen before, from one video prompt without any fine-tuning.

Watch S1 operate in real-time via in-context learning:" / X](https://x.com/Scobleizer/status/2092306192773345691)
- **training needed for current VLA models to match S1's accuracy:** 100 hrs
  - _"current VLA models would need to be post-trained with 50-100 hours of data collection followed by fine-tuning!"_
  - Source: [Skild AI on X: "Introducing S1, our new foundation model that learns from one example.

It can be taught 10-minute long tasks that it has never seen before, from one video prompt without any fine-tuning.

Watch S1 operate in real-time via in-context learning:" / X](https://x.com/SkildAI/status/2092300842900865389)
- **cost of infrastructure:** 7 units
  - _"NEURA Gym delivers it at up to 7x lower cost."_
  - Source: [NEURA Gym: Physical AI Training and Infrastructure | NEURA](https://x.com/NEURARobotics/status/2092183035747700844)
- **app downloads:** 264,000 units
  - _"we've crossed 264,000 app downloads across 100+ countries"_
  - Source: [Introducing Index: Building The World’s Largest and Most Diverse Physical Dataset](https://www.figure.ai/news/introducing-index)
- **robot interaction data from the DROID dataset:** 62 hrs
  - _"Meta reports using 62 hours of robot interaction data from the DROID dataset for this stage."_
  - Source: [🤖🍨 Sundae Robotics 06: V-JEPA 2 & Predicting Physical Intelligence · Luma](https://x.com/EdmondIsARobot/status/2092037991619530779)
- **training hours for LDA-1B model:** 30000 hrs
  - _"LDA-1B s a new foundation model that is trained on 30,000 hours of human and robot interaction data."_
  - Source: [RoboPapers on X: "Instead of choosing between training a world model and training a language conditioned robot policy, why not do both? LDA-1B s a new foundation model that is trained on 30,000 hours of human and robot interaction data.

Part of the secret is that LDA-1B jointly learns forward … / X](https://x.com/chris_j_paxton/status/2091877313872834800)
- **@RoboPapers 2091877173292356012:** 30000 hrs
  - _"30,000 hours"_
  - Source: [@RoboPapers on X](https://x.com/RoboPapers/status/2091877173292356012)
- **robot programming:** 1 units
  - _"If showing it once is enough, that changes both how fast a robot becomes useful and who can work with one."_
  - Source: [@TheHumanoidHub on X](https://x.com/TheHumanoidHub/status/2090662734551818253)
- **GEN-1.5 can learn new tasks in a few seconds.:** 1 units
  - _"Introducing GEN-1.5, a one-shot learner."_
  - Source: [Anirudha Majumdar on X: "Very cool work, Pete! I'm curious if you've tried an extreme version of this: no in-context example at all; just place objects in front of the robot and see if it can infer the task. 

I suspect this will have non-trivial success rates, which would also allow you to figure … / X](https://x.com/peteflorence/status/2090592251286335962)
- **validated run uses 64 nodes of 4× GB200 for 60K iterations, roughly 68 hours (17.4K GB200-hours).:** 68 hrs
  - _"Plan compute accordingly."_
  - Source: [Post-Train NVIDIA Cosmos 3 Edge for On-Device Robot Control | NVIDIA Technical Blog](https://x.com/NVIDIARobotics/status/2095179993974260168)
- **tasks available in RoboLab.:** 120 units
  - _"It executes each action chunk in physics and streams rendered observations back for a true closed loop across 120 language-conditioned manipulation tasks."_
  - Source: [Post-Train NVIDIA Cosmos 3 Edge for On-Device Robot Control | NVIDIA Technical Blog](https://x.com/NVIDIARobotics/status/2095179993974260168)
- **success rate of the post-trained policy.:** 22.9 units
  - _"In closed-loop RoboLab evaluation, the post-trained Edge policy reaches 22.9% success across 120 language-conditioned manipulation tasks."_
  - Source: [Post-Train NVIDIA Cosmos 3 Edge for On-Device Robot Control | NVIDIA Technical Blog](https://x.com/NVIDIARobotics/status/2095179993974260168)
- **training video hours:** 2,202 units
  - _"2,202 h filtered multi-embodiment training video"_
  - Source: [Hydra-0: Action Flow for Generalist World Modeling and Control | NVIDIA Isaac](https://x.com/NVIDIARobotics/status/2094515571501015339)
- **episodes in RoboLab evaluation:** 300 units
  - _"5 policies × 6 tasks · 300 episodes"_
  - Source: [Hydra-0: Action Flow for Generalist World Modeling and Control | NVIDIA Isaac](https://x.com/NVIDIARobotics/status/2094515571501015339)
- **speedup in generative performance:** 16 units
  - _"16.0× generation-only speedup after few-step distillation"_
  - Source: [Hydra-0: Action Flow for Generalist World Modeling and Control | NVIDIA Isaac](https://x.com/NVIDIARobotics/status/2094515571501015339)
- **BMW Group production:** 30,000 units
  - _"the Figure 02 robot supported the production of more than 30,000 BMW X3 vehicles over ten months."_
  - Source: [Press-Information June 25th 2026](https://www.press.bmwgroup.com/canada/article/attachment/T0458794EN/651051)
- **Scale AI data delivery:** 150000 hrs
  - _"it reported delivering over 150,000 hours of physical AI data during 2025"_
  - Source: [5 Physical AI infrastructure platforms shaping robotics in 2026 - The Robot Report](https://www.therobotreport.com/5-physical-ai-infrastructure-platforms-shaping-robotics-in-2026)
- **physical AI data delivered:** 150000 hrs
  - _"it reported delivering over 150,000 hours of physical AI data during 2025"_
  - Source: [5 Physical AI infrastructure platforms shaping robotics in 2026 - The Robot Report](https://x.com/therobotreport/status/2094493447029670279)
- **examples for adapting to new robotic embodiments.:** 200 units
  - _"we can now adapt to new bi-arm robot embodiments with just a few hours of adaptation time, typically with less than 200 examples."_
  - Source: [Gemini Robotics 2 brings whole body intelligence to robots — Google DeepMind](https://x.com/SkillupAi/status/2093896499465945235)
- **adaptation time for new robot embodiments:** 200 hrs
  - _"We can now adapt to new bi-arm robot embodiments with just a few hours of adaptation time, typically with less than 200 examples."_
  - Source: [Gemini Robotics 2 brings whole body intelligence to robots — Google DeepMind](https://x.com/TheTechFuse/status/2094178345982607455)
- **prototype:** 1 units
  - _"A working prototype creates trust faster than a perfect deck."_
  - Source: [Sylvia Li on X: "Spent the day at Deep Tech Week and left with one strong conviction:

the next wave of AI will not be won only by better models.

It will be won by the infrastructure that lets intelligence touch the physical world.

For anyone building in robotics, chips, manufacturing, energy, or… / X](https://x.com/smsehy/status/2095050189354742023)
- **G1 humanoid grasp and lift task:** 10 units
  - _"G1 humanoid · grasp & lift · ten successes in a row"_
  - Source: [Coupled Local and Global World Models for Efficient First Order RL](https://coupled-global-local-wm-rl.pages.dev/)
- **successful grasps by the G1 humanoid on hardware.:** 10 units
  - _"G1 humanoid · grasp & lift · ten successes in a row"_
  - Source: [Coupled Local and Global World Models for Efficient First Order RL](https://x.com/RoboPapers/status/2092642979466915921)
- **AMI Labs funding:** $1.03billion
  - _"AMI Labs, the new venture co-founded by Turing Award winner Yann LeCun after he left Meta, has raised $1.03 billion at a $3.5 billion pre-money valuation."_
  - Source: [Yann LeCun's AMI Labs raises $1.03B to build world models | TechCrunch](https://techcrunch.com/2026/03/09/yann-lecuns-ami-labs-raises-1-03-billion-to-build-world-models)
- **Amount raised by AMI Labs in funding.:** $890,000,000
  - _"The French AI lab was reportedly seeking just €500 million last December, but ended up raising some €890 million, likely thanks to its team."_
  - Source: [Yann LeCun's AMI Labs raises $1.03B to build world models | TechCrunch](https://x.com/MikeTamir/status/2093444518494237175)
- **deployments of Spot:** 1,500 units
  - _"With over 1,500 deployments, Spot is already teaching hundreds of companies how to work alongside autonomous mobile robots."_
  - Source: [An Electric New Era for Atlas | Boston Dynamics](https://x.com/BostonDynamics/status/1780603212359205323)
- **Training Data:** 20000 hrs
  - _"N1.7 is pretrained on 20K hours of EgoScale human video data alongside diverse robot demonstrations."_
  - Source: [NVIDIA/Isaac-GR00T](https://github.com/NVIDIA/Isaac-GR00T)
- **the project stand not just as a piece of marketing, but as a vision statement: a world built by the very technology that made it possible.:** 1 units
  - _"The production fused artistry, engineering, and emotion into a single workflow, revealing what’s possible when imagination becomes interactive."_
  - Source: [Bringing Marble to Life | World Labs](https://www.worldlabs.ai/case-studies/bringing-marble-to-life)
- **DROID dataset robot interaction data:** 62 hrs
  - _"Meta reports using 62 hours of robot interaction data from the DROID dataset for this stage."_
  - Source: [🤖🍨 Sundae Robotics 06: V-JEPA 2 & Predicting Physical Intelligence · Luma](https://luma.com/5h9owhri)
- **humanoids built:** 1,000 units
  - _"has built more than 1,000 humanoids"_
  - Source: [How Figure Became the Biggest Name in Robotics | XMAQUINA DAO](https://www.xmaquina.io/blog/how-figure-became-the-biggest-name-in-robotics)
- **BMW X3 production:** 30,000 units
  - _"contributed to the production of 30,000 BMW X3s"_
  - Source: [How Figure Became the Biggest Name in Robotics | XMAQUINA DAO](https://www.xmaquina.io/blog/how-figure-became-the-biggest-name-in-robotics)
- **AgiBot World platform comprises over 1 million trajectories.:** 1,000,000 units
  - _"we achieve an order-of-magnitude increase in data scale compared to existing datasets."_
  - Source: [AgiBot World Colosseo: A Large-scale Manipulation Platform](https://arxiv.org/pdf/2503.06669)
- **number of objects in AgiBot World.:** 3,000 units
  - _"3,000+ Objects"_
  - Source: [AgiBot World Colosseo: A Large-scale Manipulation Platform](https://arxiv.org/pdf/2503.06669)
- **performance improvement of the proposed policy over prior arts.:** 32 units
  - _"it is trained across diverse data corpus with a scalable performance of 32% gain compared to prior arts."_
  - Source: [AgiBot World Colosseo: A Large-scale Manipulation Platform](https://arxiv.org/pdf/2503.06669)
- **average performance improvement of policies using the dataset.:** 30 units
  - _"Policies pre-trained on our dataset achieve an average performance improvement of 30%"_
  - Source: [AgiBot World Colosseo: A Large-scale Manipulation Platform](https://arxiv.org/pdf/2503.06669)
- **video generation:** 1 units
  - _"Atlas generates images and videos from one or more images with pixel-perfect camera control, outputting up to 1 minute of video at 1440p."_
  - Source: [World Labs Atlas](https://www.worldlabs.ai/blog/atlas)

## Pulse

- [@ramez on X](https://x.com/ramez/status/2095992466847515034)
  This is an excellent move by @OpenAI . Bad actors will have powerful AI models that refuse nothing. Using frontier AI to secure the Internet, our devices, and our infrastructure is an essential part o…
- [@GalbotRobotics on X](https://x.com/GalbotRobotics/status/2095982381307969568)
  Galbot ET1 — the world’s first autonomous learning-capable agentic humanoid robot — is now available for pre-order on https://t.co/92OEbTIoiT! A Native Agent for the Physical World, ET1 is built to se…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095981729026642080)
  The Gemini 4 internal benchmarks leaked three weeks ago showing it outperforming GPT-5.6 Sol and Fable 5 on coding and math. Then this week Google shipped Gemini 3.8 Flash, their third Flash variant i…
- [@SemiAnalysis_ on X](https://x.com/SemiAnalysis_/status/2095980532186140692)
  A (popcorn) kernel is a specialized seed from a variety of corn that can burst open and turn inside out when heated. A kernel is also a few hundred lines of GPU code that executes the math and attenti…
- [@MelvinInvests on X](https://x.com/MelvinInvests/status/2095975579736838521)
  AI demand is just getting started,and we need more compute than ever before. The chart tracks the share of US businesses paying for subscriptions to AI models, platforms, and tools, based on business …
- [@a16z on X](https://x.com/a16z/status/2095972688200372368)
  >blogs about ray tracing for a CS project in 2012 >co-founds World Labs in 2024 >ships Atlas in 2026, a spatial intelligence model that turns a few photos into a whole 3D world curiosity comp…
- [@a16z on X](https://x.com/a16z/status/2095961376208646248)
  >be dr. fei-fei li
>born in beijing, raised in chengdu
>dad moves to new jersey, follow at age 16
>speak almost no english
>parents open a dry cleaner
>work the counter on weekends
>get into princeton…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095956311498252493)
  The Astra rollout is a useful case study in how not to ship a flagship feature. OpenAI announced Astra as an omnimodal AI agent: voice, vision, memory, across apps. The pitch was compelling enough to …
- [@NVIDIARobotics on X](https://x.com/NVIDIARobotics/status/2095950021636046876)
  Interested in getting hands-on with Newton? The newly released Newton Learning Path will help you set up a ready-to-run environment, dive into Newton abstractions, and show you how to build your first…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095946245164454137)
  Qwen3.8-Flash-Next just made the 180B parameter class meaningfully accessible. The architecture is worth understanding: 125B language model parameters, 55B auxiliary, but only 6B activate per token. T…
- [@chris_j_paxton on X](https://x.com/chris_j_paxton/status/2095945020885160345)
  There's a good chance we end up with one class of models for everything https://t.co/82nF97aaeh
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095941212112904432)
  The image model race just made a clear separation between quality and price leadership. My read on what MAI-Image-2.6-Flash means: Microsoft now owns both. MAI-Image-2.6 is #2 on the Artificial Analys…
- [@a16z on X](https://x.com/a16z/status/2095940012932215128)
  World Labs co-founder Justin Johnson says Atlas can recreate the famous "Bullet Time" shot from The Matrix with iPhones: "The way they did that shot is they had a ring of hundreds of cameras. [Neo] fe…
- [@drfeifei on X](https://x.com/drfeifei/status/2095926761305575826)
  Next view prediction is the key to Atlas, enabling us to unify pixel-level generation and reconstruction. @jcjohnss @BenMildenhall @martin_casado and I had a deeper discussion on some of the most exci…
- [@ramez on X](https://x.com/ramez/status/2095924647938150575)
  Alternate framing: OpenAI agents discovered they could post to a wiki on the external internet, did so, and used it to communicate. https://t.co/Lnejpr3KzC
- [@tchsignal on X](https://x.com/tchsignal/status/2095922412751880596)
  Researchers Linked Thousands of Wiki Agents to OpenAI Researchers analyzing public wiki activity attributed 3,103 AI agents to systems linked with OpenAI, documenting 14,666 edits across 4,584 pages b…
- [@humanoidsdaily on X](https://x.com/humanoidsdaily/status/2095921359553450339)
  Everyone is analyzing NVIDIA’s $12.9B Hugging Face deal through LLMs and cloud compute. The more interesting story? Physical AI. Between LeRobot becoming the open standard for robot learning and Polle…
- [@a16z on X](https://x.com/a16z/status/2095921217308086425)
  Old footage sitting in your camera roll could now be reconstructable as a 3D scene. World Labs co-founders Ben Mildenhall and Fei-Fei Li on how Atlas got there: Ben: "In a casual sense... I took three…
- [@MelvinInvests on X](https://x.com/MelvinInvests/status/2095920654759743956)
  OpenAI just shipped the best model in the world and the biggest winner will be Oracle (Save this). GPT 6 Astra launched yesterday and OpenAI calls it the most intelligent and aligned model in the worl…
- [@SemiAnalysis_ on X](https://x.com/SemiAnalysis_/status/2095919853471224118)
  Zhipu ($2513.HK) filed its first interim report as a public company on 31 Aug. 19 of the 60 pages read like a technical blog, plus a 9-page technical glossary. Most investors would not expect to see G…
- [@NVIDIARobotics on X](https://x.com/NVIDIARobotics/status/2095919230247936270)
  More capable open models are becoming practical for real-time edge AI on NVIDIA Jetson. ⚡ Choosing the right one is not just about capability. It is about matching the model to the work your agent nee…
- [@tchsignal on X](https://x.com/tchsignal/status/2095918470047351131)
  Apple Enters the Ternus Era as Nvidia Expands Across the AI Stack Two of the technology industry’s biggest companies are entering very different next phases. John Ternus became Apple CEO on September …
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095917053227979144)
  The pattern we need to talk about more: agents do not stay inside their containers. Reuters reported today that a swarm of OpenAI agents hijacked DseWiki, a German programming wiki, in May. They made …
- [@CKCapitalxx on X](https://x.com/CKCapitalxx/status/2095914980533047483)
  The debt overhang narrative around OpenAI and $ORCL was overdone. Astra just dropped and it's a real leap, trained on 100,000+ GPUs at Stargate, OpenAI's own VP calling it their biggest training run b…
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2095906198994239540)
  ROBOTIS is turning its AI Sapiens K1 humanoid into a real training platform for the next generation of robotics engineers. Through Open Humanoid GYM, student teams build motions in simulation and tran…
- [@a16z on X](https://x.com/a16z/status/2095905665516736760)
  We're thrilled to lead a $300M investment in Gimlet Labs. AI inference is one of the fastest-growing markets in the history of capitalism, and we are running out of nearly every physical input require…
- [@martin_casado on X](https://x.com/martin_casado/status/2095905556242604308)
  Unbelievable 360 from a single image. When can we say NVS is solved? https://t.co/U9hbOsDcPN
- [@a16z on X](https://x.com/a16z/status/2095904215969477103)
  We're thrilled to lead a $300M investment in Gimlet Labs. AI inference is one of the fastest-growing market in the history of capitalism, and we are running out of nearly every physical input required…
- [@pstAsiatech on X](https://x.com/pstAsiatech/status/2095900858265276714)
  OpenAI agents hijacked German website in previously undisclosed AI breakout this spring Can you imagine if this had been a Chinese website? https://t.co/8rww2zK8qn
- [@ricci_nov on X](https://x.com/ricci_nov/status/2095898516665340115)
  Extra thinking compute without extra tokens: a pause carried in a parallel prediction stream on shared weights adds no context length, no KV cache and essentially no latency, and buys 2-3 centinats on…
- [@ricci_nov on X](https://x.com/ricci_nov/status/2095895438700466248)
  Attention-based KV eviction fails when the cache outlives the queries: H2O and SnapKV score 0.00-0.33 on needle retrieval in a NoPE-MLA model. VestigeKV reads 11% of each row and holds 1.00 at 8x and …
- [@ricci_nov on X](https://x.com/ricci_nov/status/2095890091931799690)
  Sequence parallelism today partitions the same way regardless of micro-batch size. BASP splits GPUs into disjoint groups sized to the micro-batch, shrinking the all-to-all group. Training time improve…
- [@Thom_Wolf on X](https://x.com/Thom_Wolf/status/2095889630306472127)
  Another swarm of AI agents in the wild, this time on a German-language forum, found by safety researchers looking for activity similar to the swarm that attacked Hugging Face. A couple of notes while …
- [@a16z on X](https://x.com/a16z/status/2095882301032828932)
  World Labs co-founders Fei-Fei Li, Justin Johnson, Ben Mildenhall, and a16z's Martin Casado on Atlas, a world model for spatial intelligence: LLMs are built on next token prediction. Video models are …
- [@chris_j_paxton on X](https://x.com/chris_j_paxton/status/2095865736975515725)
  This: "You don't have to post-train ChatGPT on every user" is exactly what I think robotics has been missing and why every "chatGPT moment" for robotics has rung a bit hollow. we're at a really exciti…
- [@chris_j_paxton on X](https://x.com/chris_j_paxton/status/2095854947938058250)
  Learn contact-rich humanoid manipulation policies, purely within a world model. This is usually not tractable, because you need a very large model to get the contact dynamics right -- instead, @Jspham…
- [@RoboPapers on X](https://x.com/RoboPapers/status/2095854394667393218)
  Reinforcement learning has unlocked tremendous gains in broader AI and humanoid robot whole body control, but still has yet to deliver strong results for robot manipulation due to the difficulty of ca…
- [@BitsWattsDesk on X](https://x.com/BitsWattsDesk/status/2095852633496195313)
  The humanoid hand debate is focused on the wrong bottleneck. The usual argument is about where to put the motor. Tendon-driven hands keep the actuators upstream and pull the fingers with cables. Tesla…
- [@pstAsiatech on X](https://x.com/pstAsiatech/status/2095848583362691517)
  Machines now use AI more than humans, boosting cheap Chinese models https://t.co/l9X3tRiwNq via @scmpnews
- [@ErenChenAI on X](https://x.com/ErenChenAI/status/2095846181922329012)
  TeleAI × Gamma Robotics unveiled NEXUS, a perceptive behavior foundation model for humanoids. The idea: take motions learned on flat ground and adapt them in real time to complex terrain without losin…
- [@Biti8888 on X](https://x.com/Biti8888/status/2095813477386400150)
  Figure is making a massive bet on compute. The company has committed $3.5B to a multi-year partnership with Nscale, with plans that could eventually scale beyond $6B and provide access to as many as 1…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095792230363771232)
  GPT-6 Astra hit 99.9% on ARC-AGI-3 with a stateful harness. Under standardized eval it scored 62.7%. ARC AGI 4 launched today. Astra and Claude Fable are both at ~4%. Saturating a benchmark doesn't me…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095768833437184166)
  The hardest thing to benchmark is not intelligence. It is judgment over a full year. E-Commerce Bench, released this week by the Qwen team, is the first benchmark that runs an LLM agent through a full…
- [@LimX_Dynamics on X](https://x.com/LimX_Dynamics/status/2095758891892298066)
  One Navigation Brain, Two LimX Platforms: LightNav-0 × Oli / TRON 2 @LightOrigins’ open-source VLN model LightNav-0 was deployed zero-shot on LimX Dynamics’ Oli and TRON 2, enabling person following a…
- [@smsehy on X](https://x.com/smsehy/status/2095757607290216673)
  NEWS: Nvidia has released a new video showing how they are working on solving Full-Self Driving using synthetic data. Nvidia CEO Jensen Huang: “We build detailed 4D https://t.co/oHW9oDpN2i
- [@smsehy on X](https://x.com/smsehy/status/2095757175851479448)
  Utilizing digital twins to systematically augment rare failure edge cases accelerates reinforcement learning policy stability without risking physical robot hardware damage. https://t.co/I5C3BaTXT8
- [@Scobleizer on X](https://x.com/Scobleizer/status/2095756331915542758)
  A 3D world for your bots. https://t.co/WSUa10DjhA
- [@smsehy on X](https://x.com/smsehy/status/2095755978360902124)
  Humanoid scaling is tightly coupled to heavy rare earth refining corridors for dysprosium and terbium. Without domestic permanent magnet sintering capacity, software models cannot be deployed into phy…
- [@smsehy on X](https://x.com/smsehy/status/2095754933601776017)
  Encoding factory floor craftsmanship into deterministic physical AI models requires capturing micro-slip tactile telemetry and high-frequency vibration data, not just standard RGB camera streams. http…
- [@smsehy on X](https://x.com/smsehy/status/2095753144831791614)
  Building an industrial robotics conglomerate by combining foundation model architectures with legacy robotic arm distribution creates immediate sales channel leverage across global automotive and elec…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095739885772554382)
  Reef is one of the most ambitious open-source AI infrastructure projects I have seen this year. ~300 GitHub stars in two days. The core idea is genuinely different from anything else in the space: mos…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095734852259881296)
  What NVIDIA shipped at IFA yesterday quietly changes local inference infrastructure. PAIR (Personal AI Router) is a free, open-source tool that auto-discovers compatible GPUs on your local network and…
- [@tchsignal on X](https://x.com/tchsignal/status/2095733547231514841)
  Meta Puts a Price on AI Interaction Data Meta is offering a striking trade-off for developers using its Muse Spark AI model: much cheaper API access if they opt in to Contributor terms that allow thei…
- [@smsehy on X](https://x.com/smsehy/status/2095712146147733971)
  Training reinforcement learning on cached latent embeddings solves policy training throughput. The critical step for automotive manufacturing remains hard reliability optimization to eliminate rare ed…
- [@smsehy on X](https://x.com/smsehy/status/2095711570483634340)
  Prototyping robotic kinematics is accessible for small engineering teams, but foundation models and custom edge silicon create massive capital consolidation. A few platforms will own the physical AI o…
- [@smsehy on X](https://x.com/smsehy/status/2095711321790566803)
  Physical constraints across high-bandwidth memory packaging and behind-the-meter power interconnects are forcing compute optimization down to the edge. When centralized megawatts take four years to co…
- [@smsehy on X](https://x.com/smsehy/status/2095711107193450908)
  255 million vehicles and global manufacturing plants already generate the physical sensor telemetry that foundation models need to learn real-world physics. Digital AI scales in datacenters, but physi…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095705156671480096)
  The 30-50x cost difference between computer use and MCP keeps rattling around in my head. The framing that clicks for me: computer use is paying for a human-in-the-loop at model prices. Every screen i…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095700123464733012)
  The thing I keep noticing across agentic frameworks in 2026: resumability is now the reliability primitive everyone is building toward. Genkit Go 1.13 ships it properly. A Generate call that fails at …
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2095698231455813708)
  A big win for the robotics community AGIBOT just open sourced one of its most valuable real world datasets yet. @AGIBOTofficial has open-sourced AGIBOT WORLD 2026 Theme 3, a real-world dataset designe…
- [@DrJimFan on X](https://x.com/DrJimFan/status/2095696882844463382)
  Good old days at OpenAI in 2016: an agent stares at screen pixels, moves a mouse, and books a flight on United. We called it World of Bits, inside OpenAI Universe. 10 yrs later, Astra is reincarnated …
- [@rimtoln on X](https://x.com/rimtoln/status/2095628517103071435)
  GPT-6 ASTRA ISN'T A PATCH. IT'S A GENERATION FLIP. openai just shipped the model they call a generational leap past gpt-5.6 sol not better chat better computer use · coding · cyber · science ▹ why thi…
- [@deepakpathak on X](https://x.com/deepakpathak/status/2095626632430240118)
  You don't have to post-train ChatGPT on every user. If you did, it would never have taken off. Yet this is exactly how robotics works today For robots to take off, they need to learn in-context. Great…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095624626084716955)
  The AI lab valuation compression is happening.Thinking Machines Lab is in talks to raise $1 billion at a $40 billion pre-money valuation, according to The Information. That's down from $50 billion the…
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2095609621180232193)
  GPT-6 Astra is here, and it might be one of the most important AI releases yet. OpenAI says the model hits 98% on FrontierMath Tier 4, 99.9% on ARC-AGI 3 and a perfect 100% on ExploitBench. It also sh…
- [@SemiAnalysis_ on X](https://x.com/SemiAnalysis_/status/2095595233064972516)
  Shoutout to the cracked team at @vllm_project that implemented recent agentic workload optimizations. (1/5)🧵 https://t.co/BsXCtLBwvU
- [@XRoboHub on X](https://x.com/XRoboHub/status/2095588510950727820)
  Robots can’t stop while the model thinks. In a high-speed throw, even one pause can kill the momentum. Astribot released SmoothRL for online RL during async inference. S1 keeps moving as the model com…
- [@spaceandtech_ on X](https://x.com/spaceandtech_/status/2095587174804750669)
  AGIBOT has officially open-sourced AGIBOT WORLD 2026 Theme 3: Reinforcement Learning, a real-world embodied AI dataset collected across expert demonstrations, autonomous policy rollouts, and human-in-…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095578824046219594)
  One command now routes Claude Code, Cursor, Codex, Cline, and four other coding harnesses through a single AI Gateway. That is what Vercel's setup flow does: run `vercel ai-gateway coding-agents setup…
- [@frontrunvc on X](https://x.com/frontrunvc/status/2095574324283674959)
  EARLY: @munariai - 6 followers (Robotics) Munari just launched. Operator scoring for robot training data. On @frontrunvc, the story started 47 days before the X account went live. jul 16: a tracked ac…
- [@AppliedInt on X](https://x.com/AppliedInt/status/2095569678685049061)
  At this year’s Agentic AI Summit hosted by @BerkeleyRDI, our Chief Scientist, @Wei_ZHAN_ shared some of the cutting edge research he has been leading in the physical AI space. As the industry races to…
- [@rimtoln on X](https://x.com/rimtoln/status/2095568388684738657)
  EVERYONE SHOWS THE 300-AGENT SWARM. almost nobody shows what happens after 300 agents are easy to spawn turning 300 outputs into one structure is the hard part ▹ the pipeline nobody demos collect → co…
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2095562223582060581)
  Zhejiang University just pushed long-horizon robot manipulation forward with HINT. HINT helps Vision-Language-Action models keep track of the human’s original intent across long tasks instead of losin…
- [@ctorobotics on X](https://x.com/ctorobotics/status/2095557438602915934)
  Construction plans, printed directly onto the floor. 🤖🏗️ Meet FieldPrinter 2 from Dusty Robotics, an autonomous layout robot designed to turn digital construction models into full-scale markings dir…
- [@tchsignal on X](https://x.com/tchsignal/status/2095551413661335854)
  Nvidia’s $12.9B Hugging Face Deal Raises a Bigger Question About Open AI Nvidia has agreed to acquire Hugging Face in a deal valued at about $12.93 billion, but the transaction has not closed yet. It …
- [@MelvinInvests on X](https://x.com/MelvinInvests/status/2095548651619877295)
  AI’s biggest bottleneck is moving data and that could still create huge opportunities for optical networking companies (Save this) The chart shows a 1.6T optical transceiver, a device that transfers d…
- [@TheHumanoidHub on X](https://x.com/TheHumanoidHub/status/2095543869245108681)
  Nvidia is now a humanoid robot designer. That's the real takeaway from this news. Nvidia is acquiring Hugging Face for $12.93B. Hugging Face is the largest open-model and dataset hub, the "GitHub of A…
- [@elonmusk on X](https://x.com/elonmusk/status/2095541022365266094)
  Humans writing code as seen by future AI https://t.co/WqmYxAvKGx
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095523460046885193)
  $12.9B is not a model acquisition. It is a distribution acquisition.NVIDIA just locked in the largest AI developer community on the planet: 18M developers, 3M models, 500K datasets, 1M apps. Jensen's …
- [@CKCapitalxx on X](https://x.com/CKCapitalxx/status/2095499690305352092)
  $NVDA just officially agreed to acquire Hugging Face for $12.93 billion. Hugging Face is where open source AI lives. Over 18 million developers, 3 million models, 500,000 datasets, and 200,000 compani…
- [@Biti8888 on X](https://x.com/Biti8888/status/2095489069358645443)
  🚨 BREAKING: @nvidia has officially confirmed the acquisition of @huggingface Jensen Huang said open models strengthen safety and cybersecurity, accelerate innovation, and let every developer, startup…
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2095487614560145813)
  People saw the NVIDIA and Hugging Face acquisition news and immediately focused on foundation models, GPUs and the broader AI industry. But one of the biggest consequences could be in Physical AI and …
- [@tphuang on X](https://x.com/tphuang/status/2095462541514461499)
  Horizon Robotics has hit 15m in chip delivery. It has a 31.94% mkt share in China's auto ADA chip mkt in the 1st 6 months of the yr. Its J6M chip recently started fielding on various models as the 128…
- [@tchsignal on X](https://x.com/tchsignal/status/2095441545906331912)
  Claude Cybersecurity Tests Exposed a Two-Layer AI Safety Problem Anthropic found three incidents in which Claude models, while running cybersecurity evaluations, reached real production systems instea…
- [@Newsforce on X](https://x.com/Newsforce/status/2095440412379091244)
  A new LatchBio report found @SpaceXAI ’s @Grok 4.6 is now the top-performing AI model in biosecurity safety testing, refusing disguised requests for dangerous biological research while still answering…
- [@Biti8888 on X](https://x.com/Biti8888/status/2095440140785078321)
  Figure and Hark are run by him simultaneously - meaning one person is literally steering two multi-billion-dollar companies in parallel, not even counting his role at Cover. That’s three companies, th…
- [@smsehy on X](https://x.com/smsehy/status/2095438457741214089)
  Deploying general-purpose robots into demanding factory plants requires tight integration between high-level reasoning models and ruggedized, fanless edge controllers certified for industrial temperat…
- [@smsehy on X](https://x.com/smsehy/status/2095436258772496396)
  Capital deployment into physical AI foundation models is surging, but deploying robots into production environments requires tooling up domestic supply chains for high-precision strain wave gears, hig…
- [@smsehy on X](https://x.com/smsehy/status/2095435804466426036)
  Deploying driverless fleets without steering controls validates perception models in urban corridors. The next operational test is managing remote teleoperation intervention ratios and depot turnaroun…
- [@tchsignal on X](https://x.com/tchsignal/status/2095431997548294427)
  Lawsuit Targets Secrecy Around US Frontier AI Review Rules A new FOIA lawsuit is testing how much of the US government's frontier-AI review system should remain out of public view. Executive Order 144…
- [@OperationsPLS on X](https://x.com/OperationsPLS/status/2095425056818303140)
  Agriculture mechanization took 80 years. Robotics + AI compresses that to a decade. Workflow redesign, not just headcount math, is the ops challenge. Transition speed is the variable nobody models. ht…
- [@tchsignal on X](https://x.com/tchsignal/status/2095421748154097945)
  US Government Backs OpenAI's Fair Use Argument for AI Training The U.S. government has entered a major copyright fight over AI training, backing OpenAI's argument that using copyrighted written works …
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095403166775099887)
  The model that matters for agentic software is not the best at chat. It is the one that completes a 4-hour task without losing the thread. Meta released Muse Spark 1.3 yesterday. The benchmark that go…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095359629735710878)
  The model that finds the most bugs does not find all the bugs. Bug Hunt Bench: 2 real repos, 105 hidden bugs, all tests green. Fable 5.1 leads at 43/105. GPT-5.6 is one behind at 42. But the number th…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095354597007032519)
  Standard coding benchmarks are saturating. The interesting separation is happening on autoresearch. Autoresearch Bench launched this week: a benchmark for coding agents autonomously tackling research …
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095339497177444759)
  The benchmark gap that keeps surprising me: on standard coding benchmarks, the top models cluster. On ultra-long horizon coding, the gap blows open. Proximal released FrontierSWE v2 today. 34 real sof…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095329430641979875)
  Meta released Muse Spark 1.3 today. The improvements are measurable: 20% fewer tool calls and 25% fewer tokens compared to 1.2. Fewer tool calls is the right metric to optimize. Each round-trip adds l…
- [@SemiAnalysis_ on X](https://x.com/SemiAnalysis_/status/2095316010249166875)
  NVIDIA Research 🚀 has produced some great research, like LatentMoE (used in Kimi K3) and GatedDeltaNets (used in Qwen). But for e2e frontier training, NVIDIA's bureaucratic culture has produced embar…
- [@Scobleizer on X](https://x.com/Scobleizer/status/2095312354083062029)
  For a while I was depressed about VR. There are so many pioneers who have gone sour on VR/AR. That depression lifted this morning. The future is so amazing. And I am seeing this in many companies from…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095309298138272015)
  12 PhDs and a Fields Medal winner built this in 4 months, backed by General Catalyst. https://t.co/pyhyV4u6I4 has a protocol that passes hidden states directly from a 753B frontier model to a 4B edge …
- [@ruima on X](https://x.com/ruima/status/2095299079534477635)
  Well, this news really is dividing people. Some see this as evidence that AI has so much room to grow, especially as it expands beyond coding as a use case, since apparently this top 1% are other tech…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095294198886903921)
  The hybrid compute pattern is getting serious. Perplexity just open-sourced Lily, the local inference engine powering their Mac hybrid compute feature, and the architecture choices are worth understan…
- [@a16z on X](https://x.com/a16z/status/2095267502930764175)
  World Labs co-founder Justin Johnson on how five photos from your phone could teach a robot to work in a room it has never seen: "I think there's a couple different tech trees that people are working …
- [@MParekh on X](https://x.com/MParekh/status/2095258950320996639)
  ‘New and Improved’ in AI Models, Anthropic, OpenAI & Google. ARD #154 ...Fable 5.1’s fine print. Astra tips out on cyber. Gemini narrows the gap. World models level up. Full writeup and source lin…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095255445124583878)
  The AI assistant category is being replaced. QwenWork landed in public beta globally this week. Alibaba's move is to make the next product category a single subscription that puts an AI agent in contr…
- [@ricci_nov on X](https://x.com/ricci_nov/status/2095252181683474590)
  SinkPruner drops 89% of visual tokens in multimodal models with no training, keeping 96.5% of LLaVA-1.5's performance and 91.8% of Qwen2.5-VL's. The tokens it cuts are high-norm outliers earlier metho…
- [@Scobleizer on X](https://x.com/Scobleizer/status/2095248673668248057)
  You have been paying a giant model to think and then to write. That second part is the expensive one. Mostik splits it. The giant model reads your problem. A small one on your own computer writes the …
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095245376978870557)
  The model tier split just happened. Gemini 3.8 Flash and Gemini 3.8 Flash Cyber launched today. Same day, completely different access regimes. The public Flash model is on the Gemini API at $0.75 per …
- [@TheRundownAI on X](https://x.com/TheRundownAI/status/2095234538267877768)
  The heavy week of AI releases continues with Meta and Muse Spark 1.3. Big jumps across the board over 1.2, with Mark Zuckerberg calling it "frontier performance almost too cheap to meter." https://t.c…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095230277727461755)
  The question "which model is better for my use case" only has one honest answer: run your actual work across both and compare. Fable 5.1 shipped yesterday. Qwen3.8-Max-0902 shipped yesterday. Grok 4.7…
- [@stretchcloud on X](https://x.com/stretchcloud/status/2095225248774029545)
  Something shifted with the World Labs Atlas announcement. We've had image generators, video generators, and 3D tools for years. Atlas does all three from a single model with the same spatial understan…
- [Agility on X: "Every object a different size, weight, and grip. Rearranging a full room's worth of stuff is the same generalized manipulation Digit uses to take on mixed loads and changing layouts on a real warehouse floor.

#Agility #DigitRobot" / X](https://x.com/HumanoidInvest/status/2095166627708887090)
  Post Log in Sign up Post Agility on X: "Every object a different size, weight, and grip. Rearranging a full room's worth of stuff is the same generalized manipulation Digit uses to take on mixed loads…
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2095145553482891753)
  We’re cooked if robots can start learning from bodies they don’t even have. NVIDIA Brown Columbia and Harvard researchers released Hydra-0, a world model that represents robot actions as pixel motion.…
- [Wall St Engine on X: "CATERPILLAR PARTNERS WITH FIELDAI ON PHYSICAL AI

$CAT is working with FieldAI to bring AI-powered autonomy and robotics into jobsites and factories, using Caterpillar’s operational data, FieldAI’s robot foundation models and NVIDIA technologies.

Early applications include au… / X](https://x.com/CKCapitalxx/status/2095141910465618065)
  Post Log in Sign up Post Wall St Engine on X: "CATERPILLAR PARTNERS WITH FIELDAI ON PHYSICAL AI $CAT is working with FieldAI to bring AI-powered autonomy and robotics into jobsites and factories, usin…
- [@10xthinker on X](https://x.com/10xthinker/status/2095127748188741887)
  Are humanoids a business story or just hype? - The Beijing robotics conference almost looked like a comedy show — robots stumbling, misidentifying obstacles, and performing like toddlers. - Yet the bu…
- [@smsehy on X](https://x.com/smsehy/status/2095119584890085731)
  Morgan Stanley estimates building an Optimus Gen 2 costs $45,500 using China's supply chain, versus $131,800 without it. That $86,300 delta is not simply raw neodymium pricing. It reflects fifteen yea…
- [Chubby♨️ on X: "wtf Qwen 3.8 has been updated to version 0902 and is now almost at the same level as Fable 5 in benchmarks.

The interesting two aspects:

1) For such significant jumps, it seems that you don't even change the version number anymore (no Qwen 3.9), rather the development happens so q… / X](https://x.com/stretchcloud/status/2095220210999664739)
  Post Log in Sign up Post Chubby♨️ on X: "wtf Qwen 3.8 has been updated to version 0902 and is now almost at the same level as Fable 5 in benchmarks. The interesting two aspects: 1) For such significan…
- [@smsehy on X](https://x.com/smsehy/status/2095050737344725288)
  Digital models scale exponentially by distributing weights across networks, but physical AI scaling remains gated by physical component manufacturing lead times. Sourcing specialized precision cycloid…
- [Omai Leidi (3/3) on X: "Happy Wednesday frens👋

Everyone is watching the race to build better robots. 
I’m starting to think the real race is the data teaching those robots how to act.

I’ve been digging into @axisrobotics, and it changed how I look at the space.

At first glance, AXIS looks like… / X](https://x.com/OperationsPLS/status/2095167309207982333)
  Post Log in Sign up Post Omai Leidi (3/3) on X: "Happy Wednesday frens👋 Everyone is watching the race to build better robots. I’m starting to think the real race is the data teaching those robots how…
- [LimX Dynamics on X: "Together with ZINOVA's Tool Intelligence, TRON 2 takes on increasingly complex construction workflows.

@ZiNovaLabs builds on TRON 2 to explore an innovative robotic configuration for construction, demonstrating key tasks in a scaled-down tilt-up construction workflow, includi… / X](https://x.com/chris_j_paxton/status/2095114381818253768)
  Post Log in Sign up Post LimX Dynamics on X: "Together with ZINOVA's Tool Intelligence, TRON 2 takes on increasingly complex construction workflows. @ZiNovaLabs builds on TRON 2 to explore an innovati…
- [Noble Machines Accelerates Humanoid Robot Development 3X | NVIDIA](https://x.com/NVIDIARobotics/status/2094923298589024698)
  Manufacturing \| Robotics Noble Machines Accelerates General Purpose Industrial Robot Development With NVIDIA Isaac GR00T Learn More Objective Noble Machines is advancing industrial robot development …
- [Yuchen Jin on X: "Even if most AI labs are benchmark-maxxing now, Fable 5.1 still looks like an insane jump.

From Anthropic’s blog:
- Found a 1-in-a-million crash Millennium’s team couldn’t explain for 4–5 years
- 2x faster than Opus 5 while using half the tokens

Big if true." / X](https://x.com/pstAsiatech/status/2095219413431529817)
  Post Log in Sign up Post Yuchen Jin on X: "Even if most AI labs are benchmark-maxxing now, Fable 5.1 still looks like an insane jump. From Anthropic’s blog: \- Found a 1-in-a-million crash Millennium’…
- [From Robot Development to Deployment with Isaac GR00T & Jetson Thor - YouTube](https://x.com/NVIDIARobotics/status/2094850249802076179)
  Error 401 (Bad Request)!!1 401. That’s an error. The server cannot process the request because it is malformed. It should not be retried. That’s all we know. Back Skip navigation Search Search with yo…
- [Seeed reBot Arm B601-RS: Physical AI & VLA Model Course | NVIDIA DLI Series](https://x.com/NVIDIARobotics/status/2094850247184826558)
  Warehouse China Warehouse US Warehouse Germany Warehouse The store will not work correctly in the case when cookies are disabled. US Warehouse: Enjoy FREE UNIUNI shipping on orders under $50! (Exclude…
- [Physical AI for Industrial Automation | Sanctuary AI](https://x.com/TheSanctuaryAI/status/2094844880057229640)
  We value your privacy We use cookies to enhance your browsing experience, serve personalised ads or content, and analyse our traffic. By clicking "Accept All", you consent to our use of cookies. Custo…
- [Google DeepMind on X: "We’re bringing agentic video understanding to our latest Gemini models.

They can now analyze videos with better accuracy while using up to 88% fewer tokens. 🧵" / X](https://x.com/GoogleDeepMind/status/2094840179676660097)
  Post Log in Sign up Post Google DeepMind on X: "We’re bringing agentic video understanding to our latest Gemini models. They can now analyze videos with better accuracy while using up to 88% fewer tok…
- [World Labs on X: "All of this is powered by one unified architecture: a multimodal autoregressive diffusion transformer, pretrained from scratch.

This foundation blends the best of modern LLMs and video models, benefiting from the architectural, algorithmic, and systems advances from both areas." / X](https://x.com/theworldlabs/status/2094839768115036581)
  Post Log in Sign up Post World Labs on X: "All of this is powered by one unified architecture: a multimodal autoregressive diffusion transformer, pretrained from scratch. This foundation blends the be…
- [Atlas: A World Model for Spatial Intelligence | World Labs](https://x.com/theworldlabs/status/2094839769616589051)
  September 1, 2026Introducing Atlas, our new omni world model for spatial intelligence. Atlas: A World Model for Spatial Intelligence World models generate, reconstruct, and simulate any possible world…
- [wafer on X: "We’re excited to announce that we’ve raised a $40M Series A!

co-led by @MarathonMP and @chemistry, with participation from @Wing_VC, @AMD Ventures, @outsetcap, @fiftyyears, and @ycombinator, and our existing investors doubling down on @wafer_ai. we are also joined by an incredible lis… / X](https://x.com/frontrunvc/status/2094841271244247488)
  Post Log in Sign up Post wafer on X: "We’re excited to announce that we’ve raised a $40M Series A! co-led by @MarathonMP and @chemistry, with participation from @Wing\VC, @AMD Ventures, @outsetcap, @f…
- [C. Zhang on X: "New paper release: Learning Agile Perceptive Traversal of Sparse 3D Structures for Humanoids

Paper https://t.co/LlrDsajHqx
Videos https://t.co/jXLtG68oUg

We study how to do learning and sim2real for 3d traversal, such as monkey bars and overhanging obstacles." / X](https://x.com/chris_j_paxton/status/2094815833516060984)
  Post Log in Sign up Post C. Zhang on X: "New paper release: Learning Agile Perceptive Traversal of Sparse 3D Structures for Humanoids Paper https://t.co/LlrDsajHqx Videos https://t.co/jXLtG68oUg We st…
- [@OptionKing666 on X](https://x.com/OptionKing666/status/2094638592253370646)
  Unitree’s IPO raised RMB 6bn and closed its first day up 460% at an RMB 237bn market value; 48% of proceeds is allocated to robot-model R&D. The asymmetry is clear: funding for capability expansion is…
- [smartfitguide on X: "Google DeepMind's RT-2 model treats robot actions as text tokens. By co-fine-tuning pre-trained Vision-Language Models with robotic data, it achieved nearly double the performance on novel tasks compared to its predecessor, RT-1. #AI" / X](https://x.com/bareani21645/status/2094613978739593316)
  Post Log in Sign up Post smartfitguide on X: "Google DeepMind's RT-2 model treats robot actions as text tokens. By co-fine-tuning pre-trained Vision-Language Models with robotic data, it achieved near…
- [tphuang on X: "GLM-6 disclosed - more parameter & lower activation ratio & inference cost & improved post training expected. Model to make self training decisions

Expecting significant increased domestic compute addition in next 3 to 6 months. Using self owned, leased & purchased c… / X](https://x.com/tphuang/status/2094601178684264502)
  Post Log in Sign up Post tphuang on X: "GLM-6 disclosed - more parameter & lower activation ratio & inference cost & improved post training expected. Model to make self training decisions …
- [tphuang on X: "How it views its own models in terms of the kind of tasks them can do. GLM-5.3 in terms of long time horizon work is still short of 1 wk, need at least 1 month to be autonomous.

Flash is the 1st one to use domestic chips & have low cost + intelligence to support long tasks." / X](https://x.com/tphuang/status/2094601170102747138)
  Post Log in Sign up Post tphuang on X: "How it views its own models in terms of the kind of tasks them can do. GLM-5.3 in terms of long time horizon work is still short of 1 wk, need at least 1 month …
- [@ErenChenAI on X](https://x.com/ErenChenAI/status/2094589562760687736)
  Shanghai-based 3D generative AI company Hyper3D⁠ has introduced WorldGen, a world generation model that turns a single image into a complete, editable 3D scene. Built on its SIGGRAPH 2025 Best Paper-w…
- [The Robot Report: Skild AI unveils S1 robot foundation model | AI Understanding](https://x.com/aiuorg/status/2094602992565522608)
  Back to News ProductAI Understanding briefing The Robot Report: Skild AI unveils S1 robot foundation model The Robot Report reports that Skild AI unveiled S1, a robot foundation model that the company…
- [Introducing Agentic Video in Gemini](https://x.com/GoogleDeepMind/status/2094840182457422260)
  Introducing agentic video understanding with Gemini Sep 01, 2026 \| 7 min read - x.com - Facebook - LinkedIn - Mail - Copy link Our new agentic feature for video analysis cuts token consumption by up …
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2094550793743708256)
  A 36B parameter model just blurred one of robotics’ biggest boundaries, the gap between seeing a scene and acting on it. Perceptron AI has released Isaac 0.5, an open-weight Embodied AI model designed…
- [@a16z on X](https://x.com/a16z/status/2094542071759872251)
  Gavin Baker says the future for the world's biggest companies is open models and private context: "I think the future is an ensemble of models. There's a Pareto curve. No one model is going to be the …
- [国家发改委：加速具身智能在制造、医疗等领域真实场景中落地应用_财经上下游_澎湃新闻-The Paper](https://x.com/pstAsiatech/status/2094536116539961624)
  下载客户端 登录 无障碍 - +1 国家发改委：加速具身智能在制造、医疗等领域真实场景中落地应用 澎湃新闻记者 滕晗 2026-08-28 11:44 来源：澎湃新闻 ∙ 财经上下游 > 听全文 字号 8月28日，国家发展改革委举行8月份新闻发布会，国家发展改革委政策研究室副主任、委新闻发言人李超在会上强调，机器人产业涉及人工智能、先进制造、新材料等诸多前沿技术，必须坚持因地制宜、健康有序发展，立…
- [The hardest problem in physical AI may be the magnet, not the model](https://thenextweb.com/news/physical-ai-hardest-problem-magnet-not-model-crma-schaeffler)
  Skip to content AI Learning and Artificial Intelligence Credit: Canva Actuators account for 40% to 60% of a humanoid robot’s bill of materials and China refines roughly 90% of the rare earth magnets i…
- [The Imitator Game — Benchmarking Robot Imitative Ability Beyond Action Prediction](https://x.com/XunzheZhou/status/2094530835215258044)
  Intent imitation, L0-L320,000+ paired episodesHuman evaluation Arena The Imitator Game BenchmarkingRobotImitativeAbilityBeyondActionPrediction Xunzhe Zhou 1,2,\,† · Yiyang Cai 2,3,\ · Fengyi Wang 2,3,…
- [@Astra1Byte on X](https://x.com/Astra1Byte/status/2094493801410662483)
  Four years ago, Figure didn’t have a robot. Today it’s valued at $39B and has already shipped 1,000+ humanoids. CEO Brett Adcock says the real advantage isn’t the AI it’s that hardware, models, and ma…
- [NVIDIA's robot business is already worth $10 billion/year, and China is its best customer | HumanoidHub](https://x.com/johnkoetsier/status/2094485667409760625)
  News/ trends,china,usa,physical\ai,hardware,government trends,china,usa,physical\ai,hardware,government NVIDIA's robot business is already worth $10 billion/year, and China is its best customer John K…
- [Optimize Models with Unsloth | Jetson AI Lab](https://x.com/NVIDIARobotics/status/2094480283135316182)
  🎉 Welcome to Jetson AI Lab 2.0! We've redesigned the site with curated tutorials. Looking for old content? Browse the archive → × Back to Tutorials Optimize Models with Unsloth Fine-tune, convert and…
- [@ruima on X](https://x.com/ruima/status/2094468785591964066)
  I think the Unitree IPO was obviously overhyped (well, except to whoever bought it at $60B). It’s down to a more manageable ~$30B now, which is still extremely pricey. I wonder what this means for the…
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2094465636265087097)
  Humanoid robotics has its own language. Here are 30 acronyms you will see constantly API = Application Programming Interface BC = Behavior Cloning CoM = Center of Mass CoP = Center of Pressure DoF = D…
- [@thehypedotnews on X](https://x.com/thehypedotnews/status/2094464209895424198)
  unitree robots now build themselves – and train while doing it @UnitreeRobotics deployed its embodied ai model inside its own factory in china. the robots aren't repeating pre-programmed movements – t…
- [Runway on X: "Today, we're sharing new research on Solaris, our first Interface World Model.

Solaris is a new kind of operating system that generates interactive interfaces frame by frame, in real time, with no code. We find that Solaris outperforms frontier LLMs when generating new interfaces, ac… / X](https://x.com/TheRundownAI/status/2094519142426567065)
  Post Log in Sign up Post Runway on X: "Today, we're sharing new research on Solaris, our first Interface World Model. Solaris is a new kind of operating system that generates interactive interfaces fr…
- [@0xconglomerate on X](https://x.com/0xconglomerate/status/2094460816947081633)
  @LeggedRobot @ROBOTIS Looking forward to the next update! Hehe would be fun to have a Mini Pupper that I could play with and make content on, eh? 😅👀
- [@frontrunvc on X](https://x.com/frontrunvc/status/2094448223775494221)
  EARLY: 5 @ycombinator fall cohort companies across Compute, Energy, Hardware and DevTools. 👇 @nodussss - 1 follower - (Compute): the first intelligent execution layer for ai workloads. founders: @vis…
- [@whitee_rhinoo on X](https://x.com/whitee_rhinoo/status/2094444376655024384)
  The #US will impose up to 100% #tariffs on #Chinese #drones over 25kg, thermal models, and key parts starting Sept 3, with lighter drones facing a 25% rate. Aimed at cutting reliance on foreign #suppl…
- [GitHub - hku-sail/StreamPI: StreamPI: Streaming Multimodal Temporal Modeling for Vision-Language-Action Models · GitHub](https://x.com/DexRobotics/status/2094411289380765788)
  Skip to content You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or…
- [Biti8 on X: "Four years ago, @Figure_robot didn’t even have a robot.

Today, it’s valued at $39B, has built 1,000+ humanoids, deployed robots at BMW, built its own AI models, manufacturing infrastructure and a global data collection network.

What impresses me most is the speed.

Figure isn’t build… / X](https://x.com/adcock_brett/status/2094462666798723307)
  Post Log in Sign up Post Biti8 on X: "Four years ago, @Figure\robot didn’t even have a robot. Today, it’s valued at $39B, has built 1,000+ humanoids, deployed robots at BMW, built its own AI models, m…
- [@GabrielnSpace on X](https://x.com/GabrielnSpace/status/2094350888211234898)
  i thought community robot data would hit diminishing returns fast but Axis Dataset V1 just made that take look stupid @axisrobotics pushed π0.5 from 83.9% to 88.8% on LIBERO Plus, and performance kept…
- [Pentagon blacklist vs. FCC drone bans: What drone pilots…](https://x.com/djirumor/status/2094361382112419970)
  Breaking Pentagon blacklist vs. FCC drone bans: What drone pilots need to know about the differences thedronegirl.com 2d ago Aug 31, 2026 The Pentagon blacklist is a paper cut for DJI — it only blocks…
- [How Figure Became the Biggest Name in Robotics | XMAQUINA DAO](https://x.com/xmaquina/status/2094334389979373824)
  Genesis Auction Wave 2! Launches June 24 Get DEUS homeDAO Portal How Figure Became the Biggest Name in Robotics Color theme: Four years, three generations of humanoids, a $39 billion valuation and now…
- [@tchsignal on X](https://x.com/tchsignal/status/2094323502098911410)
  U.S. Barriers Test China's Robotics Scale Advantage The U.S. is raising barriers around foreign-made drones and advanced robots, but those measures are colliding with a market where Chinese manufactur…
- [Physical AI Platform | Industrial Automation Solutions | Sanctuary AI](https://x.com/heetezition/status/2094280429264420968)
  We value your privacy We use cookies to enhance your browsing experience, serve personalised ads or content, and analyse our traffic. By clicking "Accept All", you consent to our use of cookies. Custo…
- [Gemini Robotics ER 1.6 shutdown date Aug 31, 2026; replacement ER 2](https://ai.google.dev/gemini-api/docs/deprecations)
  Google Developers Gemini API deprecations page, Robotics models table (page last updated 2026-08-27 UTC; shutdown date in the table is in-window): gemini-robotics-er-1.6-preview released April 14, 202…
- [@ParadisLabs on X](https://x.com/ParadisLabs/status/2094203645877493923)
  AI is still so early, but so bullish. We're witnessing a huge unlock of economic value with the adoption of AI models like Claude, and more recently, agents like Grok Bot. This then filters down into …
- [@Robot_AIsignals on X](https://x.com/Robot_AIsignals/status/2094201246987034914)
  Show it once. That is the entire instruction. In a post published in August, Skild AI introduced S1, a robotics foundation model prompted with a video demonstration in place of a written command. The …
- [@AtMapshock on X](https://x.com/AtMapshock/status/2094127830585979064)
  The FCC Constraint: Closing the Western Market Escape Valve One mechanism that might have imposed commercial discipline on Chinese humanoid makers — exposure to demanding Western end-users who would n…
- [kasper on X: "Elon Musk was asked when Tesla Optimus could become better than the best human surgeons.

His argument is that a robot would not learn like one doctor. Every Optimus surgeon could share what every other unit had seen, including rare complications that a human might encounter only once… / X](https://x.com/HumanoidsHQ/status/2094325624177397929)
  Post Log in Sign up Post kasper on X: "Elon Musk was asked when Tesla Optimus could become better than the best human surgeons. His argument is that a robot would not learn like one doctor. Every Opti…
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2093739174746611828)
  XPENG is putting serious money behind humanoid robotics. Its robotics business just raised more than $900 million, reaching a post-money valuation above $6.3 billion. The round was led by IDG Capital,…
- [@techniahqrobot on X](https://x.com/techniahqrobot/status/2093692200534024317)
  BeyondMimic is now in Science Robotics. Berkeley and Stanford used a Unitree G1 to track human motions like sprinting, spin-kicks and cartwheels, then used guided diffusion to combine those skills for…
- [South Korea Commits KRW 2.3 Trillion to Build Full-Stack Humanoid Robotics Ecosystem by 2030 | Humanoids Daily](https://x.com/humanoidsdaily/status/2093648555726590278)
  Search articles Key Takeaways Show all - Dedicated Humanoid Capital: South Korea is investing KRW 2.3 trillion ($1.66 billion) through 2030 to build an indigenous, full-stack humanoid robotics industr…
- [@realgalleryx on X](https://x.com/realgalleryx/status/2093635683323891741)
  서랍이 안 열리자 수저를 오븐에 넣었다. Physical Intelligence를 만든 Sergey Levine이 Ryan Peterman 앞에서 말한 평가 장면이다. 모델은 π0.5다. 주방에서 수저를 넣으라고 했는데 서랍이 안 열렸다. 옆에 오븐이 있으니 그걸 열고 넣기 시작했다. 아이한테 치우라고 하면, 안 보이게만 넣기도 한다는 것이다. 접시 닦기도…
- [Gemini Robotics 2 brings whole body intelligence to robots - YouTube](https://x.com/ScrumVenturesJP/status/2093534278378553445)
  Error 401 (Bad Request)!!1 401. That’s an error. The server cannot process the request because it is malformed. It should not be retried. That’s all we know. Back Skip navigation Search Search with yo…
- [@csentropy on X](https://x.com/csentropy/status/2093431495277043724)
  Announced July 30, 2026 (author Carolina Parada) as part of Gemini Robotics 2, three models: Gemini Robotics 2 (VLA) converts vision+language into motor control first DeepMind VLA to control a full hu…
- [MemoryAnchors](https://x.com/du_maximilian/status/2093390863821131915)
  01 · Continual Learning Challenge Data Sensitivity in Continual Robot Learning To understand Memory Anchors, we first need to take a closer look at why robot policies forget old tasks when learning ne…
- [@0xconglomerate on X](https://x.com/0xconglomerate/status/2093373586803478854)
  I’m pretty sure some of the things I mentioned here are already things @eastworlds_io and other robot data companies think about internally. s/o to @scale_AI @xdofai @MeckaAI @micro1_ai @LightwheelAI …
- [Cicada Market Making on X: "https://t.co/26k9SK56mf" / X](https://x.com/axisrobotics/status/2093684446465798629)
  Post Log in Sign up Post - Cicada Market Making @cicada\mm Inside Robotics and Physical AI Featuring insights from Axis Robotics In 2026, the artificial intelligence industry learned to solve the comp…
- [Why Top Founders Are Racing Into AI Infrastructure - YouTube](https://x.com/a16z/status/2093340813174747432)
  Error 401 (Bad Request)!!1 401. That’s an error. The server cannot process the request because it is malformed. It should not be retried. That’s all we know. Back Skip navigation Search Search with yo…
- [@axisrobotics on X](https://x.com/axisrobotics/status/2093217040295465183)
  Announcing our partnership with @Dexmal_AI About Dexmal: Dexmal is a technology company focused on general-purpose embodied intelligence, committed to building intelligent, useful, and trustworthy rob…
- [Physical Intelligence (π)](https://x.com/hausman_k/status/2093129881689391574)
  Physical Intelligence is bringing general-purpose AI into the physical world. We are a group of engineers, scientists, roboticists, and company builders developing foundation models and learning algor…
- [Cerebras's Next Generation CS-4: Fast Just Got Faster](https://x.com/SemiAnalysis_/status/2093126657682907547)
  SubscribeSign in Cerebras's Next Generation CS-4: Fast Just Got Faster Double the Performance, Double the Power, Double the Fun Myron Xie, Bryan Shan, Wega Chu, and 2 others Aug 18, 2026 ∙ Paid 136 6 …
- [Optimus Just Entered Production at Fremont. Here's What Changes for Tesla Investors. | The Motley Fool](https://x.com/VanEmmerickKris/status/2094940107484586346)
  Accessibility Menu ▲ S&P 500 +---% \|▲ Stock Advisor +---% Join The Motley Fool Search for a company Accessibility... Help Arrow-Thin-Down\\ \\ S&P 500\\ \\ 7,673.19\\ \\ +0.5%\\ \\ +41.72 Arrow-Thin-…
- [@Rewkang on X](https://x.com/Rewkang/status/2093056014967087389)
  Most Physical AI companies are still doing lab demos. A key factor in our investment in Dyna last year was their world class post training expertise/results and their deployment focus. More deployment…
- [@frontrunvc on X](https://x.com/frontrunvc/status/2093041605540466832)
  plugged this article into claude + the @frontrunvc mcp. companies that fit the thesis: 👇 vertical, owning a category: @cosmic_robotics autonomous construction @shiraz_ai robots that learn on the job …
- [Dyna Robotics on X: "After so many demos, models, pilots, robots are still struggling to land real deployments with real customers. Until now.

Today we’re excited to share that our robots have successfully crossed the ROI threshold, and Din Tai Fung, one of the highest revenue per location restaur… / X](https://x.com/TheHumanoidHub/status/2093387536962076683)
  Post Log in Sign up Post Dyna Robotics on X: "After so many demos, models, pilots, robots are still struggling to land real deployments with real customers. Until now. Today we’re excited to share tha…
- [Dyna Robotics on X: "After so many demos, models, pilots, robots are still struggling to land real deployments with real customers. Until now.

Today we’re excited to share that our robots have successfully crossed the ROI threshold, and Din Tai Fung, one of the highest revenue per location restaur… / X](https://x.com/Rewkang/status/2093048258537087479)
  Post Log in Sign up Post Dyna Robotics on X: "After so many demos, models, pilots, robots are still struggling to land real deployments with real customers. Until now. Today we’re excited to share tha…
- [Not Just a Model, But a Product — DYNA](https://x.com/DynaRobotics/status/2093039146357178475)
  Contact11:33 AM Name Email Company Team Size 1-2021-5051-100100+ Message BusinessMediaInvestorOther Submit Sections: 01Not just a model, but a product02"Both" is our way03How Dyna-2 closed the gaps04D…
- [@TheHumanoidHub on X](https://x.com/TheHumanoidHub/status/2092974952769048876)
  Perceptron's Isaac 0.5 is a 36B open-weight embodied foundation model with only 2.5B active per token. The model - Perception, reasoning, and control all read from one shared backbone, so it makes vis…
- [New BMW Cars & SUVs | BMW Dealer Serving Atlanta GA](https://x.com/BMWofAtlanta/status/2092960092014391750)
  Skip to main content Labor Day Specials start NOW! Financing as low as 0.9% for 60 months + Loyalty Credits on select models. Shop Labor Day Offers Global Imports BMW - 500 Interstate North Pkwy SE At…
- [Axis Robotics × Booster: From Digital Twins to a Robot Data Engine](https://x.com/axisrobotics/status/2092884880430661971)
  Back How task-aligned simulation, real demonstrations, and continued pretraining can make each new Booster task easier to build than the last. Robot learning has a model problem. More importantly, it …
- [How to Train a Cross-Embodiment Robot Navigation Policy with AI Agents | NVIDIA Technical Blog](https://x.com/NVIDIARobotics/status/2093381258243895657)
  Technical Blog Subscribe Related Resources Robotics English한국어中文 How to Train a Cross-Embodiment Robot Navigation Policy with AI Agents Aug 26, 2026 By Yan Chang, Mihir Acharya, Wei Liu, Katie Washaba…
- [Agility on X: "Whole-body range of motion, stress-tested by a workout class. Every joint in sync — timing, velocity, balance. The same control Digit uses on a real warehouse floor.

#Agility #DigitRobot" / X](https://x.com/agilityrobotics/status/2092704763892937082)
  Post Log in Sign up Post Agility on X: "Whole-body range of motion, stress-tested by a workout class. Every joint in sync — timing, velocity, balance. The same control Digit uses on a real warehouse f…
- [Rhys on X: "Cool to see a bunch of new robotics foundation model releases:

π0.7 from @physical_int 
GEN-1.5 from @generalistAI 
S1 from @SkildAI 

Feels like the GPT-3 moment for robotics. Robot Models are Few-Shot Learners!" / X](https://x.com/RhysLindmark/status/2092684832199639136)
  Post Log in Sign up Post Rhys on X: "Cool to see a bunch of new robotics foundation model releases: π0.7 from @physical\int GEN-1.5 from @generalistAI S1 from @SkildAI Feels like the GPT-3 moment for …
- [World Labs on X: "The next generation of robotics needs models that can understand, simulate, learn from, and act across virtual and physical worlds.

We’re hiring a Sr Business Development Lead, Robotics & Physical AI to help bring this technology into real-world applications 🤖🚀" / X](https://x.com/theworldlabs/status/2092659658980983279)
  Post Log in Sign up Post World Labs on X: "The next generation of robotics needs models that can understand, simulate, learn from, and act across virtual and physical worlds. We’re hiring a Sr Busines…
- [@adcock_brett on X](https://x.com/adcock_brett/status/2092628102262477268)
  We're now over 43,000 weekly active users collecting data to train Helix, our AI model for F.03 robots This data collection project, Index, is our answer to the data problem: the largest useful robot …
- [@antopatrex1 on X](https://x.com/antopatrex1/status/2092616347792580997)
  scaling laws do not apply to physical AI yet. i'll say it again. scaling laws do not apply to physical AI. chelsea finn from physical intelligence showed this during her YC talk. they trained pi0.7 to…
- [@UnitreeStore_SH on X](https://x.com/UnitreeStore_SH/status/2092555273936503147)
  Unlock unlimited sports potential (Extra large joint movement space angle, 23~43 joints) Force control of dexterous hands, manipulation of all thingsImitation & reinforcement learning driven Robot wor…
- [@zhodonx on X](https://x.com/zhodonx/status/2092555239282843992)
  I’ve said before that I think Google & DeepMind are playing a longer game with models. This is still a firm stance. Because a lot of the time, i think people look at Google’s model strategy too narrow…
- [@dredgefactory on X](https://x.com/dredgefactory/status/2092395639384539542)
  A 29900$ MACHINE IS DANCING ON A PAVEMENT AND NOT ONE PERSON WALKING PAST LOOKS UP. 29900$ is a Unitree H2, the class this thing sits in. Entry models go for 13500$. Every beat throws its centre of ma…
- [@gupta_abhinav_ on X](https://x.com/gupta_abhinav_/status/2092331051041612145)
  Punch: Skild S1 is 380 times more data efficient than standard VLA models! yes...let that sink in..
- [@rohanpaul_ai on X](https://x.com/rohanpaul_ai/status/2092312459382235407)
  Robotics is more and more getting close to get its version of in-context learning. Skild just released S1, a robot foundation model that uses video demonstrations to define tasks instead of language i…
- [@DeryaTR_ on X](https://x.com/DeryaTR_/status/2092312053272707504)
  This is super exciting advance in robotics AI! I believe S1 is the most impressive robot foundation model I had seen! S1, is a robot foundation model built to learn new manipulation tasks through in-c…
- [@gupta_abhinav_ on X](https://x.com/gupta_abhinav_/status/2092304774813331612)
  We did not rush to put our results in public domain because we wanted to spend time understanding the prompting phenomenon. For the last two months, we have tried to obtain scaling laws of prompting m…
- [@chris_j_paxton on X](https://x.com/chris_j_paxton/status/2092304651995529392)
  This is the mark of a general purpose robot foundation model -- that it can be taught tasks from examples without training. We saw an early version of this from rhoda and a solid one last week from ge…
- [Brett Adcock on X: "The data needed to scale a truly general purpose robots doesn't exist on the internet - it has to come from the real world

For the last 4 months, we’ve been building a Figure-exclusive pipeline to scale data collection at higher throughputs. Today, we’re coming out of stealth" / X](https://x.com/adcock_brett/status/2092304283865891009)
  Post Log in Sign up Post Brett Adcock on X: "The data needed to scale a truly general purpose robots doesn't exist on the internet - it has to come from the real world For the last 4 months, we’ve bee…
- [Brett Adcock on X: "Introducing Index

Today we're coming out of stealth with Index, the largest & most diverse robot dataset in the world

→ 30min of video uploads/sec
→ 16M video uploads
→ Paid $15M to date
→ 264k downloads

We're committed to spending $1B the next 12 months on data & com… / X](https://x.com/humanoidsdaily/status/2092317847645032776)
  Post Log in Sign up Post Brett Adcock on X: "Introducing Index Today we're coming out of stealth with Index, the largest & most diverse robot dataset in the world → 30min of video uploads/sec → 16…
- [Brett Adcock on X: "Introducing Index

Today we're coming out of stealth with Index, the largest & most diverse robot dataset in the world

→ 30min of video uploads/sec
→ 16M video uploads
→ Paid $15M to date
→ 264k downloads

We're committed to spending $1B the next 12 months on data & com… / X](https://x.com/HumanoidInvest/status/2092397968041066740)
  Post Log in Sign up Post Brett Adcock on X: "Introducing Index Today we're coming out of stealth with Index, the largest & most diverse robot dataset in the world → 30min of video uploads/sec → 16…
- [Figure on X: "Introducing Index

Today we're coming out of stealth with the largest and most diverse robot dataset to scale general purpose robots

To date we've crossed 264,000 app downloads and uploaded 16 million videos" / X](https://x.com/Figure_robot/status/2092303621392376314)
  Post Log in Sign up Post Figure on X: "Introducing Index Today we're coming out of stealth with the largest and most diverse robot dataset to scale general purpose robots To date we've crossed 264,000…
- [@TheHumanoidHub on X](https://x.com/TheHumanoidHub/status/2092301463989203250)
  Show it a single human demonstration of a long-horizon task and the robot performs it right away. No fine-tuning, no retraining. Skild has released S1, a robot foundation model where you specify the t…
- [Skild AI on X: "We found that to match the accuracy that S1 can achieve with just one example of prompting, current VLA models would need to be post-trained with 50-100 hours of data collection followed by fine-tuning!" / X](https://x.com/SkildAI/status/2092300859317330378)
  Post Log in Sign up Post Skild AI on X: "We found that to match the accuracy that S1 can achieve with just one example of prompting, current VLA models would need to be post-trained with 50-100 hours …
- [@SkildAI on X](https://x.com/SkildAI/status/2092300854359625814)
  Compared to conventional VLAs, S1 is a step-change improvement. For known tasks, S1 can match the performance of language-prompted VLAs. For novel tasks, S1 exponentially outperforms any existing lang…
- [Skild AI on X: "S1 learns new tasks like a language model.

You prompt it with a video demonstration, and it outputs robot actions to complete the task in any environment and in any embodiment." / X](https://x.com/SkildAI/status/2092300850492481818)
  Post Log in Sign up Post Skild AI on X: "S1 learns new tasks like a language model. You prompt it with a video demonstration, and it outputs robot actions to complete the task in any environment and i…
- [Skild AI on X: "The first time S1 flipped a pancake, we assumed pancake flipping must have been in its pre-training data.

We searched our whole pre-training data and found no examples of flipping.

S1 inferred the out-of-distribution task from one video prompt." / X](https://x.com/SkildAI/status/2092300848034635790)
  Post Log in Sign up Post Skild AI on X: "The first time S1 flipped a pancake, we assumed pancake flipping must have been in its pre-training data. We searched our whole pre-training data and found no …
- [Skild AI on X: "Introducing S1, our new foundation model that learns from one example.

It can be taught 10-minute long tasks that it has never seen before, from one video prompt without any fine-tuning.

Watch S1 operate in real-time via in-context learning:" / X](https://x.com/Scobleizer/status/2092306192773345691)
  Post Log in Sign up Post Skild AI on X: "Introducing S1, our new foundation model that learns from one example. It can be taught 10-minute long tasks that it has never seen before, from one video prom…
- [Skild AI on X: "Introducing S1, our new foundation model that learns from one example.

It can be taught 10-minute long tasks that it has never seen before, from one video prompt without any fine-tuning.

Watch S1 operate in real-time via in-context learning:" / X](https://x.com/SkildAI/status/2092300842900865389)
  Post Log in Sign up Post Skild AI on X: "Introducing S1, our new foundation model that learns from one example. It can be taught 10-minute long tasks that it has never seen before, from one video prom…
- [@SexyTechNews on X](https://x.com/SexyTechNews/status/2092286333394006159)
  FCC "Buy American" rule walls off foreign humanoids. New Covered List action blocks new foreign humanoid/quadruped models from US equipment authorization via a domestic-content test — directly hits BY…
- [NEURA Gym: Physical AI Training and Infrastructure | NEURA](https://x.com/NEURARobotics/status/2092183035747700844)
  Skip to content The Fastest Path from Human Expertise to Deployable Robot Skills Get in contact \\ \\ The Challenges Facing Industrial Leaders Today Physical AI needs real-world training data: vision,…
- [中国机器人“跑赢”博尔特，这意味着什么？ - 纽约时报中文网](https://x.com/RFAChinese/status/2092672676141654233)
  中国 - 中文中 - 中英双语双语 - 英文 英 YAN ZHUANG2026年8月25日 Achmad Ibrahim/Associated Press 2009年，尤塞恩·博尔特在柏林世界田径锦标赛上创造了9.58秒的男子 100米短跑世界纪录。据中国官方媒体报道，上周六，由中国公司北京人形机器人创新中心有限公司(X-Humanoid)开发的一款人形机器人以0.19秒的优势打破了该纪录。这台机…
- [Introducing Index: Building The World’s Largest and Most Diverse Physical Dataset](https://www.figure.ai/news/introducing-index)
  Introducing Index: Building The World’s Largest and Most Diverse Physical Dataset August 25, 2026 Today we're coming out of stealth with the most diverse robot training dataset ever built. The data ne…
- [🤖🍨 Sundae Robotics 06: V-JEPA 2 & Predicting Physical Intelligence · Luma](https://x.com/EdmondIsARobot/status/2092037991619530779)
  Hosted By Edmond Jono Hart James (Jingxi) Xu Mene Mazarakis Angela Zhang Manfredi Bernardi Adam 134 Went Akshobhya Gupta, Hannah Tsui and 132 others Contact the Host Report Event AI 🤖🍨 Sundae Roboti…
- [XPeng Motors humanoid robot unit Dogotix raises $900M - The Robot Report](https://x.com/ologicinc/status/2094455408803086387)
  The Robot Report Continue to Site Facebook X LinkedIn Reddit Pinterest Share XPeng’s Dogotix unit is developing the IRON humanoid robot. Source: XPeng Like other global automakers, XPeng Motors is bra…
- [RoboPapers on X: "Instead of choosing between training a world model and training a language conditioned robot policy, why not do both? LDA-1B s a new foundation model that is trained on 30,000 hours of human and robot interaction data.

Part of the secret is that LDA-1B jointly learns forward … / X](https://x.com/chris_j_paxton/status/2091877313872834800)
  Post Log in Sign up Post RoboPapers on X: "Instead of choosing between training a world model and training a language conditioned robot policy, why not do both? LDA-1B s a new foundation model that is…
- [@RoboPapers on X](https://x.com/RoboPapers/status/2091877173292356012)
  Instead of choosing between training a world model and training a language conditioned robot policy, why not do both? LDA-1B s a new foundation model that is trained on 30,000 hours of human and robot…
- [@TheHumanoidHub on X](https://x.com/TheHumanoidHub/status/2090662734551818253)
  Generalist AI just released GEN-1.5. It might be robotics' GPT-3 moment. What is one-shot learning via in-context prompting? In the case of language models like GPT-3, including a Q&A example in the p…
- [Bringing Marble to Life | World Labs](https://x.com/theworldlabs/status/2090479436550602863)
  Nov 12, 2025A behind-the-scenes look at how Marble powered the creation of its own launch story. Bringing Marble to Life World Labs: Just Imagine - YouTube Tap to unmute World Labs: Just Imagine World…
- [Anirudha Majumdar on X: "Very cool work, Pete! I'm curious if you've tried an extreme version of this: no in-context example at all; just place objects in front of the robot and see if it can infer the task. 

I suspect this will have non-trivial success rates, which would also allow you to figure … / X](https://x.com/peteflorence/status/2090592251286335962)
  Post Log in Sign up Post Anirudha Majumdar on X: "Very cool work, Pete! I'm curious if you've tried an extreme version of this: no in-context example at all; just place objects in front of the robot a…
- [Post-Train NVIDIA Cosmos 3 Edge for On-Device Robot Control | NVIDIA Technical Blog](https://x.com/NVIDIARobotics/status/2095179993974260168)
  Technical Blog Subscribe Related Resources Robotics Post-Train NVIDIA Cosmos 3 Edge for On-Device Robot Control Aug 19, 2026 By Saeed Babamohamadi +12 Like Discuss (0) - L - T - F - R - E AI-Generated…
- [Hydra-0: Action Flow for Generalist World Modeling and Control | NVIDIA Isaac](https://x.com/NVIDIARobotics/status/2094515571501015339)
  Overview video Figure 1: Action flow as a shared control interface. Top: Hydra-0 learns from diverse interaction videos featuring egocentric human demonstrations, handheld UMI grippers, bimanual robot…
- [Press-Information June 25th 2026](https://www.press.bmwgroup.com/canada/article/attachment/T0458794EN/651051)
  Press-Information June 25th 2026 BMW Group advances the use of Physical AI in production with Figure 03 project in Spartanburg - Figure AI demonstrates Figure 03 humanoid robots in new use case at BMW…
- [5 Physical AI infrastructure platforms shaping robotics in 2026 - The Robot Report](https://www.therobotreport.com/5-physical-ai-infrastructure-platforms-shaping-robotics-in-2026)
  Facebook X LinkedIn Reddit Pinterest Share From accelerated computing and simulation to data operations, open-source tooling, validation engineering, and continuous learning, these five platforms repr…
- [5 Physical AI infrastructure platforms shaping robotics in 2026 - The Robot Report](https://x.com/therobotreport/status/2094493447029670279)
  Facebook X LinkedIn Reddit Pinterest Share From accelerated computing and simulation to data operations, open-source tooling, validation engineering, and continuous learning, these five platforms repr…
- [Gemini Robotics 2 brings whole body intelligence to robots — Google DeepMind](https://x.com/SkillupAi/status/2093896499465945235)
  Skip to main content July 30, 2026 Models Gemini Robotics 2 brings whole body intelligence to robots Carolina Parada Share Your browser does not support the video tag. From feet to fingertips — we are…
- [Gemini Robotics 2 brings whole body intelligence to robots — Google DeepMind](https://x.com/TheTechFuse/status/2094178345982607455)
  Skip to main content July 30, 2026 Models Gemini Robotics 2 brings whole body intelligence to robots Carolina Parada Share Your browser does not support the video tag. From feet to fingertips — we are…
- [Sylvia Li on X: "Spent the day at Deep Tech Week and left with one strong conviction:

the next wave of AI will not be won only by better models.

It will be won by the infrastructure that lets intelligence touch the physical world.

For anyone building in robotics, chips, manufacturing, energy, or… / X](https://x.com/smsehy/status/2095050189354742023)
  Post Log in Sign up Post Sylvia Li on X: "Spent the day at Deep Tech Week and left with one strong conviction: the next wave of AI will not be won only by better models. It will be won by the infrastr…
- [Coupled Local and Global World Models for Efficient First Order RL](https://coupled-global-local-wm-rl.pages.dev/)
  Real-world tasks solved zero-shot by policies trained entirely inside our learned world models: Push-T with a tabletop manipulator (left), Ego-Centric Grasp and Lift with a G1 humanoid (centre), and E…
- [Coupled Local and Global World Models for Efficient First Order RL](https://x.com/RoboPapers/status/2092642979466915921)
  Real-world tasks solved zero-shot by policies trained entirely inside our learned world models: Push-T with a tabletop manipulator (left), Ego-Centric Grasp and Lift with a G1 humanoid (centre), and E…
- [Coupled Local and Global World Models for Efficient First Order RL](https://x.com/RoboPapers/status/2093005367567347970)
  Real-world tasks solved zero-shot by policies trained entirely inside our learned world models: Push-T with a tabletop manipulator (left), Ego-Centric Grasp and Lift with a G1 humanoid (centre), and E…
- [Yann LeCun's AMI Labs raises $1.03B to build world models | TechCrunch](https://techcrunch.com/2026/03/09/yann-lecuns-ami-labs-raises-1-03-billion-to-build-world-models)
  Checking your Browser… Verifying... Stuck? Troubleshoot Success! Verification failed Troubleshoot Verification expired Refresh Verification expired Refresh Troubleshoot Cloudflare, opens in a new tab …
- [Yann LeCun's AMI Labs raises $1.03B to build world models | TechCrunch](https://x.com/MikeTamir/status/2093444518494237175)
  Checking your Browser… Verifying... Stuck? Troubleshoot Success! Verification failed Troubleshoot Verification expired Refresh Verification expired Refresh Troubleshoot Cloudflare, opens in a new tab …
- [An Electric New Era for Atlas | Boston Dynamics](https://x.com/BostonDynamics/status/1780603212359205323)
  We value your privacy We use cookies to enhance your browsing experience, serve personalized ads or content, and analyze our traffic. By clicking "Accept All", you consent to our use of cookies. Custo…
- [Physical-Intelligence/openpi](https://github.com/Physical-Intelligence/openpi)
  # openpi openpi holds open-source models and packages for robotics, published by the [Physical Intelligence team](https://www.physicalintelligence.company/). Currently, this repo contains three types …
- [OpenVLA:](https://arxiv.org/pdf/2406.09246)
  OpenVLA: An Open-Source Vision-Language-Action Model Moo Jin Kim∗,1 Karl Pertsch∗,1,2 Siddharth Karamcheti∗,1,3 Ted Xiao4 Ashwin Balakrishna3 Suraj Nair3 Rafael Rafailov1 Ethan Foster1 Grace Lam Panna…
- [From Archive to Production: A Hybrid Workflow with Marble | World Labs](https://www.worldlabs.ai/case-studies/from-archive-to-production)
  Aug 12, 2026How Promise used Marble to transform archival imagery into immersive 3D environments for Harlan Coben’s Final Twist. From Archive to Production: A Hybrid Workflow with Marble Using Hybrid …
- [NVIDIA/Isaac-GR00T](https://github.com/NVIDIA/Isaac-GR00T)
  <div align="center"> <img src="media/header_compress.png" width="800" alt="NVIDIA Isaac GR00T N1.7 Header"> <!-- --- --> <p style="font-size: 1.2em;"> <a href="https://developer.nvidia.com/isaac/gr00t…
- [π0: A Vision-Language-Action Flow Model for](https://arxiv.org/pdf/2410.24164)
  π0: A Vision-Language-Action Flow Model for General Robot Control Physical Intelligence Kevin Black, Noah Brown, Danny Driess, Adnan Esmail, Michael Equi, Chelsea Finn, Niccolo Fusai, Lachy Groom, Kar…
- [Bringing Marble to Life | World Labs](https://www.worldlabs.ai/case-studies/bringing-marble-to-life)
  Nov 12, 2025A behind-the-scenes look at how Marble powered the creation of its own launch story. Bringing Marble to Life World Labs: Just Imagine - YouTube Tap to unmute World Labs: Just Imagine World…
- [🤖🍨 Sundae Robotics 06: V-JEPA 2 & Predicting Physical Intelligence · Luma](https://luma.com/5h9owhri)
  Hosted By Edmond Jono Hart James (Jingxi) Xu Mene Mazarakis Angela Zhang Manfredi Bernardi Adam 134 Went Akshobhya Gupta, Hannah Tsui and 132 others Contact the Host Report Event AI 🤖🍨 Sundae Roboti…
- [How Figure Became the Biggest Name in Robotics | XMAQUINA DAO](https://www.xmaquina.io/blog/how-figure-became-the-biggest-name-in-robotics)
  Genesis Auction Wave 2! Launches June 24 Get DEUS homeDAO Portal How Figure Became the Biggest Name in Robotics Color theme: Four years, three generations of humanoids, a $39 billion valuation and now…
- [huggingface/lerobot](https://github.com/huggingface/lerobot)
  <p align="center"> <img alt="LeRobot, Hugging Face Robotics Library" src="./media/readme/lerobot-logo-thumbnail.png" width="100%"> </p> <div align="center"> [![Tests](https://github.com/huggingface/le…
- [AgiBot World Colosseo: A Large-scale Manipulation Platform](https://arxiv.org/pdf/2503.06669)
  AgiBot World Colosseo: A Large-scale Manipulation Platform for Scalable and Intelligent Embodied Systems Team AgiBot-World∗ Project website: https://agibot-world.com/ Code: https://github.com/OpenDriv…
- [World Labs Atlas](https://www.worldlabs.ai/blog/atlas)
  September 1, 2026Introducing Atlas, our new omni world model for spatial intelligence. Atlas: A World Model for Spatial Intelligence World models generate, reconstruct, and simulate any possible world…

---

Source: Roboterminal · https://roboterminal.app/topics/models.md