Exploring the moon's surface with AI. 🌙 @FireflySpace, an #NVIDIAInception member, is preparing to launch its Blue Ghost Mission 2, including: 🔭 Lunar lander carrying science and technology instruments to detect faint signals from the cosmic dark ages 🛰️ Elytra spacecraft in lunar orbit for a five-year mission 📸 Ocula moon imaging service, powered onboard by the NVIDIA Jetson edge AI platform This builds on Firefly's Blue Ghost Mission 1, launched in March 2025, that downlinked ~120GB of raw lunar surface imagery and video data to Earth. Explore AI on the moon ➡️ nvda.ws/3R8XScQ

Scientist develop diving suits for cyborg cockroach swarms - and it could enable them to explore Mars dailymail.com/sciencetech/ar… #robotics @NVIDIARobotics @UBTECHRobotics @21stRobotics @roboticsfeed @TaiyoRobotics #spcaexploration @SpaceX #marsexploration @_SpaceWeather_ @space @CNN

cool to see robotics stepping up safety while also getting crowd support—shows where interest really lands

a $600k chunk is solid momentum. crowdfunded hardware scaling feels rare these days

How is AI changing robotics? 💭 Featuring: @openmind_agi, @NEURARobotics, @FANUCAmerica, @ABBRobotics, and @in_bolt 🦾

Happy 4th of July! Here's what happened in Physical AI this week: - @VerkadaHQ announces investment from @NVIDIARobotics to accelerate development and deployment of its Physical AI solution in schools, hospitals, retailers, and manufacturers leveraging NVIDIA Cosmos and Physical AI Data Factory - The Korean Conglomerates - @Hanwha_Official Group, @HMGnewsroom, @Samsung and SK Telecom - roll out $201B in Physical AI and infrastructure investments across small modular reactors, satellites, launch vehicles, AI data centers, autonomous driving, manufacturing and aviation. In addition, Samsung, GS Group and NAVER Corp will spend $519B to establish semiconductor complex and AI data centers in country's southwestern region - @luxonis raises $14M in Series A funding to scale production of its OAK cameras and cloud perception ecosystem for industrial use cases in agricutlure, defense, industrial automation, medtech and warehousing - @PuduRobotics deployed its four of its commercial service robots at Davos Tech Summit 2026 across three locations: SPAR Supermarket, Hotel Services and Train Station - @AmbiRobotics and @pickle_robot deploy integrated Physical AI solution to solve automated truck unloading and palletizing within logistics. Pickle unloads boxes from the truck onto conveyor belt. Ambi handles the palletizing by picking up boxes from the conveyor belt. - CarbonSix raises $40M in Series A funding led by DSC Investment, LB Investment to accelerate development and deployment of its specialized SigmaKit gripper single-arm package that can be trained using ego-centric UMI gripper data collection across manufacturing use cases. - @Kawasaki_JPN Heavy Industries, Ltd. to raise $1B via new shares and bonds to fund AI investments including its four-legged Corleo vehicle aka wolf rider bike - Thai startup Amity Robotics raises $7M seed round led by @eastventures with participation from 500 Global to launch its ARC Move product, an autonomous mobile robot with a giant touch screen for telecommunication We're doing some cool stuff too @MeckaAI.

Yes I just left a comment on the tweet linking the paper.

Nvidia CEO Jensen Huang's signature black leather jacket is about to be auctioned at Sotheby's. This "tech world battle robe," estimated at over 400,000 RMB, not only symbolizes the AI ​​revolution but will also be used to charitably support the next generation of tech talent. Sotheby's will publicly display the item in its New York showroom from July 7th to July 16th, with online bidding closing on July 17th. The auction house estimates the final sale price to be between $40,000 and $60,000 (approximately 272,000 to 408,000 RMB at the current exchange rate).#nvidia #jensenhuang #jacket @nvidia @NVIDIARobotics

Did you read this critique of the prevailing conception of world models (including JEPA):

I'll be there !

NVIDIA AI Introduces ASPIRE: A Self-Improving Robotics Framework Reaching 31% Zero-Shot on LIBERO-Pro Long Tasks Most robot-coding agents throw away everything they learn. Solve a task, discard the fix, start the next one cold — the agent on its 100th task is no smarter than on its first. NVIDIA's ASPIRE draws a clean line between that and an agent whose experience actually compounds. They introduced ASPIRE (Agentic Skill Programming through Iterative Robot Exploration) — a code-as-policy system where a coding agent (Claude Code, Claude Opus 4.6, 1M-token context) writes and debugs its own robot programs against a fixed perception/planning/control API, and distills every validated fix into a reusable skill library, with no fixed perception-plan-execute pipeline anywhere in the loop. Here's what's actually interesting: → The execution engine logs per-primitive multimodal traces — RGB keyframes, grasp candidates, object poses, motion plans, return status — so the agent localizes the failing primitive, not just the failed rollout → Validated fixes distill into a text skill library (failure signature when-to-apply guard repair sketch), not weights — and the agent is barred from reading sim ground truth, so the skills transfer to real hardware → Evolutionary search proposes K candidate programs per round, conditioned on surviving programs residual failure traces — beyond single-trajectory tuning → LIBERO-Pro Object under perturbation: 98 vs 22 for CaP-Agent0 → Robosuite bimanual handover: 92 vs 20 for CaP-Agent0 → LIBERO-Pro Long zero-shot: 31 vs 4 for prior methods (skills learned on LIBERO-90, no test-time retries) On a real bimanual robot with a different embodiment and API (OpenAI Codex GPT-5.5), transferred skills took soda-can lifting to 19/20 at ~10x fewer tokens, and drawer opening from 0/20 to 11/20. Full analysis: marktechpost.com/2026/07/03/… Paper: arxiv.org/pdf/2607.00272 Project page: research.nvidia.com/labs/gea… @NVIDIAAI @nvidia @NVIDIARobotics @DrJimFan @guanzhi_wang @UMichCSE @ECEILLINOIS @berkeley_ai @CMU_Robotics @runyulu_x @yubowu25 @letian_fu @_wenlixiao @AjayMandlekar @YinzhenXu @GuanyaShi @Ken_Goldberg @mosharaf @yukez @DrJimFan

A virtual physics world model, with symbolic modelling seems to be where the puck is going. Many trajectories are pointing in that direction, including Marble from @theworldlabs and JEPA from @ylecun Isaac Sim from @NVIDIARobotics and many independent devs. A common vocabulary is yet to emerge but I expect it will arrive shortly.

lmao @openmind_agi is the only robot puppy interviewee

$bb $nvda @nvidia @NVIDIARobotics @john_giamatteo Imho matter of time before Nvidia takes a stake in BB around $25-50, BB stock shoots to $50-100. Maybe a week, month or year. Sooner the better so as partners they dominate Physical AI. bb $500-1,000 per share by 2030-2033?

Hi @NVIDIARobotics! I ran into an issue while setting up my machine to the latest Jetson ISO r39.2. Unfortunately, it required me to install an older version 38 then upgrade to the latest r39.2 This isn't mentioned anywhere in the documentation guide docs.nvidia.com/jetson/agx-t…