Your NVIDIA Jetson Orin Nano creations inspired us. 💡 With Jetson Orin Nano 2 coming, the possibilities are just beginning. What will you build? 🛠️
NVIDIA Robotics
Computer Hardware Manufacturing
Santa Clara, California 583,642 followers
Inspiring visionaries and developers to create the next gen of AI-driven robots and explore the world of physical AI.
About us
The NVIDIA Robotics platform accelerates the development of AI-driven robots, streamlining processes from design and simulation to deployment. It enables key functions like navigation, mobility, grasping, and vision, supporting robotics across industries such as manufacturing, agriculture, logistics, and healthcare.
- Website
-
https://www.nvidia.com/en-us/industries/robotics/
External link for NVIDIA Robotics
- Industry
- Computer Hardware Manufacturing
- Company size
- 10,001+ employees
- Headquarters
- Santa Clara, California
Updates
-
Ready to take robots from development to deployment? Join our livestream on August 26 at 9 a.m. PT to learn how to: 🧠 Develop and deploy robots with NVIDIA Isaac GR00T and Jetson Thor 🏭 Apply sim to real workflows for industrial robotics 🦾 Turn teleoperation data into real robot actions 🎙️ See Noble Machines and Seeed Studio demonstrate practical robotics workflows across humanoids and robotic arms.
From Robot Development to Deployment with Isaac GR00T & Jetson Thor
www.linkedin.com
-
NVIDIA Robotics reposted this
Last year's ~1T parameter frontier model intelligence is now available in small and medium models - these models run real-time on Nvidia Jetson. This amazing capability is unlocking use-cases that we could only dream of previously.
-
NVIDIA Robotics reposted this
Announcing 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. How? 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. By composing skills learned in pretraining or creating new ones, it can make coffee, pot a plant, fry pancakes, and more – tasks never seen during pretraining. 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 of a million hours and found no examples of flipping. S1 inferred the out-of-distribution task from one video prompt. S1 does not blindly replay the video demonstration. Instead, it displays common-sense understanding that goes beyond the video prompt. It can withstand perturbations and improvise upon mistakes, even if the human didn’t. At times, S1 executes with more precision than the human in the video prompt 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 language-prompted VLAs as we scale the pre-training. 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! We believe this is a promising direction towards establishing scaling laws for robotics. We're excited to deploy S1 with our limited industrial partners today, rolling out to more customers over the next few months. S1 is the first light on a new path towards general intelligence grounded in the physical world. Technical blog: https://lnkd.in/g7ANeDkb
-
Introducing NVIDIA Jetson Orin Nano 2, built to bring frontier intelligence to entry-level edge AI. ⚡ It delivers 2x the inference performance of Jetson Orin Nano Super while matching its performance using 40% less power in 15W mode, all in the same compact form factor. Jetson Orin Nano 2 will give developers the room they need to build advanced AI at the edge. Read the announcement 👉 https://nvda.ws/4gRL1G0
-
-
Last week at MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), more than 85 students, postdocs, and faculty joined NVIDIA for a Robotics Day featuring remarks from Dr. Daniela Rus and Dr. Song Han. 🙌 Thank you to everyone who took part in the conversations, collaboration, and hands-on learning. We’re excited to see the next generation of robotics researchers build what’s next. 🦾
-
-
NVIDIA Robotics reposted this
Last week, I had the chance to join my NVIDIA colleagues Moritz Reuss and Danfei Xu for a livestream on WAMs and VLAs for robot learning. One question kept coming up: Do we have to choose between them? My take is that the choice might not come down to one single model. As hardware continues to improve and agentic systems mature, hybrid approaches could combine the strengths of WAMs and VLAs—with a controller coordinating the right models for the task. Here’s a short clip from our discussion. Thanks to everyone who joined live and shared such thoughtful questions. Watch the full livestream: https://lnkd.in/gJEwr5NH
-
More edge AI capability starts with getting more from the hardware you already have. Teams building with NVIDIA Jetson are proving it: • UBTECH Robotics reduced memory use by up to 15GB. • SandStar AI Retail moved from 16GB to 8GB NVIDIA Jetson Orin NX with no performance loss. • Connect Tech Inc.'s CTai LABS moved a dual-model AI system from the Jetson AGX Orin 64GB to the 32GB module while improving throughput. • GROOVE X and NoTraffic are applying the same approach to robotics and traffic management. Learn how NVIDIA Jetson Agent Skills can help developers get more from their existing systems in the full EE Times | Electronic Engineering Times article: https://nvda.ws/3SMmJE8
-
-
Humanoid robots are learning to move with more agility and control. Learn what SONIC is and how it’s helping advance whole-body control. 👇
Nvidia’s SONIC Teaches Humanoids to Move: Nvidia’s model uses real-time human demonstrations and training data to give operators a one-stop shop for humanoid motion.
-
NVIDIA Robotics reposted this
Last week, I had the chance to discuss current efforts in robotics foundation models with Thomas Ran Tian and Danfei Xu as part of the Cosmos Labs series of NVIDIA Robotics. We talked about World Action Models, how they compare with VLAs, and how the two approaches might work together. Here’s a short clip from the conversation. You can watch the full discussion here: https://lnkd.in/eTdWHhZU
