Jonathan Stephens

11K posts

Jonathan Stephens banner
Jonathan Stephens

Jonathan Stephens

@jonstephens85

Synthetc Data | World Models | Radiance Fields | Computer Vision | GenAI | Chief Evangelist at @LightwheelAI

Salem, OR Katılım Şubat 2009
1.6K Takip Edilen12.9K Takipçiler
Jonathan Stephens
Jonathan Stephens@jonstephens85·
Wow! I was not expecting to see myself in this amazing video that @Scobleizer produced at @NVIDIAGTC! You can find me at 19:07 into the episode. The agentic AI enthusiasm was so high at the Build-A-Claw! Very exciting!
China Asset Management@ChinaAMC_HQ

Live from NVIDIA GTC 2026-The Rise of Physical AI: Robots Enter the ChatGPT Moment Live from @nvidia GTC 2026, our correspondent Robert Scoble @Scobleizer is on the ground exploring how AI is moving beyond models and into infrastructure, robotics, and real-world deployment.— here's what stood out on Day 3: OpenClaw is emerging as a new AI infrastructure entry point —a faster, more accessible local inference on open models. Key takeaways from the floor: • Robotics is entering an AI-driven "renaissance" — from hard-coded to adaptive systems • Humanoids = the next computing platform • Agentic AI is reshaping how robots are designed & deployed • NVIDIA's ecosystem strategy is expanding across startups, open models & physical AI Watch our Day 3 field report and stay with us as we continue exploring the future of AI. #AI #GTC2026 #Robotics #OpenModels #PhysicalAI #NVIDIA

English
0
0
0
151
Jonathan Stephens
Jonathan Stephens@jonstephens85·
This was my busiest @NVIDIAGTC ever! I had aspirations of posting all day long, but I was lucky to just stop and eat for 5 minutes. I enjoyed every minute of it though! Can’t wait for GTC 2027! I’ll be posting more content soon.
Jonathan Stephens tweet mediaJonathan Stephens tweet media
English
0
0
2
85
Jonathan Stephens retweetledi
Lightwheel
Lightwheel@LightwheelAI·
We’re live at @NVIDIAGTC. 🚀 All week at Booth 1406 right by the entrance to Hall 2, the energy around @LightwheelAI has been incredible. If you’re at @NVIDIAGTC, come say hi.
English
0
1
6
632
Chris Paxton
Chris Paxton@chris_j_paxton·
Robot Olaf walking the gtc floor
English
12
182
940
49.3K
Jonathan Stephens
Jonathan Stephens@jonstephens85·
I can’t even begin to describe my day 2 at @NVIDIAGTC. The momentum in Physical AI is building. The speed at which @NVIDIARobotics continues to innovate with partners like @LightwheelAI is truly impressive. Deep thanks to NVIDIA for hosting an amazing conference.
Jonathan Stephens tweet mediaJonathan Stephens tweet mediaJonathan Stephens tweet mediaJonathan Stephens tweet media
English
0
2
11
564
Jonathan Stephens
Jonathan Stephens@jonstephens85·
@LightwheelAI @NVIDIAGTC What a honor to represent our incredible incredible work at Lightwheel and share the stage with true innovators in Physical AI.
English
0
0
0
20
Lightwheel
Lightwheel@LightwheelAI·
@LightwheelAI at @NVIDIAGTC : At the Center of the Robotics Spotlight ✨ At NVIDIA GTC, @jonstephens85 , Chief Evangelist at @LightwheelAI, was featured in a standout robotics session alongside leading voices from across the industry. Physical AI is moving from research milestones to real-world deployment, and Lightwheel is at the center of that shift. At NVIDIA GTC, Jonathan brought that momentum to one of the industry’s most visible stages. Great to share the stage with @akhil_docca of @nvidia , @mack2it of @workr_official , Rajesh Radhakrishnan of @ServeRobotics , @kevinmpeterson1 of @BedrockRobotics, and @Yao__Lu of @physical_int. #NVIDIAGTC #PhysicalAI #Robotics #EmbodiedAI #Leadership #Ecosystem
Lightwheel tweet mediaLightwheel tweet mediaLightwheel tweet media
English
1
1
7
898
Jonathan Stephens
Jonathan Stephens@jonstephens85·
What an honor it was to share the stage with a few of the best innovators in Physical AI today at @NVIDIAGTC! If you missed my talk, I explained how @LightwheelAI is building the three pillars of simulation infrastructure that makes physical AI successful: World, Behavior, and Eval. Still want to learn about what we are building? Come meet me at our booth Wednesday or Thursday this week. Or just come say hi! #PhysicalAI #robotics
Lightwheel@LightwheelAI

@LightwheelAI at @NVIDIAGTC : At the Center of the Robotics Spotlight ✨ At NVIDIA GTC, @jonstephens85 , Chief Evangelist at @LightwheelAI, was featured in a standout robotics session alongside leading voices from across the industry. Physical AI is moving from research milestones to real-world deployment, and Lightwheel is at the center of that shift. At NVIDIA GTC, Jonathan brought that momentum to one of the industry’s most visible stages. Great to share the stage with @akhil_docca of @nvidia , @mack2it of @workr_official , Rajesh Radhakrishnan of @ServeRobotics , @kevinmpeterson1 of @BedrockRobotics, and @Yao__Lu of @physical_int. #NVIDIAGTC #PhysicalAI #Robotics #EmbodiedAI #Leadership #Ecosystem

English
0
2
2
602
Jonathan Stephens
Jonathan Stephens@jonstephens85·
The physics of simulation just got a lot more serious. @NVIDIARobotics announced Newton 1.0 GA at @NVIDIAGTC this week, an open, GPU-accelerated physics engine built for contact-rich manipulation, locomotion, and deformable dynamics at scale. It's built on Warp and OpenUSD, integrates with Isaac Lab and Isaac Sim, and brings together solvers from @GoogleDeepMind (MuJoCo Warp) and Walt Disney Imagineering (Kamino) under one unified architecture. But what caught my attention isn't just the engine, it's what the Newton ecosystem is starting to demand. Faster simulation isn't enough anymore. Questions I keep getting at GTC is: how physically accurate is it? Can a cable bend the way a real cable bends? Can a connector seat the way it does on an actual factory floor? That gap between synthetic approximation and physical ground truth is where deployment-ready robotics lives or dies. That's exactly where @LightwheelAI is focused. I'm proud to share that Lightwheel has joined the Newton Technical Steering Committee under The Linux Foundation, alongside NVIDIA, Google DeepMind, Disney Research, and Toyota Research Institute. Our role is to lead SimReady schema development, drive real-to-sim calibration pipelines, and build the physically measured, solver-validated assets the Newton ecosystem needs to train policies that actually transfer. The Samsung cable manipulation demo in the blog is a concrete example of what this looks like in practice. Lightwheel built the SimReady assets behind it. Simulation infrastructure is becoming foundational to physical AI. We're building the standards and pipelines that make it real. developer.nvidia.com/blog/newton-ad…
English
0
2
18
1.1K
Jonathan Stephens
Jonathan Stephens@jonstephens85·
Wow!!! What a turn out for our @LightwheelAI Physical AI party! It was an honor to have @lukas_m_ziegler be our host. He encouraged everyone to say hi to new faces and have a good time. If you were at the party and have a few great photos, I would love if you share with me!
Jonathan Stephens tweet mediaJonathan Stephens tweet mediaJonathan Stephens tweet media
English
0
0
6
309
Jonathan Stephens retweetledi
Chris Paxton
Chris Paxton@chris_j_paxton·
This is the first time that I have seen Skild doing live demos in public! Demonstrating neural nets for precision manufacturing
English
16
48
389
103.3K
Jonathan Stephens retweetledi
Lightwheel
Lightwheel@LightwheelAI·
@NVIDIAGTC Behind Lightwheel’s SimReady Architecture: Measure → Solve → Generate As robotics models scale, the challenge is no longer just training, it is producing physically accurate simulation assets at scale. At @LightwheelAI , we are building simulation infrastructure for Physical AI to solve this problem. Our SimReady System introduces a unified architecture for generating physics-grounded simulation assets through a closed-loop pipeline: Measure → Solve → Generate • Measure Capture real-world properties such as friction, stiffness, deformation, and contact dynamics using our Physical Measurement Factory. • Solve Translate measured parameters into reliable simulation behavior through advanced physics solvers. • Generate Produce OpenUSD-based assets and environments at scale for robot learning, teleoperation data collection, and evaluation. Together, this architecture enables physically accurate simulation assets at scale, grounding robot learning in real-world physics. We are also collaborating with @NVIDIARobotics to advance simulation physics using the Newton Physics Engine, and enabling SimReady assets to integrate seamlessly with NVIDIA Isaac Sim and Isaac Lab workflows. Learn more about the SimReady Architecture: lnkd.in/gjjkNy59 learn more about OpenUSD in Physical AI simulation in NVIDIA’s GTC talk: lnkd.in/g__xxm9X #NVIDIAGTC #NVIDIA #NVIDIARobotics
Lightwheel tweet media
English
0
3
8
651
Jonathan Stephens retweetledi
張小珺 Xiaojùn
張小珺 Xiaojùn@zhang_benita·
Tokens are the new commodity, and compute is revenue:)
張小珺 Xiaojùn tweet media張小珺 Xiaojùn tweet media張小珺 Xiaojùn tweet media張小珺 Xiaojùn tweet media
English
11
6
106
12.6K
Jonathan Stephens
Jonathan Stephens@jonstephens85·
Token king!!!! This made me laugh. Extreme codesign is the way.
Jonathan Stephens tweet mediaJonathan Stephens tweet media
English
0
0
3
503
The Humanoid Hub
The Humanoid Hub@TheHumanoidHub·
We are the media now!
Ilir Aliu@IlirAliu_

I’ll be at the GTC this week, here with my friend Devang (@TheHumanoidHub) looking forward to see great sessions on robotics. Waiting for Jensen right now. One session I’m especially curious about today: ‘From Concept to Production: Humanoid Robotics at Scale’. nvidia.com/gtc/session-ca… With @pathak2206 (Skild AI)a @chelseabfinn (Stanford / Physical Intelligence), Pras Velagapudi (Agility Robotics), and Amit Goel from @Nvidia. Humanoid robotics needs to move from research to production. Let’s see what the people actually building it have to say….

English
4
3
47
7K

Keşfet

@Scobleizer @NVIDIAGTC @cixliv @chris_j_paxton @LightwheelAI @BenjaminDEKR @ErenChenAI @NVIDIARobotics