I'll be in Austin for #CoRL2026 this week! ๐ค
On Wednesday, I'll share some of our recent progress in physical AI for manufacturing.
I'm also interested in visual RL, post-training, and multimodal foundation models. Would love to connect with anyone exploring these areas, so come say hi or DM me to grab a coffee โ
See you there!
Path Robotics retweeted
๐ท Thanks again to Patron Sponsor @PathRobotics! See you in Austin at #CoRL2026 ๐ค ๐ค
Path Robotics retweeted
Replying to @PathRobotics
@PathRobotics CRO explains how physical AI can help manufacturers extend operations, even if labor and floor space are limited.
For small and midsize manufacturers in particular, this can mean extending shifts, increasing throughput, and taking on more orders without the time and capital needed to expand their physical footprint.
Watch the full episode: bit.ly/4jUXqea
Path Robotics retweeted
First speakers announced for the AI+ Robotics Summit, part of the SCSP AI+ Summit Series.
On Oct 14 in Washington, DC, leaders from robotics, manufacturing, investment, and research will examine how the U.S. can turn world-class AI into industrial capacity and deploy intelligent machines at scale.
Panels, fireside chats, tech talks, and demos. Speakers include leaders from @FANUCAmerica, @nvidia, @BostonDynamics, @agilityrobotics, @IntrinsicAI, MxD, and more.
More speakers coming soon. Registration link in the replies.
Path Robotics retweeted
Rating Robots at Actuate 2026
Scott (@GoingBallistic5) spent a day at Actuate in San Francisco getting hands on with the robots on the show floor and giving each one a score.
0:00 Actuate 2026
0:15 IO-AI Tech
1:50 @DexmateAI
3:00 @PathRobotics
3:40 @trossenrobotics
5:25 @UFBots
6:50 Final rankings
Path Robotics retweeted
I visited @PathRobotics yesterday to see how its physical AI technology is complementing highly skilled workers and creating new opportunities to expand American manufacturing.
๐Columbus, Ohio
Path Robotics retweeted
I spoke with CNBC Morning Call to discuss our $900M deal with HII and how physical AI is helping shipbuilders increase production capacity. Watch here: lnkd.in/gb4ysYtD.
Thanks to Morgan Brennan and team for having me on the show!
Path Robotics retweeted
Replying to @PathRobotics
@PathRobotics is building robotic tools to increase the productivity of Americaโs shipbuilding workforce, keeping Ohio at the forefront of American manufacturing.
Path Robotics retweeted
America is running out of welders. By 2035, weโll lose 43% of America's welding workforce. @PathRobotics is building robots to fill the gap.
On episode 75, S3 went through Path's weld school and captured their journey from a $300K bet with their first customer to speaking before Congress.
Path Robotics retweeted
Weโre putting investors to the test: @JacklouisP from @RoboStrategy tried to out-weld the robots today.
Who do you think did it better - Rove or Jack?
Spoiler: Rove won.
Path Robotics retweeted
Path Robotics has developed a legged robot called Rove for large-scale industrial welding.
It moves to the worksite and scans welding seams to understand their shape.
Then, its AI performs welding and adjusts the process in real time.
Saronic is rapidly advancing US shipbuilding to deliver autonomous ships at speed and scale.
Excited to share our work with @PathRobotics to integrate physical AI for welding robotics into our shipbuilding workflows in Franklin, LA โ bringing together our world-class welders and advanced technologies to drive gains in safety, efficiency, and production quality.
Learn more about how weโre utilizing advanced manufacturing capabilities and techniques to accelerate production โ today in Franklin, and tomorrow at our next-generation shipyard, Port Alpha: medium.com/saronic-technologโฆ
Path Robotics retweeted
๐๐ถ๐ด๐ต-๐บ๐ถ๐
๐บ๐ฎ๐ป๐๐ณ๐ฎ๐ฐ๐๐๐ฟ๐ถ๐ป๐ด ๐ฏ๐ฟ๐ผ๐ธ๐ฒ ๐๐ฟ๐ฎ๐ฑ๐ถ๐๐ถ๐ผ๐ป๐ฎ๐น ๐ฟ๐ผ๐ฏ๐ผ๐๐.
For decades, robots thrived in high-volume, low-mix production.
But the moment factories moved to high-mix, low-volume, traditional automation stopped making sense.
On this episode of The Neuracore Podcast, we chatted with @NimaGard, Head of AI at @PathRobotics, who explains why setup time is now the real bottleneck - and why AI-powered automation is the path forward.
Listen to the full conversation with Stephen James here:
youtu.be/ih_zDd2tusk?si=SQbvโฆ
Path Robotics retweeted
We just cleared 10,000 sq-ft for an insane new facility addition.
The future of defense tech manufacturing coming Feb 17th...
Path Robotics retweeted
Thanks to @NimaGard and the team at @PathRobotics for taking part in our inaugural episode!
Today we are excited to introduce ๐ฅ๐ผ๐ฏ๐ผ๐ ๐๐ฒ๐ฎ๐ฟ๐ป๐ถ๐ป๐ด ๐ถ๐ป ๐๐ป๐ฑ๐๐๐๐ฟ๐, a new Neuracore original series exploring how machine learning is transforming robotics across every major industry.
In this series, we speak directly with the engineers, researchers, and technical leaders deploying ML in the physical world.
From factory floors to warehouses, from hospitals to energy sites, and across any environment where robots must operate with real-world variability.
To launch the series, we sat down with someone building that future from the inside.
@NimaGard is the Director of AI at @PathRobotics, the company pioneering autonomous welding systems and developing one of the first foundation models for industrial manufacturing.
Welding is one of the hardest problems in robotics. The variability, the precision, the real-world messiness all of it breaks most automation systems. And yet Path believes that with the right data, the right models, and the right hardware, welding can become the proving ground for true physical AI.
Nima is leading a team training large-scale models on massive, unsupervised datasets collected from robots deployed across global factory floors. They are building custom scanning hardware. They are unifying data pipelines from worldwide fleets. They are pushing machine learning into places where failure is not theoretical, it is molten metal.
So we asked him:
โข Why welding is the perfect proving ground for industrial foundation models
โข The importance of high-fidelity scanning before, during, and after welds
โข How every data point, good or bad, strengthens the model and accelerates improvements
โข The technical infrastructure required to train at production scale across a global fleet
โข How their methods extend far beyond welding to other core manufacturing workflows
Teams across robotics-driven industries are searching for clarity about what machine learning can truly deliver today.
This series exists to provide exactly that clarity.
youtu.be/caqlk8LCK7Q?si=QR5Xโฆ
Path Robotics retweeted
Today we are excited to introduce ๐ฅ๐ผ๐ฏ๐ผ๐ ๐๐ฒ๐ฎ๐ฟ๐ป๐ถ๐ป๐ด ๐ถ๐ป ๐๐ป๐ฑ๐๐๐๐ฟ๐, a new Neuracore original series exploring how machine learning is transforming robotics across every major industry.
In this series, we speak directly with the engineers, researchers, and technical leaders deploying ML in the physical world.
From factory floors to warehouses, from hospitals to energy sites, and across any environment where robots must operate with real-world variability.
To launch the series, we sat down with someone building that future from the inside.
@NimaGard is the Director of AI at @PathRobotics, the company pioneering autonomous welding systems and developing one of the first foundation models for industrial manufacturing.
Welding is one of the hardest problems in robotics. The variability, the precision, the real-world messiness all of it breaks most automation systems. And yet Path believes that with the right data, the right models, and the right hardware, welding can become the proving ground for true physical AI.
Nima is leading a team training large-scale models on massive, unsupervised datasets collected from robots deployed across global factory floors. They are building custom scanning hardware. They are unifying data pipelines from worldwide fleets. They are pushing machine learning into places where failure is not theoretical, it is molten metal.
So we asked him:
โข Why welding is the perfect proving ground for industrial foundation models
โข The importance of high-fidelity scanning before, during, and after welds
โข How every data point, good or bad, strengthens the model and accelerates improvements
โข The technical infrastructure required to train at production scale across a global fleet
โข How their methods extend far beyond welding to other core manufacturing workflows
Teams across robotics-driven industries are searching for clarity about what machine learning can truly deliver today.
This series exists to provide exactly that clarity.
youtu.be/caqlk8LCK7Q?si=QR5Xโฆ




























