General and robust AI robots that power the future of the physical economy.
541411.1K344K381
Tags
<500KAI agentProduct launchSeries AB2BDemoUS
TLVC Rating
3.60
Hook
Editing / Creativity
Copy
Sentiment of launch
Distribution strategy
Community Rating
No ratings yet
to rate this launch.
Dyna Robotics builds general-purpose robotic arm systems powered by its own foundation models, aimed at businesses that need repetitive dexterous work done reliably. This launch marks the moment its robots crossed the ROI threshold in a paying customer deployment at Din Tai Fung, one of the highest revenue per location restaurant chains, moving the company past the pilot stage that has stalled much of the humanoid and manipulation field. For founders and operators evaluating physical AI, the significance is that a fleet is now covering its cost inside a real, high-throughput kitchen rather than a staged demo.
The company was founded in 2024 by Lindon Gao, Jason Ma and York Yang and is based in Redwood City. Lindon and York are repeat founders who previously built Caper AI, an AI-powered smart cart company which they sold to Instacart for $350M. Jason Ma is a former researcher at DeepMind and NVIDIA . Their approach centers on task-specialized foundation models trained and refined in production, with the model deployed at multiple customer sites, supporting generalization and commercial viability as it continues to learn and improve from on-the-job experience .
Under the hood, Dyna's robots work up to sixteen hours a day in hotels, restaurants, laundromats, and gyms , handling tasks like folding and food preparation. Pricing follows a robots-as-a-service structure, where hardware, software, maintenance, and continuous model updates are bundled into a single monthly payment, reducing ROI uncertainty for customers while giving Dyna recurring revenue and each deployment generates training data that enhances the foundation model for all customers . The Din Tai Fung milestone is the first public proof that this loop closes on a customer P&L.
Comments (15)
Sign in to join the discussion.
K
Kwabena O.Aug 28, 2026
Dumplings are the ultimate stress test. If the robot can pleat without crying, it's already ahead of half my coworkers.
O
Olamide ShawAug 28, 2026
A robot cannot tip the busser. The physical economy will notice.
M
Marisol VegaAug 28, 2026
The cut where the robot arm syncs with the steamer lid opening is chef's kiss editing. Whoever cut this trailer deserves the xiao long bao.
Y
yukiAug 28, 2026
Voiceover pacing on this launch is unusually calm for a robotics reveal. No swelling synth build, no drone shot of a warehouse. Refreshing.
D
Dev ParikhAug 28, 2026
'Crossed the ROI threshold' is doing a lot of heavy lifting in that tweet. Show me hours per station per shift or I'm assuming it means one location on a Tuesday.
N
Nikhil R.Aug 28, 2026
I've watched three robot-in-restaurant videos this year and all of them cut away right before the plating. What happens after the pleat?
L
Linnea LundqvistAug 28, 2026
Every robotics launch video is shot at 60fps in golden hour and I refuse to believe any of these arms move that smoothly at 2am on a Saturday rush.
T
Toshiro K.Aug 28, 2026
Curious what the failover looks like when a dumpling skin tears mid-pleat. That's the demo I actually want to see.
A
Amara UdoAug 28, 2026
Din Tai Fung as a design partner is a flex. Their QC standards are brutal, so if the robot survived that kitchen it survived anything.
P
Petros M.Aug 28, 2026
Robotics companies love the phrase 'real deployment' the way crypto folks loved 'real utility' in 2021. Show me the third location.
B
Benoit CarrèreAug 28, 2026
Reminds me of what Physical Intelligence and Skild are chasing but with an actual customer instead of a research paper. Bullish on anyone who ships to a kitchen.
R
Rashida KhatriAug 28, 2026
Before I get excited: what's the story on food-contact compliance and how does the customer handle liability when the arm fumbles an order? Not rhetorical.
C
Chen YuAug 28, 2026
Every founder tweet in robotics starts with 'after so many demos' now. It's the new 'we've been quietly building.'
S
sveaAug 28, 2026
The tweet threading is smart, hook then customer then proof. Way better structure than the usual 12-tweet manifesto robotics folks post at 6am.
I
Ingrid V.Aug 28, 2026
The bar for robotics launches used to be 'look, it moved.' Now it's 'look, someone paid for it.' Progress feels slow until it isn't.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
Dyna Robotics
General and robust AI robots that power the future of the physical economy.
General and robust AI robots that power the future of the physical economy.
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
Dyna Robotics builds general-purpose robotic arm systems powered by its own foundation models, aimed at businesses that need repetitive dexterous work done reliably. This launch marks the moment its robots crossed the ROI threshold in a paying customer deployment at Din Tai Fung, one of the highest revenue per location restaurant chains, moving the company past the pilot stage that has stalled much of the humanoid and manipulation field. For founders and operators evaluating physical AI, the significance is that a fleet is now covering its cost inside a real, high-throughput kitchen rather than a staged demo. The company was founded in 2024 by Lindon Gao, Jason Ma and York Yang and is based in Redwood City. Lindon and York are repeat founders who previously built Caper AI, an AI-powered smart cart company which they sold to Instacart for $350M. Jason Ma is a former researcher at DeepMind and NVIDIA . Their approach centers on task-specialized foundation models trained and refined in production, with the model deployed at multiple customer sites, supporting generalization and commercial viability as it continues to learn and improve from on-the-job experience . Under the hood, Dyna's robots work up to sixteen hours a day in hotels, restaurants, laundromats, and gyms , handling tasks like folding and food preparation. Pricing follows a robots-as-a-service structure, where hardware, software, maintenance, and continuous model updates are bundled into a single monthly payment, reducing ROI uncertainty for customers while giving Dyna recurring revenue and each deployment generates training data that enhances the foundation model for all customers . The Din Tai Fung milestone is the first public proof that this loop closes on a customer P&L.
Comments (15)
Dumplings are the ultimate stress test. If the robot can pleat without crying, it's already ahead of half my coworkers.
A robot cannot tip the busser. The physical economy will notice.
The cut where the robot arm syncs with the steamer lid opening is chef's kiss editing. Whoever cut this trailer deserves the xiao long bao.
Voiceover pacing on this launch is unusually calm for a robotics reveal. No swelling synth build, no drone shot of a warehouse. Refreshing.
'Crossed the ROI threshold' is doing a lot of heavy lifting in that tweet. Show me hours per station per shift or I'm assuming it means one location on a Tuesday.
I've watched three robot-in-restaurant videos this year and all of them cut away right before the plating. What happens after the pleat?
Every robotics launch video is shot at 60fps in golden hour and I refuse to believe any of these arms move that smoothly at 2am on a Saturday rush.
Curious what the failover looks like when a dumpling skin tears mid-pleat. That's the demo I actually want to see.
Din Tai Fung as a design partner is a flex. Their QC standards are brutal, so if the robot survived that kitchen it survived anything.
Robotics companies love the phrase 'real deployment' the way crypto folks loved 'real utility' in 2021. Show me the third location.
Reminds me of what Physical Intelligence and Skild are chasing but with an actual customer instead of a research paper. Bullish on anyone who ships to a kitchen.
Before I get excited: what's the story on food-contact compliance and how does the customer handle liability when the arm fumbles an order? Not rhetorical.
Every founder tweet in robotics starts with 'after so many demos' now. It's the new 'we've been quietly building.'
The tweet threading is smart, hook then customer then proof. Way better structure than the usual 12-tweet manifesto robotics folks post at 6am.
The bar for robotics launches used to be 'look, it moved.' Now it's 'look, someone paid for it.' Progress feels slow until it isn't.