RoboHubđ€
@XRoboHub
Worldâs largest AI Ă robotics channel. Direct access to 200+ robotics companies. Original, tech-grounded reporting. 500K+ followers. đŹ xrobohub@gmail.com
1.1K Following 22.2K Followers
These robots really bring the stage to life! Honestly, they're even more eye-catching than the human backup dancers.
AGIBOT's X2 robots joined the performers at the Greater Bay Area Film Concert, dancing in sync.
Wouldn't this be a lot less fun to watch without them?
Show more
This will get you hyped... itâs a real humanoid robot band. đ€đž
CASBOT BAND was easily one of my favorite things at WRC. Four full-size CASBOT 02 humanoids actually play guitar, bass, keyboard and drums, staying in sync while performing with a human lead singer.
If I ever open a bar, Iâm absolutely getting a robot band. Then no matter how badly I sing, Iâll always have a band backing me up. lol
Show more
Anyone heading to Pittsburgh for IROS?
WUJI Salon | Pittsburgh · #IROS2026#
Toward an Embodied Future
Bringing together robotics researchers, developers and industry partners for a CEO keynote, guest talks, dinner and networking.
Sept 29 · 6:30â9:00 PM EDT
Scan the QR code to register.
#EmbodiedAI# #Robotics# #WUJI# #WUJIHand#
#DexterousHand#
Show more
She looks stunning⊠U1 at Beijing Fashion Week.
UBTECH CEO Zhou Jian just shared this clip from a customer event. U1 series deliveries are already underway.
Is this the robot girlfriend or boyfriend youâve been waiting for?
UBTECH has officially launched UWORLD, its consumer full-size hyper-realistic humanoid robot series.
Pricing starts at RMB 119,800, about $17.6K, for the lighter half-body U1 Lite.
U1 Pro is RMB 169,800, about $25K.
U1 Ultra goes much higher: RMB 990,000, about $146K, for the male version, and RMB 880,000, about $130K, for the female version.
UBTECH says the first-generation bionic robots are built around âemotional value,â mainly for single users and seniors over 60 who need companionship and care.
The goal is more natural human-robot interaction, emotional companionship, and basic life assistance.
They also showed more than 50 full-size hyper-realistic humanoid designs, with different male and female looks, body types, and heights from 1.60 m to 1.85 m.
Thatâs the bigger signal here: humanoid robots are starting to move from one standard body into personalized consumer products.
U1 is already open for pre-order on JD, with more than 13,361 orders so far.
UBTECH says it will push mass production and delivery, aiming to complete those 10,000+ orders this year.
Consumer humanoid robots just entered a real market test.
Show more
Tesla is moving Optimus toward mass production, with orders already going to Chinese suppliers.
Chinaâs 21st Century Business Herald reports that Tesla began a new round of supplier audits on September 17, after its team arrived in Ningbo. The reported work includes checking production consistency and helping suppliers set up equipment.
Earlier audits covered joint modules, precision parts and assembly quality, using automotive-style requirements for production speed and yield.
July supply-chain reporting described parts orders covering hundreds of Gen3 robots and supplier expansion plans based on 100,000 robots.
Suppliers are expanding capacity and developing new production processes. Meanwhile, component costs are being broken down against an eventual selling-price target of around $20,000, making faster production and fewer defective parts central to the ramp.
The production plan calls for converting Fremontâs Model S/X lines for Optimus, targeting a start by the end of 2026 and eventual capacity of 1 million robots a year. Initial deployments are intended for Teslaâs own factories.
Show more
Figure AI just unveiled Helix 2.5 to get humanoids working in everyday homes.
Humanoids have largely relied on training for specific environments. Helix 2.5âs biggest advance is zero-shot generalization: using what it already knows to navigate unfamiliar rooms and handle new objects, without retraining.
Figure rented 30 Bay Area homes the robot had never seen, where Figure 03 autonomously tidied living rooms, made beds and folded towels. Figure calls this a first for humanoids at this scale.
Behind it is Index, a global program turning human experience into robot training data, with more than 90,000 contributors each week. In a controlled test, Index pretraining alone pushed zero-shot task success from 9% to 56%.
More human data also brought predictable gains in robot-action prediction, suggesting humanoids could scale like LLMs.
Robot learning still has hard problems to solve, but signs of general physical intelligence are emerging. Figure has committed $3.5B in compute to scale up Helix training.
Humanoids working in everyday homes no longer seem so far off. Chinese media report that Teslaâs Optimus team has also begun supplier audits in China, with production orders already placed.
Show more
The holy grail for robotics is being able to generalize: doing work in unseen places
We rented 30 homes in the Bay Area and are doing tasks without any new training
Japanâs ugo just unveiled Nova, a wheeled dual-arm robot with mass production planned for 2027.
Nova has a 22-DOF humanoid upper body on mecanum wheels, with a 4 kg payload per arm and up to 10 kg using both. Runtime is about six hours.
Operators can record demonstrations using force-feedback or VR controllers for AI training. Trained models can run onboard on NVIDIA Jetson Thor.
ugo is also offering help with what comes after buying the robot: collecting task data and adapting models for a factoryâs handling, picking and assembly work.
Show more
This is too funny⊠did Thing from The Addams Family get a twin? đ
ETH Zurichâs team taught a WUJI robotic hand to crawl around and get back up when it tips over, and yes, it can use a keyboard too.
Show more
I taught a robotic hand to walk on its fingertips.
Usually an arm takes the hand to the work. We wanted it to get there without one, then use the same fingers to do the job.
Meet Fingers as Legs.
Video:
Paper:
Show more
Congrats! Check out Erenâs robot shop.
Iâm excited to announce that my online shop is now live.
A curated collection of robotic hardware, now available with simple, all-in pricing.
No hidden costs. No extra domestic shipping fees.
Orders ship domestically from the U.S. with short lead times.
Explore the shop:
Show more
All three of these companies are using dexterous gloves of various sorts for high quality data collection
Agility just unveiled Digit 5. Still green, but it's looking a little more Atlas these days.
It's designed to work alongside people without safety fences. It detects nearby people and can avoid them, stop or sit down, backed by an independent safety controller.
Payload rises 40% to 50 lb (22.7 kg). A 9-minute recharge supports 90 minutes of work. Swappable grippers are part of the plan to expand beyond hauling totes into machine tending, inspection and palletizing.
Digit 4 has already logged 65,000+ operating hours, and Agility reports $300M+ in multiyear orders for Digit 5.
Show more
Introducing Digit 5.
Agilityâs next-generation humanoid, engineered for cooperatively safe work at scale, allowing it to work in close proximity to people without the physical safety barriers required by traditional automation.
Explore Digit 5:
Watch the full video:
Show more
More robotics teams are getting behind the same idea: let robots learn straight from people.
Sundayâs ACT-1 learned long-horizon household tasks from human demonstrations. X Squareâs TwinDEX used a few hundred robot-free episodes to train a policy that powered a 24-step chemistry experiment.
Now @RewardAI_âs OM-1 learns manipulation from human demonstrations alone. The same policy runs across industrial arms and humanoids, handling contact-rich tasks that require coordination, force control and recovery when things go wrong.
Reward says OM-1 can pick up a new task from less than 30 minutes of human demonstration data.
Show more
Introducing OM-1, our first robot foundation model, zero-shot generalizing to any robot: table-top arms, industrial arms and humanoids.
- learned directly from human manipulation data
- no teleop/robot data
- close to human-level dexterity and efficiency
- multi-robot collab
Show more
Okay, the speakerphone moment got me. đ
Jensen Huang took Trumpâs call onstage at All-In and got roasted for his phone skills.
He held a mic to his phone: âMr. President, youâre now talking to the planet.â
A few minutes later, Trump: âI have no idea who the hell Iâm talking to.â lol
Show more
MUST SEE: Amazing moment as President Trump calls Nvidia CEO Jensen Huang while he's on stage at the All-In Summit.
@POTUS on AI Doomerism: âI'm telling you, it's all a hoax⊠and we're not going to let that happen.â
Show more
RMB 9,800 (about $1,460) to get started.
BLUE WORM just unveiled Mantis Standard. Detach an arm for your desk, build a dual-arm workstation, or add the mobile base and lift for a wheeled humanoid.
The arms, base, lift and head use mechanical and electrical quick-connects. Change the configuration, keep the same robot brain: BW-Brain.
The full 1.4 m build has 22 degrees of freedom, a rated 7 kg payload per arm and 550 mm of vertical lift.
It supports ROS 2, Isaac Sim and LeRobot, with compatibility for Ï0.5, ACT and SmolVLA.
You can simulate a task, collect VR demonstrations, fine-tune a policy and deploy it on the hardware. For a small lab, reusing those components and development tools could mean fewer separate rigs to buy and less integration work.
Show more
A humanoid every 10 minutes. UBTECHâs new Liuzhou factory is up and running, with a line designed for that pace and planned annual capacity above 10,000 robots.
Last December, this site celebrated its 1,000th Walker S2. Less than nine months later, itâs gearing up for five-figure annual production.
UBTECH reported 921 full-size humanoids sold in the first half of 2026. On September 10, it announced over RMB 50M (about $7.5M) in recent overseas orders from Europe, Japan and South Korea, including Walker C1 and U1 robots.
Show more
UBTECH Hits 1,000 Walker S2 Units, Eyes 10,000 by 2026 đ€đ
Just as its 1,000th Walker S2 rolled off the Liuzhou production line, UBTECH announced two new major wins, totaling over RMB 130 million (~$18M USD).
This includes contracts for a data acquisition center in Huizhou (RMB 59.62 million) and an embodied intelligent technology center in Hohhot (RMB 77.80 million).
This pushes UBTECHâs total humanoid robot orders for 2025 close to RMB 1.4 billion (~$193M USD). They are rapidly scaling up, with plans to reach an annual production capacity of 10,000 units by 2026.
Source: UBTECH
Show more
Thereâs been a lot of talk lately about whether roboticsâ GPT moment could be GPT itself.
Astra can already use tools to control robots, build simulations and write control code. That still seems a long way from a GPT moment, but AI is already taking on parts of a robotics engineerâs job.
On @genrobotics_aiâs GRID, agents determined that a stirring task needed human demonstrations, requested the data and fine-tuned a policy. When perception couldnât keep up with a moving beaker during pouring, they fine-tuned and integrated a faster tracker.
Those integrations, models and tested fixes stay available for the next task. The team reports about four hours to get the first skill working on a fresh Flexiv setup, with later skills on that same setup deployable in as little as 10â15 minutes.
So I donât think a stronger GPT automatically replaces embodied AI models. The teams building them can use it to train, test and deploy faster. That could help bring the breakthrough closer.
Show more
What if a robot could take a goal and determine how to achieve it?
Today, weâre introducing Auto-Engineering on GRID.
Robotics has more models, simulators, and hardware than ever. Turning them into working deployments still takes specialized engineering.
For two years, weâve been building that robotics know-how into unified intelligence GRID - spanning robots, skills, models, and proven approaches.
Auto-Engineering puts it to work by completely automating how robot intelligence is built, tested, and deployed.
Show more
Skild just crossed a $100M annual revenue run rate, ten months after its first commercial deployment.
It now has 60+ paying customers and is working with NVIDIA and Foxconn to deploy robot arms for Blackwell system assembly.
Those deployments helped shape S1. Customers keep changing parts, layouts and workflows. Having to collect data and retrain every time gets expensive fast.
S1 lets operators teach a new task with a video prompt, without updating the modelâs weights. It addresses a practical problem: keeping robots useful as the work changes.
Skildâs bigger bet is that experience from different customers can improve the general model, so future deployments need less customization. The business grows by deploying robots; those deployments help build the next version of the brain.
Congrats to @deepakpathak and the whole team!
Show more
This is the first IRON robot coming off XPENG Roboticsâ new production line.
XPENG just started building humanoid robots like cars. And yesâit walked off by itself. The line automates 80%+ of core processes and uses automotive-grade quality control.
Mass production is set for year-end, with stores and campuses first. China and overseas deliveries follow in 2027.
$900M+ raised at a $6.3B+ valuation just two weeks ago. Now the production line is running. XPENGâs humanoid push suddenly looks a lot less like a side project.
Show more
$900M round. $6.3B valuation. XPeng Robotics just set a new funding record in Chinaâs embodied AI race.
The round was led by IDG Capital, with Tencent, Alibaba, and Gaorong joining. XPeng retains control.
The money goes straight into scaling IRON and its Physical AI stack: hardware, data, mass production, and commercial sales in China and overseas slated for 2027.
More important than the funding itself is where XPeng is starting from.
Like Tesla, it already spent years building in automotiveâAI chips, world models, manufacturing, and a mature supply chain. A lot of that can carry straight into robotics.
Itâs easy to look at a $6.3B valuation and say itâs too high. But investors just put more than $900M behind this bet.
Theyâre not only betting on the robots XPeng has today, but on where the entire robotics industry is headingâand whether XPeng can turn its automotive engineering into a scalable, real-world Physical AI business.
And yes, $6.3B this early is huge for a robotics business. But Unitree just went public and closed its first day at roughly a $50B market cap.
Robotics is still incredibly early. Morgan Stanley sees humanoids alone becoming a $5T market by 2050âpotentially twice the size of the auto industry.
If that direction plays out, I donât think robotics ends with one trillion-dollar company. There could be many.
Show more
This is wild⊠humanoid robots can actually fight on their own now! đ„
Unitreeâs UnifoLM-X2-1.0 reads the opponent, predicts the next move and reacts instantly. Itâs the first time a world model has controlled a fully autonomous humanoid fight in real time.
Robot fighting is brutal. The other robot keeps moving and hits back, leaving almost no time to decide what to do next.
Show more
Unitree Breakthrough: The Worldâs First Real-Time World Model-Driven Fully Autonomous Humanoid Robot Combatđ„
UnifoLM-X2-1.0 breaks through world-action foundation models' bottlenecks in instant planning, decision-making, and dynamic interactive execution, achieve high dynamics, strong interaction, real-time prediction and planning of the future, achieve fully autonomous combat for humanoid robots. This validates the fundamental feasibility of large-scale deployment of world model-driven humanoid robots.
Show more
One robot hand. Four gripping modes. No tool changes.
Leju Robotics and HeymanDex teamed up on Dasheng 1 (SG100), a reconfigurable hand that switches between dexterous, long-stroke, four-finger and opposing-finger modes.
It has 11 active DoF, â„24 kg (53 lb) of hook-grip capacity per hand and â€0.03 mm repeatability. In the demo, the same hands handle tiny parts, soft packages, irregular components and bins.
For factories, that means one end effector and one control model across far more SKUs. Basically a Swiss Army knife for industrial robots.
Show more
The first 10,000-unit humanoid robot production line is now live â built by Leju Robotics and Dongfang Precision. đ€
One robot every 30 minutes, fully digital, fully tested before it ships.
This is what scaling actually looks like.
Show more
