One of the bottlenecks in training robot policies in simulation is the sim-to-real gap. We are working on something that addresses the fundamentals of physics and here is a sneak peek and how it looks in Isaac Sim from @NVIDIARobotics

What matters in Physical AI? This webinar co-organised by Bota and Kinova highlights the most important factors for every step of the workflow towards automation with AI.

Is haptic teleoperation the most promising method to collect the right data for Physical AI? We don’t have an answer to this question but we know it works and is the most cost effective solution at the moment to program a robot! The Leader feels exactly what the Follower is feeling, and this is be transferred to a Physical AI. Complex trajectories over unstructured environments become much easier to program. We will be talking about this and other data collection methods and their future over the upcoming webinar with on 20th of May. The spots are slowly filling up. Make sure you subscribe ASAP so we prepare the ground. kinovarobotics.com/resource/… At Bota Systems AG we enable teleoperation, kinesthetic teaching, handheld devices and more with the most reliable force sensing solutions.

We are going to have a breakthrough soon and it will be enabled by the sense of touch. Our technology is developed with endurance in mind, maintaining the performance required to get the highest quality data.

In the upcoming webinar we will explore the “secret sauce” behind Physical AI. There is still time to register. Either for research or deployment, constistency and quality are keys. The upcoming webinar with Kinova is on 20th of May at 11:00 AM (EST). kinovarobotics.com/resource/…

How do we solve the data gap in robotics? Come listen to the ones driving the conversation, including Bota Systems Co-founder and CEO Klajd Lika Join us on May 6th at 1:30pm. Only a few spots left - link in the comments.  📍 Zürich, Switzerland

Possible with the sense of touch

Force data will play a critical role across every stage of Physical AI! • Demonstration – Recording episodes becomes more intuitive, safer, and significantly more precise. • Training – Models learn not just motion, but interaction—capturing the nuances of contact through force signals. • Inference – With real-time force sensing, models can achieve near-perfect success rates, even in complex, contact-rich tasks. In the video below, we demonstrate this by training a simple model that leverages force data end-to-end. As manufacturing evolves, Physical AI will accelerate the adoption of robotics. The key question isn’t if this shift will happen—but how it scales across each stage and what the underlying economics will look like. That’s where the real transformation begins. At Bota Systems AG we believe that anything humans can coordinate with their eyes and control with their hands can be replicated by an AI-driven robot — provided it can feel the world as we do.

“Don’t skip force sensors” - Fully aligned here at Bota. Thanks for sharing insights from the technical journey of @Tesla_Optimus , @K0nstantin0s_ . It was a great talk addressing all technical challenges and the way forward for humanoids that make sense. An artist of motor optimisation and one of the few of his kind. Happy to see National Technical University of Athens alumnies like Konstantinos still sharpening their engineering excellence throughout the biggest technology companies worldwide and making the impossible - possible. @Tesla might be the only company that has all required resources, infrastructure, know how, and momentum for humanoid technology to make financial sense. A technology that might be able to match or surpass the impact level of Ford and Toyota in manufacturing. At Bota we enable humanoids and other robotic systems to interact with the maximum precision, safety and usability in the era of Physical AI. Many thanks to @ethroboticsclub for organizing this Inspire Talk.

@BotaSystems is 10x Faster at Real-World Data Collection Force sensitivity is the key to making robots truly collaborative and capable. To push the boundaries, we’ve been working on fine-tuning a VLA model that changes the game for data efficiency: botasys.com/physical-ai

Better together. 🤝 Starting from Athens last Christmas we did the same in Zurich. This time the crowd was massive. More than 400 attendees and many from around the world 🌍 joined us. Organizing yesterday’s Zurich Robotics Meetup with my friend @_dario_b_ and TheRoboticsClub was a highlight of February. We started this with a simple goal: to bring the local robotics community closer together. It became bigger than we expected 😅 It wouldn’t have been possible without our entertaining guests Marco Hutter from ETH Zürich, Edo Jelavić from @GravisRobotics , Eris Dhionis Sako from @Duatic , Giorgio Valsecchi from @RIVR, Manuel Yves Galliker , Dario Bellicoso from RAI Institute , and myself (😂) from @BotaSystems . @leggedrobotics Lab Alumni sent a message to how a strong robotics community looks like. Seeing the deep-dive technical conversations and the new connections made over drinks made all the planning worth it. Stay tuned and see how the Zurich Robotics community will create an impact and add socioeconomic value. If you missed this one, don't worry—we’re just getting started. Swipe left to see some highlights from the night! 📸

Checkout how you can use our New Robot Joint Torque Sensor to enhance your robot with safety and precision. botasys.com/torque-sensors

Zurich in February will be covered in Roboticists! 250 RSVPs! 14 days to go! Don’t miss it! Still spots available Subscribe here: luma.com/oouk1ecu

We come from research first and foremost and we want to give back to fellow researchers; robotics is being reimagined with AI and modern tools and approaches are pushing the industry forward unlocking unpredented reliability, precision, dexterity and ease of use. Again starting from research; our platform has been enabling researchers from Stanford, Berkeley, Carnegie Mellon, UCLA, Cornell, UTA, Texas A&M University, ETH Zurich, EPFL, Delft, Tsinghua University, Peking University, Zhejiang University, Harbin Institute of Technology, and many many more to explore what’s possible with modern robotics; We’re excited to expand our research program, enabling researchers from similar labs to use the latest and greatest of our products, including MiniONE Gen A, for research purposes; if you are a researcher that contributes on AI for robotics, you are building AI models with the sense of touch or you plan to do as of TOMORROW, you are working for a university and would acknowledge Bota contribution in dissemination then apply here to get access to our products: botasys.com/physical-ai

Tactile Sensing based on MiniONE FT sensors with: - minimized creep, - temperature dependency, - degradation of the sensing element, - and a costless elastic touching surface made with 3d printers designed to your shape of preference

The @UnitreeRobotics G1 model is now equipped with an additional sense of touch, reaching unprecedented manipulation capabilities.

Many were asking if our force controllers and Kinova Gen 3 could go faster and match human speeds. It turns out that @Klajd_Lika couldn’t go faster, but the robot could.

1x speed Superhuman dexterity We developed interaction intelligence based on physics, control theory, and force sensing. Closing the loop at 1 millisecond is more than enough for humanlike physical interaction experience. Now we can teach it by hand guidance some Primitive Robotic Skills that run at lower frequencies like: - grasping, - no-slip control, - mobility, - contact-based actions, and - continuous contact manipulation. In the near future we will climb a layer up and teach it some combined skills. These can be still be considered equivalent to “machine language” in CS. Later on, we can teach it a higher level of commands that involve the world and interpretation of human commands. This we can also leave for CS people that are experts at. ;) Stay tuned.

Well... this sensing is coming to stay. For whole-body manipulation, for locomotion, for wrench estimation, and for any robot actuator that has strain-wave, planetary or any other speed reducer, our torque sensors will be a gold standard that will be produced at scale. It will come at a cost that make sense without sacrificing a tiny little bit of performance. Stay tuned at botasys.com

We’re excited to announce that our ecosystem supports the @FRANKAROBOTICS FR3, delivering full 6-DOF force feedback for advanced teleoperation and kinesthetic teaching through the SensONE Gen-A FT Sensor. Learn more in our product page: botasys.com/robot-accessorie…