Motionlayer

Tinymovr is now MotionLayer. After 8 years building motor controllers for robots across 30+ countries, we're expanding our focus. Same team, same hardware—now building motion infrastructure for embodied AI. First release: ML-101, an open-source leader arm for VLA research. More soon.

Physical AI sounds like one big idea. But where are the real opportunities? I’ve seen many ways to break down Physical AI. I personally care more about where the opportunities actually land in the market. The breakdown below feels the most straightforward to me. The four opportunity layers: 1️⃣ Hardware - Robot bodies / embodied systems - Module-level components - Embedded compute & edge systems 👉 This layer sells physical capability. 2️⃣ Software - Robotics / Physical AI models - Capability software - Runtime & operations 👉 This layer sells intelligence and system reliability. 3️⃣ Hardware + Software - Vertical solutions - System integration & deployment - Repeatable systems 👉 This layer sells real-world outcomes. 4️⃣ Data - Training data - Operational / deployment data - Evaluation & failure data 👉 This layer sells feedback loops and long-term advantage. Curious how others break this down — where does your work fit, or do you see a better way to classify the space?

Great to see Tinymovr's ROS 2 motor driver included in this week's robotics roundup by Open Robotics! Lots of exciting developments - from iRobot's benchmarking suite to new simulation tools and Rust for ROS 2 Check out the full list 👇

ROS 2 support is here 🚀 We've just released official ROS 2 integration for Tinymovr motor controllers. The robotics community has been asking for this, and we delivered. Whether you're building research platforms, mobile robots, or manipulators – Tinymovr now plugs directly into your ROS 2 stack. Check out the repo: https://lnkd.in/deui2x34 Building the motion infrastructure for the next generation of robots, one commit at a time. #ROS2 #ROS #Robotics #OpenSource #MotorControl #MotionControl #PMSM #BLDC

We've just released Tinymovr 2.3.4 - our first firmware release since like forever 🫣. This release consolidates several months of improvements including a firmware fix for calibration reliability, Studio enhancements (conditional PID gains, better logging, Python 3.12 support), and comprehensive technical documentation covering architecture, safety constraints, and development workflows. All board revisions supported (R3.x, R5.x, M5.1/M5.2). Release notes: https://lnkd.in/dMRVBxzx #Robotics #MotorControl #OpenSource #Embedded #FOC #BLDC #MotionControl

Thanks to Microchip Makes for the feature! The MCP25625 is the perfect fit for bringing robust CAN communication to Arduino Nano projects. #Arduino #CANBus #EmbeddedSystems #Microchip

We're launching 2026 with the first and only CAN Bus shield designed specifically for the Arduino Nano. No more breakout boards and jumper wires—this mounts directly to the Nano. Built around the MCP25625 with an onboard boost converter for 3.3V microcontrollers. Physical switch for 120Ω termination. Works with standard MCP2515 libraries. €26, ready to ship. Product details: https://lnkd.in/di2bPmzk GitHub: https://lnkd.in/dG9AEj2E More updates coming soon. 2026 is going to be an interesting year for motion control and robotics. #Arduino #CANBus #Robotics #EmbeddedSystems #HardwareDesign #MotionControl

WELCOME 2026!!!! 🥳 To humanoid or not to humanoid, that is the question. Humanoids are currently the hottest topic in robotics. No question about it. What to pick: a fancy biped humanoid or a specialized mobile manipulator for a specific use case or task? This post is not intended to criticize humanoids. 🚫 I'm looking for applications where I'll say 'well, a conveyor belt and a 6-axis robot won't work here' or 'aha, that's where humanoids belong'. Some more challenging points to consider: → Wheels are consistently more efficient than legs in most scenarios. Many environments, including those designed for consumers, are better suited to wheeled systems. → When weighing cost against benefit, wheeled robots can deliver 80% of the functionality of a humanoid robot at just 20% of the cost. → General-purpose robotics does not necessitate humanoid designs. AI-powered robots can be versatile and effective without adopting a humanoid form factor. → Safety is a significant challenge with legged locomotion. If a humanoid robot were to fall, it could pose serious risks to people nearby, especially children. This concern is far less pronounced with wheeled robots that have a stable base. What is the ultimate killer application for humanoids? 🦿 P.S. The market is developing so fast that I have to ask this question once in a while. ~~ ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Happy holidays from Tinymovr! As 2025 wraps up, we're grateful for our community and excited about what we're building for 2026. See you in the new year!