9 Notable Advancements in Robotics that We Can't Miss

Robotics seem to be everywhere now. And this week Twitter was full of popping-up info about new robots and new features in robotics. We couldn’t miss this news stream and gathered for you the most interesting pieces.

So let’s discover 9 quiet recent notable updates in robotics:

1. Boston Dynamic’s Atlas as enterprise-grade robot:

   Now Atlas is officially turning into a real product, designed to work day-to-day in factories and warehouses. Built with Hyundai as its first customer, and now with Google DeepMind-powered learning and fleet management via Orbit, Atlas is made for scale, safety, and ROI. It can lift heavy loads, share space with humans, swap its own batteries, and integrate directly with MES/WMS systems.→ Read more

2. Siemens and NVIDIA accelerate the next wave of automation with physical AI:

   This partnership is embedding AI across the entire industrial lifecycle, from design and simulation to manufacturing and operations. By combining Siemens’ industrial software with NVIDIA’s AI and simulation stack, they aim to make digital twins adaptive and operational in real time. → Read more

3. RoboPocket from Neomatrix turns an IPhone into a portable data collection system for embodied AI, combining LiDAR, IMU, and vision to capture spatially accurate, real-world interaction data. It provides real-time feedback on data quality and supports multi-device synchronization → Read more

4. Aletta is a fully autonomous robot that does the entire blood draw end-to-end. It finds a vein with infrared + ultrasound, inserts the needle with sub-millimeter precision, handles the tubes, and applies the bandage. One supervisor is able to oversee up to three devices in busy clinics. Already CE-certified in the EU, it is now being used in clinical trials to push labs toward fully automated workflows → Read more

5. IRON from XPeng:

   Applies XPeng autonomous driving stack – perception, planning, and control – to a human-shaped robot. IRON is positioned as a scalable platform for real-world tasks. Interesting features are full-body perception, coordinated bimanual manipulation, and whole-body motion control, plus IRON learns to walk like humans → Read more

6. NVIDIA, in general, expends its robotics stack with open Physical AI models like Cosmos Transfer, Cosmos Predict, and GR00T (a Vision-Language-Action model), alongside Isaac Lab-Arena for large-scale simulation. These tools aim to reduce task-specific scripting and accelerate general-purpose robot learning. → Read more

7. VIGX π6 AI exoskeleton:

   Adds real-time terrain perception VIGX unveiled in the π6 AI-powered exoskeleton that uses onboard cameras and AI to analyze upcoming terrain before each step, adjusting assistance on the fly. This makes wearable robotics more adaptive and responsive for outdoor and mixed terrain use→ Read more

8. UBTECH Walker S2 robot keeps multiple rewards separate during learning, preventing them from collapsing, and helping models learn user preferences more accurately and stably. It also introduces a newcoming feature: this robot can swap its battery autonomously → Read more

9. DEEPRobotics’ LYNX M20:

   An industrial wheeled-legged robot designed for hazardous and hard-to-reach environments, combining the speed of wheels with the agility of legs. It autonomously adapts its gait to terrain, operates from –20 °C to 55 °C with IP66 protection, and handles tasks like inspection, emergency response, and logistics across rubble, stairs, tunnels, and uneven ground. → Read more