The Role of Data in Robotics Transformation

Robotics is no longer about hardware — it’s about who owns the data. When RealMan Robotics opened its data training center in Beijing, it wasn’t just creating another research hub — it was signaling the industry’s shift from hardware-driven innovation to data-driven performance. For years, robotics progress centered on building faster autonomous mobile robots (AMRs), stronger arms, and more precise sensors. But performance is hitting diminishing returns. The real bottleneck? Training data. Robots generate massive sensory streams, but without curated datasets, algorithms don’t improve. That’s why structured data pipelines have become the “conveyor belts” of the AI era. We’re seeing this across industries and geographies: • ROVR Network’s open dataset is democratizing access, much like ImageNet did for computer vision. • NVIDIA and Qualcomm are embedding compute at the edge, ensuring robots learn from local feedback in real time. • Siemens’ Shanghai plant reported a 350% efficiency leap — not from better robots, but from data-driven orchestration across fleets. • John Deere’s acquisition of GUSS Automation signals that agricultural robotics is following the same path: the value is in data loops, not just machines. The strategic play is clear: robotics differentiation is moving from hardware to data-driven adaptability. For operators, this means three things: • Data becomes the moat. Running robots today train the models that will dominate tomorrow. • Partnerships will shift. Expect joint ventures between OEMs, logistics providers, and manufacturers to pool datasets. • Competitive advantage compounds. Companies that build proprietary feedback loops will accelerate past those who just buy off-the-shelf robots. The critical question for leaders: are you just automating tasks—or building the dataset that future-proofs your operation? I put together the most comprehensive weekly newsletter on automation, robotics and innovation driving intralogistics, supply chains, and e-commerce. ↓ https://lnkd.in/evrCrS2P

Is data going to solve robotics? Ken Goldberg asked us to ponder this at ICRA 2025 Keynote this year alongside Daniela Rus, Russ Tedrake, Aude G. Billard, Leslie Kaelbling and Frank Park. Turns out that this is a surprisingly polarizing, especially in robotics. As I built my arguments for the debate, I have refined the ideas in 4 key lessons. 1. Too much structure hurts! We have seen it time and again that engineered inductive biases may not be competitive as data & compute scale. The argument to building a complex solution from a model-first approach suffers that same challenge of picking the correct hypothesis space! 2. Data helps with Ambiguity & Robustness The real world is far-too ambiguous to specify in terms of clean objective! Moreover, achieving generalization in terms of robustness to variability even within task-family has long been a challenge to model-based methods. 3. Data leads to a Unifying Perspective! Robotics has been a community of communities! Despite the nuances, we are still excited by the same problem - enabling intelligent behavior in a physical environment. Yet robotics community has deep fissures across subareas. Soft-inductive biases as well as newer AI perspectives have united many communities across Sciences. Roboticists can turn over a new leaf by adopting learning based tools to unite the variety of problems and representations across the domain. 4. Data-First ⇏ Lack of Modularity Adopting a perspective that scales Data/Compute does not eschew modular abstractions. The fabric of computing has been built on abstractions. Robotics is the next generation of computing. There is no reason to believe that the abstractions will be discarded altogether, yet a new paradigm is needed as each layer is now guided by a learning based technique. Scaling Data appears to be necessary, but insufficient to "solve" robotics problems! Yet, data will take us very far in our quest for intelligence, and then we will turn scientists to "study" our creation to understand and extend it Slides: https://lnkd.in/ePp9z2U3

1. Scan 2. Demo 3. Track 4. Render 5. Train models 6. Deploy What if robots could learn new tasks from just a smartphone scan and a single human demonstration, without needing physical robots or complex simulations? [⚡Join 2400+ Robotics enthusiasts - https://lnkd.in/dYxB9iCh] A paper by Justin Yu, Letian (Max) Fu, Huang Huang, Karim El-Refai, Rares Andrei Ambrus, Richard Cheng, Muhammad Zubair Irshad, and Ken Goldberg from the University of California, Berkeley and Toyota Research Institute Introduces a scalable approach for generating robot training data without dynamics simulation or robot hardware. "Real2Render2Real: Scaling Robot Data Without Dynamics Simulation or Robot Hardware" • Utilises a smartphone-captured object scan and a single human demonstration video as inputs • Reconstructs detailed 3D object geometry and tracks 6-DoF object motion using 3D Gaussian Splatting • Synthesises thousands of high-fidelity, robot-agnostic demonstrations through photorealistic rendering and inverse kinematics • Generates data compatible with vision-language-action models and imitation learning policies • Demonstrates that models trained on this data can match the performance of those trained on 150 human teleoperation demonstrations • Achieves a 27× increase in data generation throughput compared to traditional methods This approach enables scalable robot learning by decoupling data generation from physical robot constraints. It opens avenues for democratising robot training data collection, allowing broader participation using accessible tools. If robots can be trained effectively without physical hardware or simulations, how will this transform the future of robotics? Paper: https://lnkd.in/emjzKAyW Project Page: https://lnkd.in/evV6UkxF #RobotLearning #DataGeneration #ImitationLearning #RoboticsResearch #ICRA2025

For simulation to be viable for robotics, especially for creating AI-generated robot behavior, it is critical that robot behaviors created in simulation can be seamlessly translated to the real world AND the data from the real world can get back to that simulation! This allows for closed-loop learning of robot behaviors in simulation, using information about how the system actually performs. In this latest video, we show how READY Robotics' ForgeOS and NVIDIA Omniverse can provide this closed loop, by enabling robot programming in simulation, seamless transfer to the real world, and providing a path for data to be sent back to simulation. For modern AI algorithms to perform correctly, they need data not just of the robot's movements, but everything that the robot interacts with in its environment. This is why ForgeOS not only sends robot motion back to Omniverse but also the state of all of the tooling, as shown by the tool changer's behavior being accurately represented when mirroring the real system. ForgeOS is also able to surface sensor data, machine state, object locations, and more from the real system back to Omniverse. The ability to exfiltrate the traditionally siloed data in a robotic cell in the factory is something that ForgeOS does out of the box, without any additional IoT devices, and it ties directly back to NVIDIA's Isaac Sim. #ai #ml #manufacturing #robotics #automation #futureofwork

🦾 The latest World Robotics report reveals that manufacturing automation has hit a new milestone: 162 robots per 10,000 workers globally in 2023—double the density from 2016. As robot populations grow, manufacturers face the challenge of managing these complex operational environments. Data and AI can simplify the complexity: 📉 AI/ML models with real-time monitoring of robot telemetry data to detect performance issues before failures occur (downtime, defects, etc.) ♊ Digital twins and AI path planning help prevent costly collisions in dense robot environments 🧠 AI agents that enable dynamic optimization of multi-robot workflows 🔑 The key takeaway? Successful automation at scale requires robust data infrastructure and AI capabilities. Hardware density is just the beginning—intelligent orchestration drives real value. #Manufacturing #Robotics #AI #Automation #DigitalTransformation