Robotics and Artificial Intelligence Synergy

'A roadmap for AI in robotics' - our latest article (https://rdcu.be/euQNq) published in Nature Machine Intelligence, offers an assessment of what artificial intelligence (AI) has achieved for robotics since the 1990s and proposes a research roadmap with challenges and promises. Led by Aude G. Billard, current president of IEEE Robotics and Automation Society, this perspective article discusses the growing excitement around leveraging AI to tackle some of the outstanding barriers to the full deployment of robots in daily lives. It is argued that action and sensing in the physical world pose greater and different challenges for AI than analysing data in isolation and therefore it is important to reflect on which AI approaches are most likely to be successfully applied to robots. Questions to address, among others, are how AI models can be adapted to specific robot designs, tasks and environments. It is argued that for robots to collaborate effectively with humans, they must predict human behaviour without relying on bias-based profiling. Explainability and transparency in AI-driven robot control are essential for building trust, preventing misuse and attributing responsibility in accidents. Finally, the article close with describing the primary long-term challenges, namely, designing robots capable of lifelong learning, and guaranteeing safe deployment and usage, as well as sustainable development. Happy to be co-author of this great piece led by Aude G. Billard, with contributions from Alin Albu-Schaeffer, Michael Beetz, Wolfram Burgard, Peter Corke, Matei Ciocarlie, Danica Kragic, Ken Goldberg, Yukie NAGAI, and Davide Scaramuzza Nature Portfolio IEEE #robotics #robots #ai #artificial #intelligence #sensors #sensation #ann #roadmap #generativeai #learning #perception #edgecomputing #nearsensor #sustainability

Not long ago, solving a Rubik’s Cube was considered a mark of human intelligence and spatial reasoning. Can you solve the Cube that fast? Today, AI-powered robots can do it in 0.103 seconds, thanks to ultra-fast cameras capturing 4,500 frames per second and motors executing rotations in under 10 milliseconds. It’s more than a party trick — it’s a signal of how far robotics and AI have come. 📈 Processing Power: Since 2010, compute performance for AI workloads has grown by over 1 million×. ⚙️ Robotics Precision: Modern servomotors can reach accuracy levels below 5 microns, enabling surgical precision. 🧠 Learning Efficiency: Reinforcement learning models can now train 10× faster using GPU and accelerator platforms like AMD Instinct and ROCm. 🌐 Adoption Rate: Over 70% of manufacturers are investing in autonomous robotics or cobots to boost productivity and safety. The Rubik’s Cube isn’t the story — it’s the metaphor. Machines have evolved from replicating human logic to outpacing it, not through brute force but through speed, adaptability, and self-optimization. 🔹 Robots that invent their own challenges to learn faster. 🔹 AI systems that design and test hardware in simulation before humans even prototype it. 🔹 Collaborative robotics that co-create with humans — blending creativity, empathy, and logic. AI and robotics are no longer about automation; they’re about amplifying imagination. #AI #Robotics #Innovation via @cuberx5w #MachineLearning #FutureTech #Automation #ReinforcementLearning

From Agentic AI to Embodied AI Agents: The Future in Motion AI agents today exist mostly in digital form. These systems can plan, reason, and adapt, but they remain confined to software. Meanwhile, robotics has advanced in parallel, with humanoids demonstrating increasingly fluid motion, dexterous control, and real-world interaction. The intersection of these two trajectories, Agentic AI and robotics, is where embodied AI agents will emerge. When AI systems gain physical autonomy and the ability to perceive, reason, and act in the real world, they will become more than tools; they will be true agents, capable of navigating and executing complex real-world tasks. In this video, ENGINEAI’s PM01 humanoid robot is learning to dance. This may seem like a simple demonstration, but it represents something much bigger: the increasing ability of robots to learn, refine movements, and execute tasks with dynamic adaptability. The Convergence of AI and Robotics ⭐ Agentic AI: Advanced decision-making, planning, and adaptability in digital environments. ⭐ Humanoid Robotics: High-degree-of-freedom motion, real-world interaction, and dexterity. ⭐ Embodied AI Agents (The Future): AI that doesn’t just process information but moves, interacts, and autonomously operates in physical space. PM01: A Glimpse into the Future At $12K, ENGINEAI’s PM01 is pushing the boundaries of motion learning, autonomy, and real-time adaptability. While not yet an AI agent, it showcases the building blocks of embodied intelligence, the ability to move fluidly, respond to changing inputs, and execute precise physical tasks. As robotics continue to advance and AI agents grow more autonomous, the gap between intelligence and embodiment will close. Soon, AI won’t just be something we talk to, it will be something that moves, collaborates, and coexists in the real world. This future is taking shape (literally). How do we prepare for it?

As we step into 2025, NVIDIA's foray into humanoid robotics marks a pivotal moment in the convergence of AI, robotics & computing. It isn't just about creating advanced machines; it's about reshaping the very fabric of our technological landscape & redefining human-machine interaction. NVIDIA's introduction of NIM micro-services & the OSMO orchestration service represents a significant leap forward in robotics development. By reducing deployment times from weeks to minutes & streamlining complex workflows, these tools are set to accelerate innovation in the field exponentially. The AI-enabled teleoperation workflow, which generates synthetic datasets from minimal human demonstrations, addresses one of the most pressing challenges in robotics: the need for vast amounts of training data. This advancement aligns closely with the vision of companies like CynLr, which aims to create universal factories capable of manufacturing diverse products using versatile robots. CynLr's recent $10M Series A underscores the growing interest & investment in this sector. Their focus on visual object intelligence for industrial robotics complements NVIDIA's efforts, potentially leading to a synergistic relationship between AI-powered vision systems and advanced humanoid robots. The implications of these developments extend far beyond the tech sector. As humanoid robots become more sophisticated & versatile, they have the potential to transform industries ranging from manufacturing to healthcare and education. However, this rapid advancement also raises important questions about the future of work, ethical considerations in human-robot interactions and the need for robust regulatory frameworks. As we witness this technological revolution unfold, it's crucial to consider both the immense potential & the challenges that lie ahead. The convergence of NVIDIA's computing prowess with the innovative approaches of companies like CynLr could pave the way for a future where humanoid robots are not just tools, but collaborative partners in various aspects of our lives. This evolution promises to bring about unprecedented changes in how we work, live, & interact with technology, making 2025 a truly transformative year in the field of robotics and AI. #Technology #Robotics #AI #Venture #Perspective #NVIDIA

With the global AI market now valued at nearly $400 billion, and 97 million people working in the field this year, AI adoption is reaching new heights, transforming industries worldwide. The robotics sector is booming in parallel — expected to more than double in size from $71.78 billion in 2025 to $150.84 billion by 2030, fueled by the integration of advanced AI that enables sophisticated autonomy and real-time adaptation. An astonishing 83% of companies rank AI as a top priority in their business strategies, further accelerating the pace of innovation. Against this backdrop of rapid advancement, Skild AI, a robotics startup backed by Amazon and SoftBank, has just unveiled its general-purpose AI model, Skild Brain, designed to operate seamlessly across various types of robots—from factory machines to humanoids. According to Reuters, Skild’s model enables robots to "think, navigate and respond more like humans," marking a major leap toward versatile physical AI. Real-world demonstrations have shown Skild-powered robots climbing stairs, maintaining their balance when pushed, and retrieving items from chaotic environments—capabilities that demand true spatial awareness, dexterity, and adaptability. Key innovations from Skild AI’s approach include: 1. Unified "shared brain": Robots using Skild Brain share and contribute data, continually enhancing the collective model’s intelligence and functionality. 2. Humanlike adaptability: Robust performance in diverse and unstructured environments—moving beyond today’s narrow, single-task robots. 3. Safety and collaborative design: Built-in safeguards to control force and enable safe operation alongside humans in industrial and service settings. 4. Ecosystem traction: Early customers include LG CNS and unnamed logistics and industrial partners, signaling real-world validation. As we witness this transformative moment, it’s clear that the convergence of AI and robotics is poised to redefine the future of work, automation, and human-machine collaboration. How do you see general-purpose AI and robotics impacting your industry or daily life in the years ahead?

Robotics + AI need more than better models. They need structured observability. Most robotics stacks today operate without meaningful introspection. ROS 2 offers pub/sub flexibility, but not the analytics tooling to support debugging, performance analysis, or adaptive learning at scale. Meanwhile, in modern AI infra, real-time telemetry is table stakes. We track: Token-level latency in LLMs A/B drift across ranking models Fine-grained reward traces in RL pipelines System health, usage stats, model fallback rates In contrast, most robotics teams are still: Dumping unstructured bag files Parsing logs manually Lacking any notion of trace-level attribution or failure clustering This gap is limiting iteration speed — and worse, preventing reinforcement learning, self-correction, and scalable fault diagnosis. A robotics-native analytics layer should: Ingest and index sensor + action traces in real time Tag transitions with failure/success labels (auto or semi-auto) Enable embedding-based similarity search across logs Integrate with ROS 2 and edge compute to enable online introspection If you want autonomous systems to adapt in the real world, you need structured visibility into their behavior. Not just what happened, but why — and how often. The tooling that transformed backend software (Prometheus, Datadog, Sentry) or LLMOps (Langfuse, W&B) hasn’t made its way to embodied AI. Yet. We’re working on this problem. If you are too — or if you’ve hit scaling walls with ROS logs, introspection, or closed-loop learning — I’d love to talk no sales pitch, just tell me what your dream product is!

In a recent post about Oshkosh’s AI journey, I was asked how I see AI integrated into other digital technologies like robotics and IIoT.    At Oshkosh, we see a close correlation between cognitive and physical automation through AI. We are advancing our cognitive and physical automation journey in parallel. AI driven smart robots apply the perfect amount of paint on our machines and cobots like automated guided vehicles are using real-time data to navigate the shop floor, along with the potential of fully intelligent humanoid support. The connection of cognitive and physical automation with AI helps develop an intelligent enterprise that optimizes both human and machine capabilities. Continuing to embrace the partnership between AI and team members will drive smarter, more responsive operations across the value chain.    #dataanalytics #AIjourney #generativeAI #advancedanalytics #digitaltransformation #datascience 

We are entering a world where people, AI agents and robots will work side by side. McKinsey Global Institute's latest report, “Agents, robots, and us”, shows that while 57% of US work hours are technically automatable, most human skills remain relevant. What shifts is how we use them. Demand for AI fluency has grown sevenfold in just two years, signalling that the ability to work confidently alongside AI is quickly becoming essential. The most meaningful impact will come from redesigning entire workflows so people, AI agents and robots operate together. This is where the projected 2.9 trillion dollars in annual value by 2030 could be unlocked. Human centred skills such as leadership, coaching and care remain remarkably stable, while digital and information oriented skills continue to evolve. Managers may increasingly guide hybrid teams and focus on orchestration and judgment. For individuals, transferable skills combined with AI capability offer the strongest path to long term resilience. My view: treating AI as a collaborative tool within our work, rather than something separate from it, will help us navigate this shift with more confidence and clarity. You can read the full report here: https://lnkd.in/dfHnx9i9 #artificialintelligence #futureofwork #leadership #innovation #technology Image from 'Agents, robots, and us: Skill partnerships in the age of AI'.

Industrial software competition isn’t just about features anymore— it’s about who can integrate AI, robotics, and enterprise systems into an open architecture ready for Agentic AI and a personalized user experience 🤖⚙️. At Industrial X, the IFS strategy was clear: AI agents, robots, and enterprise workflows converging into one execution loop. ➡️ Partnerships with Boston Dynamics and 1X® Technologies show the robotics side….. but IFS could go even further by leaning on Poka Inc. as the execution layer to bridge skills, work, and AI-driven action. Rockwell Automation Fair showcased solid automation and analytics advancements…. but meaningful physical AI and robotics partnerships and investments didn’t get announced. ➡️ Matthew Littlefield pointed out an overly conservative AI strategy at a time when the market is clearly moving toward deeper autonomy and robotics-enabled operations. Vendors and investment firms are betting more on open, interoperable software solutions and robotics as the Future of Industrial Work #FOIW ecosytem evolves. ⏳ Here’s where I see market signals aligning: • FANUC America Corporation + NVIDIA partnering for robotics + accelerated AI computing enabled a decent market bump • IFS has had strong Industrial AI messaging and has moved towards robotics partners for integrated physical + digital workflows • Rockwell Automation → strong automation base, but limited moves toward physical AI partnerships. • QAD has strongly positioned Redzone as the human-centric execution layer for the larger portfolio + Champion AI, but lacks the robotics partnerships for physical AI. The market shifting fast, and your strategy for AI, an intelligent supply chain, and robotics matters. For manufacturers: 1️⃣ Demand proof, not promises. AI- and “agent”-washing is exploding and vendors must show how AI outputs translate into execution, not dashboards. 2️⃣ Prioritize open, interoperable systems. Closed systems will struggle to support autonomous workflows or cross-platform AI agents, with CFW the preferred UI for the Intelligent Supply Chain. . 3️⃣ Scale workforce capability alongside AI. Agentic systems amplify human judgment — if your workforce isn’t ready, the technology won’t deliver ROI. For vendors: 1️⃣ Differentiate through interoperability. Platforms must unify humans, automation, robots, and AI into a single operational loop. 2️⃣ Move beyond advisory AI. Systems should detect, decide, and initiate action — including physical actions via robotics. 3️⃣ Form partnerships that accelerate autonomy. Wait for the ecosystem to mature will allow vendors engineering it themselves the competitive advantage. It’s no longer about feature parity — it’s about who builds the most interoperable, robotics-aware foundation for the next generation of industrial work! 🔔 Connect with me for more LNS Research thought leadership! #IndustrialAI #Robotics #AgenticAI #IndustrialTransformation #Leadership #FutureOfWork #CFW #OperationalExcellence