On the horizon The robotics ecosystem is advancing with physical AI, featuring humanoid robots like Figure’s Figure 01, Unitree’s G1/H1 and Boston Dynamics’ Atlas. These robots combine AI with whole-body control systems to navigate human environments and perform dexterous tasks. Autonomous drones are also evolving into adaptive swarms capable of independent navigation and complex missions even without GPS. A major challenge in physical AI development is data scarcity, unlike language models that have vast datasets. This is being addressed through simulation platforms like NVIDIA’s Omniverse and Isaac Sim, where robots learn in virtual worlds before applying skills in reality. These simulations include synthetic data generation, multi-agent interaction and photorealistic rendering, enhancing robots’ learning and adaptability. Integrating LLMs allows natural language instruction and semantic understanding, lowering the entry barrier into robotics and enabling rapid design iteration. The future of robotics involves developing foundation models for general-purpose capabilities across various platforms, similar to LLMs in language processing. Companies like Physical Intelligence and Skild AI are creating these models by training on diverse datasets from simulations and real-world interactions. These models will enable robots to generalize from prior experiences, reducing deployment time and engineering effort. The cost of building humanoid robots is expected to drop significantly, from approximately $35,000 in 2025 to around $17,000 by 2030.25 This reduction is driven by advancements in component design, economies of scale, and cost-effective materials and manufacturing techniques. As costs decrease, robot adoption will accelerate, enabling businesses to automate tasks in manufacturing, logistics, healthcare and agriculture, facilitating innovation and efficiency at scale. Progress in advanced materials is fundamentally transforming how technological solutions are designed, manufactured and implemented across industries. This technology domain encompasses a diverse range of advanced materials, meticulously engineered to meet the evolving needs of various industries. These materials stand out due to their enhanced properties and functionalities, enabling novel applications and convergence with other technology domain, driving improvements in performance, sustainability and efficiency in novel applications. The integration of AI has further unlocked new avenues for developing innovative materials, expanding possibilities across a wide array of industries and use cases. New developments in self-healing materials and biofabricated materials are enhancing adaptability and sustainability, while thermoelectric materials are optimizing energy efficiency in various applications. Meanwhile, established technologies such as energy storage materials are improving battery performance and longevity, piezoelectric materials are enabling energy harvesting and advanced sensing capabilities, and biocompatible implants are revolutionizing medical treatments.2.6 Advanced materials domain Technology Convergence Report 39

Technology Convergence Report 2025