Agentic AI Moves Into Physical Robotics and Logistics
The application of agentic AI is expanding from conversational interfaces to embodied robotics and industrial workflows. LG has embedded its Exaone foundation model into humanoid robots and launched a new Physical AI Lab to bridge LLMs with physical agents. In logistics, Pallet launched an AI platform that uses intelligent agents to orchestrate end-to-end warehouse workflows, while Rufus Labs released a platform to unify data from human and robotic workforces.
- LG's "One LG" framework for physical AI involves consolidating capabilities from its affiliates, including LG Electronics for humanoid robots, LG Innotek for robot sensors supplied to competitors like Boston Dynamics and Figure AI, and LG Energy Solution for robot-optimized batteries. The new Physical AI Lab is moving beyond vision models to focus on "action models," which are more advanced AIs that translate perception into physical motion and are essential for humanoid robot commercialization. - Pallet's AI platform targets "exception-heavy" logistics workflows like cross-border shipments and invoice auditing, which are often resistant to automation. Early enterprise clients have reported up to a 15% reduction in operating costs by using Pallet's logistics-trained AI models and workflow agents to manage these complex processes. The company has raised a total of $50 million from backers including General Catalyst and Bain Capital Ventures to scale its AI workforce. - The Rufus Labs WorkHero Universe platform provides a unified dashboard for warehouse KPIs by integrating with WMS and ERP systems. It allows managers to track metrics such as pick rate, cost per pick, orders fulfilled, and time spent on specific tasks like kitting and putaway for both human and robotic workers. - The broader humanoid robot market is seeing early commercial deployments, though shipment volumes are still nascent. Chinese firms Agibot and Unitree led 2025 shipments, while U.S. companies like Figure AI, Agility Robotics, and Tesla are focused more on innovation and functionality for industrial applications before scaling production. - Venture capital investment in robotics is concentrating on fewer, larger deals with clear commercial applications, a shift from broad investment in autonomous vehicle platforms. In the first seven months of 2025, robotics startups raised over $6 billion, with significant rounds going to companies in warehouse automation, surgical systems, and humanoid robotics. - The Department of Defense's AI Acceleration Strategy, detailed in a January 2026 memorandum, prioritizes the development and experimentation of AI agents for battle management and decision support. This initiative, along with a focus on AI for supply chain resiliency, signals a key demand driver for agentic systems within the defense sector, a core market for defense tech professionals. - Agentic AI systems in industrial settings are moving beyond task automation to autonomous workflow orchestration, with companies like Fanuc operating factories where robots build other robots with AI agents acting as supervisors. This approach uses multi-agent systems where different AIs can negotiate tasks, delegate workloads, and schedule their own maintenance to minimize downtime, a model Siemens has used to reduce unplanned downtime by 25%.
