World Internet Conference Spotlights Embodied AI
The World Internet Conference hosted a special forum in Barcelona focused on Embodied AI. The event, themed "Leading a New Paradigm of AI Development," signals growing international focus on AI systems that can physically interact with the world, a key step towards advanced robotics and automation.
The Barcelona forum, held alongside Mobile World Congress on March 3, 2026, featured speakers from Qualcomm, Intel, Alibaba Cloud, and China Mobile. The discussion underscored a global push to solve the hardware and software challenges of moving AI from digital spaces into the physical world. A core technical hurdle is the latency in the perception-to-action pipeline. Frameworks like "Corki" are tackling this through algorithm-architecture co-design, predicting future trajectories rather than discrete frame-by-frame actions to dramatically reduce LLM inference frequency and speed up response times. In manufacturing, the technology is already being piloted. SAP’s "Project Embodied AI" uses cognitive robots at refrigeration specialist BITZER to autonomously manage warehouse tasks 24/7 with minimal human input, connecting directly to SAP's Extended Warehouse Management system. The automotive sector is also deploying this tech; China's Zeekr factory now uses UBTech humanoid robots for a range of tasks on its assembly line, including lifting, assembly, and quality inspections, demonstrating real-world industrial application beyond the lab. Deploying these systems on edge devices remains a significant challenge, requiring the translation of high-level instructions into executable actions under tight computational and energy constraints. This demands intense hardware-software optimization to ensure the real-time performance necessary for safe operation. The market is fragmented, with key industrial players like ABB and KUKA competing alongside tech-focused companies such as Boston Dynamics and Tesla. Heavy investment is flowing from both private corporations and governments, particularly in the US and China, signaling a strategic global focus on leadership in physical AI systems.
