
Advanced Technology Insights: Physical AI and humanoid robots
Welcome to What Tech Can Do, the podcast series from Roland Berger's Advanced Technology Center that provides insights into technological and industrial shifts and helps build the foundation for successful business strategies based on operational realities. This episode dives into humanoid robots and what truly separates science fiction from industrial reality. Host Carina Lainer is joined by Thomas Kirschstein (Semiconductor Practice), Jonas Zinn (Manufacturing & Operations Practice), and Shuai Shi (Automotive Practice, Shanghai) to unpack why humanoid robots are technically capable, yet still far from scalable deployment. This episode covers why AI and data, not hardware, are now the biggest bottleneck for humanoid robots. It explores the economics behind a projected USD 2 per hour operating cost, and what that means for labor markets. It also examines why China is deploying 15,000 units this year, 30 times North America's volume, and whether this points to two separate technology ecosystems emerging, similar to semiconductors. On the operations side, this episode breaks down what humanoid robots can and cannot do in factories today, from bin picking to kitting, and where the real value unlocks once robots can switch between tasks. It also discusses the "data-for-discount" trade-off operators face when partnering with robot OEMs. Finally, this episode looks at the business opportunity ahead: a potential USD 300-750 billion market by 2035, how component suppliers can position themselves as development partners rather than vendors, and whether Europe can still catch up in the global humanoid robotics race. 🎧 Listen now to understand what it takes to move humanoid robots from pilot to scale, and what it means for operators, OEMs, and suppliers. For more information, visit our website at https://www.rolandberger.com/en/Insights/Podcasts/















