Mapping the Robotics Market on One Whiteboard

Robotics had a big summer, and it is genuinely hard to tell who is doing what. Claire Zau met with Lightspeed partner Amber Yang in front of a whiteboard to draw the market on two axes. How general is the model, and how general is the body. Yang explains what actually changed at the model layer, why robots cannot just be trained on the internet the way language models were, and why she thinks humanoids are not about to get cheap. Roughly 40 to 60 percent of a humanoid build is actuators, and that supply chain runs through China. Then she fills in the map with examples. Dexterity, Tacta and Reflex Robotics building their own models for their own hardware. Ultra skipping training entirely and fine-tuning on Physical Intelligence. Skild and Physical Intelligence chasing one brain for many bodies. Sunday Robotics, 1X and Figure on the humanoid side. She also makes the case for what she calls DeployCos, companies that take someone else's model into places like a restaurant or a nursing home without the customer ever knowing what is under the hood. Her closing point is the one we believe worth sitting with. Robotics still has no shared benchmark, so every company grades its own homework, and in a field this tied to safety that gap may be an opportunity. Stay in touch: www.lsvp.com X:   / lightspeedvp   LinkedIn:   / lightspeed-venture-partners   Instagram: https://www.instagram.com/lightspeed/... The content here does not constitute tax, legal, business or investment advice or an offer to provide such advice, should not be construed as advocating the purchase or sale of any security or investment or a recommendation of any company, and is not an offer, or solicitation of an offer, for the purchase or sale of any security or investment product. The views expressed by our guests do not necessarily represent the views of Lightspeed. For more details please see lsvp.com/legal. CHAPTERS: 0:00 Why Older Robotics Companies Struggled 0:40 Mapping The Robotics Landscape Across Two Axes 1:16 Model Layer Breakthroughs From VLAs To World Action Models 4:20 One Brain Many Bodies And The Value Of Specific Models 5:27 In-Context Learning And How Robotics Companies Collect Training Data 7:40 Hardware Developments And Dependence On China 9:24 Specific Model Specific Embodiment Quadrant 13:05 The Deploy Co Angle And Open Router For Robotics 13:55 General Model Multiple Bodies Quadrant 15:50 General Model Specific Embodiment Quadrant 17:24 Where Things May Be Commoditized And Infrastructure Opportunities 20:45 Robotics Observability And Evaluation 22:33 Learning From Autonomous Vehicles The Difference Between LLMs and Robotics Data    • Mapping the Robotics Market on One Whiteboard