The Future & Funding of Physical AI | Humanoids Summit Panel

In this talk

1. 2:14

   Where investors see opportunity

   Sam Carter points to sensor fusion, environmental data, and communication among machines as promising areas, noting that one oil rig can generate more data in a month than the internet. Jake highlights foundation models, better training data, and dexterous hands as unmet needs. Kang adds data management, tactile sensing, and low-latency edge chips for delicate tasks.

2. 6:59

   Evidence a robot can work in the field

   Kang says a strategic deployment partner can give a startup access to real-world data and proof of a use case. Jake wants either a clearly differentiated enabling technology or evidence that a specialized robot can do useful work and earn revenue. Carter prioritizes reliable operation, scalable sales, and connections to established buyers and integrators.

3. 11:43

   Demos, costs, and the path to production

   Carter says customers respond to teams that understand operational problems and can explain constraints that operators may have missed. Jake cautions that demonstrations can be unreliable and raises manufacturing readiness, safety, and cost concerns, citing a reported $40,000 sensor bill of materials and costly hand repairs. Kang says demos can still suggest where robots might first be useful, such as flipping puzzles or drilling parts.

4. 18:03

   Regional strengths and investment styles

   Kang describes the United States as strong in AI talent and China in manufacturing supply chains and lower component costs; he sees Korea’s manufacturers and flexible deployment environment as a useful test bed. Jake says U.S. investors often back ambitious technology for longer, while Korean startups can reach revenue faster through local corporate partners and pressure to monetize. Carter suggests resource constraints and concentrated customers in some markets could help shape practical approaches to physical AI.

5. 26:12

   Milestones and advice for founders

   Jake frames the investment case in two ways: show revenue and scalability when technology is not strongly differentiated, or pursue fundamental breakthroughs such as better robot memory. Carter expects return on investment, serviceability, telemetry, and interoperability to matter as deployments grow. Panelists advise Korean founders to secure strategic partners, demonstrate measurable progress, and focus on a market where they can stand out.