Humanoid robots: 4Ps that will determine who wins

Humanoid robotics is in the middle of a massive investment boom—but where is the real opportunity, and where is the hype?

Welcome to Humanoid Daily #17, a long-form HD in which I chat with Ankur Saxena, Investment Director at TDK Ventures about the “four Ps” of physical AI:

- perception

- planning

- performance

- platform.

We dig into why perception may be the most underestimated problem in robotics, why humanoid valuations could be overheated, and why industrial deployments will likely come long before robots become commonplace in our homes.

We also discuss Agility Robotics, Figure, Atlas, NEO, edge AI, robot economics, vertical integration, and the question every humanoid company eventually has to answer: can the ROI actually justify the robot?

Plus: what would it take for Ankur to spend $20,000 on a humanoid for his own home?

00:00 Meet Ankur Saxena of TDK Ventures

01:30 The four Ps of physical AI

04:40 Robotics’ most underestimated problem: perception

07:45 Why AI compute consumes so much robot power

09:15 Edge AI and reducing the NVIDIA tax

10:45 The humanoid robotics gold rush

13:15 Agility Robotics, Figure and the companies to watch

16:15 Are humanoid valuations overheated?

18:00 When will robotics have its ChatGPT moment?

20:15 Why factories come before homes

22:15 What would make a $20,000 home robot worthwhile?

24:30 Where humanoids actually make sense in factories

27:00 The economics of replacing human workers

30:00 Why humanoid companies are subsidizing their robots

32:15 Atlas and the race to simplify robot hardware

34:30 Where the biggest robotics opportunities are now

36:30 Vertical integration vs. off-the-shelf components

39:00 The real path to defensibility in robotics

40:30 Would you buy a humanoid for your home?