Summary
Recorded at Machina, a Paris conference focused on real-world AI deployment, this All-In Podcast segment features Jay Calacanis interviewing the CEOs of multiple leading robotics companies about where the industry stands and where it is heading.
The first interview features Dr. Peter Fankhauser, co-founder and CEO of ANYbotics, whose four-legged inspection robots are deployed across hundreds of industrial sites running missions up to 40 times per day. He explains why the quadruped form factor won for industrial inspection — stability in harsh environments, multi-sensor payloads including thermal cameras and gas concentration sensors — and how the economics work: avoiding even one hour of downtime at a major facility can exceed the robot’s entire cost. A key theme is that the value proposition is not labor replacement but superhuman sensing capability.
The second major interview covers 1X, a humanoid robotics company whose central bet is training foundation models on general internet video data rather than robot-only datasets. The CEO argues that making humanoid robots physically identical to humans — down to flesh-like material properties and hand proportions — is the only way to unlock the full corpus of human video as training data. The conversation addresses the recursive data flywheel problem: you need deployed robots to generate robotics-specific training data, but you need good models to deploy robots at scale. Across both interviews, the $1/hour robot labor cost framing drives discussion of timelines, competitive strategy, and what the labor market disruption will look like.
📺 Source: All-In Podcast · Published July 29, 2026
🏷️ Format: Interview
