Why Lead Time and Cost (Not Capability) Are Holding Robotics Back

Physical AI, the growing category of AI systems built to act in the real world rather than just process data, is one of the fastest-moving corners of the robotics industry. Companies building autonomous robots, automated lab equipment, and warehouse deployments are scaling quickly. But according to Ethan Wicko, Founder and CEO of US startup Transcend Mechanics, many of them are facing the same challenge: hardware cannot iterate at the speed their software does. We sat down with Wicko, and co-founder Alex McLeod, to learn about their solution to this obstacle.

A Software Mindset Meets a Hardware Challenge

“The biggest problem we heard from people, mostly online, was lead times and cost,” Wicko explained. “The hardware cycle time is way too long, on the order of weeks, while a lot of physical AI companies scaling up come from the software world and are used to much faster iteration cycles.”

The 2-finger small Nibbler grippers for the YAM arm.

That mismatch is compounding as physical AI scales. Cost, Wicko explained, is not just a barrier for smaller or early-stage teams. It becomes the central bottleneck once a company tries to deploy physical intelligence, robotic arms, grippers, and other components, at any meaningful volume.