Harry Mellsop

The through-line

Across every appearance, Harry Mellsop returns to one claim: physical AI, not another wave of software AI, is the next major technology shift, and the bottleneck holding it back is the inability to safely and cheaply test robots and autonomous systems before they are built [1][4][5][6]. His own path from building self-driving cars in Silicon Valley to co-founding Antioch is presented as a direct continuation of that problem: the tools that let the driving industry mature are now being generalised into a platform meant to let any company building physical AI systems test new features in simulation rather than in the real world [2][3][7]. The other consistent note is genuine surprise at how wide the demand for this has turned out to be, extending well past the vehicles and drones he expected into consumer hardware [7].

On physical AI as the next gold rush

Mellsop frames the current moment as a "physical AI gold rush," positioning robotics and autonomous systems as the frontier that follows the software AI boom, with Antioch built specifically to lay foundations for that wave rather than to chase it [1][4][5][6]. The framing draws directly on his background building self-driving cars, treating that earlier effort as the first proof point for physical AI and Antioch as the next, broader step [1][6].

On simulation as the route to better robots

The core of Mellsop's pitch for Antioch is that companies building AI robots need a way to develop and test new features without doing so on physical hardware in the physical world. He describes Antioch's technology as bringing artificial intelligence into the physical world by letting companies build new or better AI robots through software simulation, so that features can be tried and refined virtually before they are deployed [2][3]. He positions this simulation layer as applicable across a spread of industries rather than tied to one use case, specifically pointing to autonomous driving, drone technology, and manufacturing as sectors this approach will help develop [7].

On where the demand is actually coming from

Mellsop is explicit that the market response has not matched his own expectations. Beyond the industries he anticipated, autonomous driving, drones, and manufacturing, he has been struck by how much interest has come from smart security and consumer smart-home companies. In his own words: "I think something that's personally surprised me is a lot of interest from smart security companies, so smart home sensors, to smart doorbells, those sorts of applications as well" [7]. This point stands somewhat apart from his framing of physical AI around industrial and mobility use cases, and suggests his view of Antioch's addressable market has broadened since founding it around the more obvious autonomous-vehicle and robotics applications.

Takeaways

• Mellsop treats physical AI, not further software AI progress, as the next major technology wave, and describes Antioch as building foundational infrastructure for it [1][4][5][6].
• Antioch's core offering is letting companies test new AI robot features in software simulation rather than on physical hardware, which he frames as bringing AI into the physical world [2][3].
• He names autonomous driving, drone technology, and manufacturing as the key industries Antioch's simulation approach is meant to help develop [7].
• He has been surprised by strong interest from smart security and smart-home companies, including smart doorbells and home sensors, an application area beyond his original expectations [7].