I left Tesla over two years ago, and in the background, I’ve been quietly building my research and data company, Physical AI Data (physicalaidata.com). Whenever I looked for data on, say, robotaxi vehicle counts, ride counts, collision rates, or other data points, I realized there was no global “one-stop shop” dataset across robotaxis, humanoids, and other physical AI industries. So I decided to build one. And of course, once you build a super-detailed dataset with your team, it makes sense to turn that data into research and conclusions that help physical AI companies and investors understand what to focus on and where things are heading. At Tesla, I had the enormous privilege of spending quality time with some of the world’s best engineers and learning directly from them. Undoubtedly, Tesla was developing physical AI before “physical AI” became a term. So building this platform felt like a natural next step. In my view, physical AI, starting with robotaxis (our main focus thus far), will impact our daily lives so profoundly that it makes sense to build a platform dedicated to explaining it in plain language.