We are honored to partner with brilliant entrepreneurs like Tony Zhang who are creating new market categories with AI.
I’m beyond excited to finally share Tera AI (tera-ai.com) with the world. Tera is the spatial reasoning AI research + product company we’ve been building for the past year. As some of you know, I’ve been long obsessed with cognitive maps and world models – AI with general-purpose spatial reasoning like humans. Humans are great at navigating new physical environments, and we have this innate ability to form models of our surroundings (without us even knowing) that allow us to go wherever we want and recognize places we’ve been in the past. We can even read a map we’ve never seen before, in 2D, and navigate a new city. Yet this capability is almost completely missing in robotics today. We’ve spent a lot of time doing fundamental research in AI to understand how to even develop a system that exceeds humans in its ability to understand the spatial world around us. This meant novel algorithms, datasets, and training new models. This promises to be incredibly useful in autonomy/robotics, where today, custom software and expensive sensors make navigation solutions possible in constrained pre-mapped settings and require significant effort to scale across new environments and platforms. Our goal? Enable a single piece of software that “just works” across all autonomy platforms that already has a camera and a GPU. This would eliminate the need to retrofit additional hardware and simplify the solution to an over-the-air update (just like your iPhone). In the last year, we’ve become increasingly convinced that this future is not only possible but fast approaching. This would eliminate costly hardware and custom software, and allow robotics to reach its full scaling potential in doing work that is tedious, costly, and dangerous – all at significantly lower cost, weight, and time-to-deployment. I’m excited to share that we have partnered up with some of the best investors we’ve had the fortune to get to know, including Felicis (Viviana Faga, James Detweiler), Inovia Capital (Steven G. Woods, Kory Jeffrey), Caltech + Wilson Hill Ventures (Michael Borrus + Ajay Kshatriya), and Naval Ravikant. The future of spatial reasoning will only be unlocked by those who deeply invest in both novel fundamental research and GTM + product insights. So, if you’re an AI researcher, embedded AI engineer, or robotics product/GTM person, please reach out.