Chapter 2 of Building Robotic Space
We Started by Mapping the World
Peter Atalla, CEO, Robotic Data
November 2026 · 5 min readDraft
Chapter 2 · 2:30
Peter Atalla
For most of our history, we were called Voxelmaps. We built 3D maps of the world for autonomous systems. Street by street, city by city, country by country.
Mapping for machines is a different job from mapping for people. A person needs to know which road to take. A machine needs to know exactly where the curb is, how high the overhang is, where the pole stands, and whether any of that changed since yesterday. A map that's good enough for a phone is nowhere near good enough for a car with nobody in the driver's seat.
Chapters 2 onward are for subscribers. It's free. Tell us who you are and you'll get each new chapter from Peter.
Robotic Data
Building robots that need to understand the real world?
Robotic Data captures and engineers the real-world data Physical AI learns from: street-level, aerial and indoor world data, and human task data from people doing real work. Eight years, 26 countries, from one site to an entire country.
Next: Chapter 3
The City Changes Overnight
Why a perfect map can be wrong by morning.