The R&D phase was fundamentally unpredictable. Massive breakthroughs in CV, prediction, planning, and safety were needed to deliver a generalizable, safe self-driving car. Each required experimentation and iteration, with failure the default. Tough to provide accurate dates!

In late ‘21, we launched our first fully driverless car. R&D was no longer the limiting factor. With our Continuous Learning Machine & sim, we further proved our tech can generalize to new cities it’s never driven in before. Since then, we’ve scaled cars & customers each week.

(To be clear, we still invest deeply in R&D, and the tech you haven’t seen yet is mind-blowing. It’s just not the limiting factor anymore!)

Scaling is a challenging, yet more predictable problem. It’s fundamentally an infrastructure challenge. We’re building both the physical and virtual infra to support and scale a fleet of novel vehicles (hi Origin!) in new markets.

It’s physical infra like standing up facilities, storage, maintenance, and charging, or virtual infra like fleet tools, CS, mapping, and remote assistance. These are very difficult challenges, but the solutions and timelines are understood. No R&D breakthroughs required.