Try the prototype
← All resources

Why we're building autonomy for places without roads

November 19, 2026

FieldSim began in the Department of Energy’s BOOST program, built on simulation technology licensed from Idaho National Laboratory. The first idea was simple: before a sprayer or tractor enters a field, simulate hundreds of possible routes and hand the operator the best one.

What the field taught us

This summer we took that idea to Montana State University’s Lutz Farm and ran it against a real spray operation. The simulation held up. What we learned was about how the work actually happens:

  • Machines have real constraints. Ground speed is set by the application rate, not by the optimizer.
  • Guidance is set once. Most operators load a field’s guidance a single time and reuse it, so a better line isn’t worth paying for again every season.
  • The opportunity moves with the machine. The bigger opportunity is vehicles that plan for themselves, mission after mission: drones first, then ground robots.

Where we’re going

So we adapted. We now fly drones to map fields and worksites and feed that data back into our simulations. We’re building a rover with Prism Labs, and its first jobs will be mining inspection and agricultural scouting. Most of our 2027 work is in mining, and later we’ll add ground-penetrating radar surveys in the same environments.

The long-term goal is to prove these use cases in the hardest places, then offer FieldSim’s simulation and path planning as a platform for the robotics companies building the next generation of off-road machines, whoever makes the vehicle.

If you’re working on autonomy in mining, agriculture, or forestry, we’d like to hear from you.