Aigen Launches Alchemy, the First Generative World Model for the Open Field
Key takeaways
- Aigen has built the first generative world model for the open field, the hardest and largest setting for embodied AI.
- Alchemy generates unlimited, self-labeling environments with pixel-perfect accuracy.
- It is working now with a Fortune 500 company.
- Aigen's chemical- and fossil-fuel-free, fully autonomous robotic fleet weeds commodity crops at the same cost as spraying herbicides.
- Aigen is uniquely able to achieve these results because of large scale "truly autonomous" (via solar) data collection in the outdoors.
- The same capability extends to other outdoor and off-road domains, and to robotic action and autonomous control.
Redmond, WA — August 3, 2026
Aigen, the company building autonomous, solar-powered robots for agriculture, today launched Alchemy, the first generative world model for the open field. Alchemy learns how the hardest and largest environment in embodied AI behaves and generalizes conditions on demand, instead of waiting months or years to capture them in the real world. A Fortune 500 company is already using Aigen's world model.
Most embodied AI, from self-driving cars to warehouse robots, is built for bounded, repeatable spaces. Outdoor environments are the opposite, where no two fields, seasons, or days are alike. This makes them among the hardest environments for AI to learn and the poorest in usable training data. Collecting that data in the real world requires human labor and is season-dependent, while generic synthetic data doesn't generalize enough to be useful in training. Aigen built Alchemy to change that.
Aigen is currently using Alchemy to scale chemical- and fossil-fuel-free weed control, at a cost competitive with herbicides. Alchemy generates training data for any crop and condition, and then Aigen's Element robotic fleet successfully weeds fields and protects crops. Because the model creates and labels its own environments, Aigen can deploy its fully autonomous fleet on a new row crop in under a week, instead of the months or years that human-labeled data requires. Aigen's Element robots are trained on datasets that are up to 100% synthetic.
"Agriculture's hardest problem with robotic autonomy was never the machine. It was teaching a machine to understand an environment that never stops changing. Our world model, Alchemy, can teach any machine to operate in any open environment, and no one else can do that. We're proving it with our own robots weeding farms today and expanding to other outdoor environments soon." Kenny Lee, co-founder and CEO, Aigen
"This is a really incredible milestone for all of agriculture. Thanks to Alchemy, our robotic fleet can weed all kinds of crops completely autonomously with no chemicals and no fossil fuels, and compete with the cost of herbicides. Growers can now get affordable and effective weed control, and we all get fewer chemicals on the food we eat." Richard Wurden, co-founder, Aigen
Q&A for context
Is it actually working, or is this a research result?
Alchemy is working now. It has successfully generalized to plants and conditions it has never encountered and trained Aigen's fully autonomous crop robots using up to 99% synthetic data. Learn more at aigen.io.
Why is this cheaper and faster than how other robots are trained?
Traditional approaches depend on capturing and hand-labeling real-world data, which takes months and hundreds of thousands of dollars for each environment. Alchemy creates and labels its own environments, so Aigen's Element robots can be trained on any new crops or region in under a week.
How does this scale?
Alchemy, Aigen's world model, generalizes across any crop or condition, instead of needing humans to collect the data in the real world. That lets Aigen bring its chemical- and fossil-fuel-free weed control to new crops and geographies in weeks, at a cost that competes with herbicides. Aigen is adding new crops this season.
What's next?
Agriculture is just the beginning. The same model can generalize to outdoor and off-road domains beyond the farm and support new capabilities like robotic action and autonomous control. It is available for live demonstration now.
About Aigen
Founded in 2020 by Richard Wurden and Kenny Lee, Aigen builds fully autonomous, solar-powered robots that deliver chemical-free, fossil-fuel-free weed control and crop intelligence for large-scale agriculture. Wurden, a former Tesla engineer, and Lee, a data and security leader, founded the company to remove harmful chemicals from the food supply and build the data foundation for autonomous agriculture. Aigen sells to enterprise OEMs, agribusiness, and farmers. Learn more at aigen.io.