The newest foundation model doesn’t read books or web pages — it reads plants. Electric Plant Company emerged from stealth on October 20 with $11.2 million in seed funding and the debut of Bombadil, billed as the first AI foundation model for plant electrophysiology.
In brief: plants carry faint electrical currents through their tissues, and those signals can be recorded continuously, in the field, by the thousands. Bombadil is trained to interpret them. The company says the model has seen more than 4 million hours of recordings across 100 species and a range of environments — the most species-diverse model of its kind, it claims — and that its data corpus has since grown to 7 million hours.
Instead of inferring a plant’s health from soil, weather or what the eye can see, Bombadil reads the plant’s health state directly from its own electrical activity. To make that practical, Electric Plant built the full stack: a sensor called Bud that captures a tree’s electrical signals, and the model layer that translates them into real-time health results.
The company was co-founded in 2023 by CEO Justin Mast, who previously founded and led Bloomscape, and chief scientist Lee von Kraus, a Ph.D. neuroscientist who has worked with DARPA and the U.S. EPA. Their pitch mirrors the logic of large language models: a foundation model for plant biosignals becomes more useful as it learns from more species and more hours of recordings.
Measured accuracy and commercial timelines are still unreported — so far the news establishes the announcement — but the ambition is striking: an AI translator for a biological channel that has always been invisible to human observers.
Source: Business Wire via Markets Financial Content, Renascence
