Electric Plant Company Emerges From Stealth With 11.2 Million Dollars and an AI That Reads Plant Electrical Signals

Plant care has always been an exercise in indirect observation: soil moisture probes, satellite imagery, and agronomists making educated guesses from what a plant looks like. A San Francisco startup called Electric Plant Company is proposing a more direct channel, and it launched from stealth on Tuesday with 11.2 million dollars in seed funding and a foundation model for plants named Bombadil.

The system has two parts. Bud, a sensor that attaches to a plant, records its electrical activity in real time. Bombadil, the foundation model, is trained on electrical signal data from 100 plant species and translates those signals into a direct answer, such as water, nutrients or attention, replacing costly in-person inspection. The company says the approach builds on plant electrophysiology research, the field studying the voltage signals plants use to respond to their environment.

The founders pair horticulture distribution experience with neuroscience. CEO Justin Mast previously founded Bloomscape, an online plant retailer, while chief scientist Lee von Kraus is a neuroscientist who has worked with DARPA and NASA and co-founded Halo Neuroscience. The pairing is deliberate: reading electrical signals from living systems is closer to neurobiology than to traditional agronomy.

The seed round was led by Onto Ventures and Climactic, with participation from Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC and angel investor Dug Song. Tom McQuillen, a partner at Onto Ventures, said the thesis is that all plants share an underlying electrical language that can be learned across species, turning previously unreadable noise into a foundation model that improves as more plants are connected.

The business model is licensing rather than replacement. Electric Plant sells its technology to operators already running plant care businesses, from greenhouse growers to agribusinesses, and is accepting additional founding operators as it expands crop by crop. Early deployments with hundreds of sensors in the field are already running at commercial sites.

Most AI in agriculture looks outward, at weather, drones and satellite imagery. Electric Plant is betting the more valuable signal comes from the plant itself, and that a single model trained across species can scale intimate, per-plant care to commercial agriculture. The claims rest on company-run deployments so far, so independent validation of how accurately Bombadil translates signals into diagnoses will be the metric to watch. If it holds, the startup is outlining a genuinely new interface between growers and the trillions of dollars of economic activity that depend on plants.