Electric Plant builds AI model to decode plant electrical signals
Electric Plant raises $11.2 million seed funding to develop sensors and Bombadil, a large language model decoding electrical signals from 125 plant species.
Stock photo for illustration only, not from the actual event
- Electric Plant raised an $11.2 million seed round led by Onto Ventures and Climactic on October 20, 2026.
- The startup built "Bombadil," a large language model designed to interpret electrical signals across 125 plant species.
- Sensors cost tens of dollars, sized like a folded iPhone, attaching via stainless steel screws into the cambium layer.
- The system helps boost agricultural yields while reducing water consumption by 20% or more.
Justin Mast's perspective shifted dramatically after he read scientific literature about a houseplant wired to a motorized stand that moved toward light using electrical signals generated by its living tissues. This intriguing discovery led him to partner with neuroscientist Lee von Kraus, a former DARPA collaborator, to establish Electric Plant.
Over the past three years, the founders focused on gathering data to train a large language model named "Bombadil." The model is engineered to decipher subtle electrical signals moving through plant stems and translate those signals into actionable grower insights, bringing researchers closer to tools that could eventually let humans "speak" with plants.
Stock photo for illustration only, not from the actual event
To bring its proprietary sensors and the Bombadil model to market, Electric Plant recently secured an $11.2 million seed investment round. The funding was led by Onto Ventures and Climactic, with contributions from Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC, and Dug Song.
The hardware relies on sensors roughly the size of a regular iPhone folded lengthwise, secured to a plant using stainless steel screws. These screws penetrate the cambium layer directly beneath the bark to detect microvolt fluctuations inside the plant. Collected data transmits to a base station via low-power wireless networks before processing in the cloud.
"This is the most diverse model ever created on biological electrical signals. No one’s ever done anything equivalent to this in the animal world."
Justin Mast, Co-founder and CEO of Electric Plant
Traditional agricultural tools like soil probes, handheld spectroscopes, and satellite imagery often lag behind real-time plant needs. Electric Plant's hardware offers a much cheaper alternative priced at tens of dollars compared to traditional equipment costing thousands. Deploying just a couple of sensors per acre is enough to gather accurate data processed by Bombadil in the cloud.
Translating biological electrical impulses using modern machine learning architectures represents a fundamental shift in agricultural technology. Moving from macro environmental monitoring to direct internal plant telemetry allows growers to address biological stress before physical symptoms appear, bridging a crucial gap in precision farming.
Bombadil currently interprets signals from over 125 different species, spanning ferns, redwoods, tomatoes, and dahlias. Mast noted that optimizing schedules based on the model's feedback can boost yields while decreasing water usage by 20% or more, opening broad applications across agriculture and forestry.
Source: TechCrunch
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment