Plants can speak using electricity, so this startup built an AI model to decipher their language

Electric Plant, founded by CEO Justin Mast and neuroscientist Lee von Kraus, has spent three years collecting data to train a model it calls Bombadil that interprets the tiny electrical signals moving through a plant's stem. It says the model works across about 125 species and can tell growers what a plant needs. The company just raised an $11.2 million seed round led by Onto Ventures and Climactic.

If it works at scale, this is a practical climate and food story: real-time insight into crops from sensors Mast says cost tens of dollars, compared with tools that generally run thousands. Mast claims the system can lift yields while cutting water use by 20% or more.

We like this one. It's AI pointed at a physical, measurable problem, with an obvious payoff if it pans out. Growers today mostly react to plants after the fact using soil probes, handheld spectroscopes and satellite imagery. A sensor that reads what's happening inside the plant, as it happens, could let farms water and fertilize more precisely, and saving water is about as concrete a benefit as AI gets.

The design is also sensibly frugal. Mast says a couple of sensors per acre are enough, data travels over a low-power wireless network to a base station, and Bombadil interprets it in the cloud. Cheap hardware matters for adoption in agriculture, where margins are thin.

Now the caution. The model was trained using readings from the very tools it's meant to beat, and Mast says the company knows it will surpass them. That's a hypothesis, not a result. The 20% water savings figure is also the founder's claim, and the company won't name its first customers. We'd want independent field trials across seasons and crops before treating this as proven.

The framing deserves a gentle check too. Calling Bombadil a large language model and talking about speaking to plants is good storytelling, and Mast himself concedes the company is listening more than speaking. What's really on offer is signal interpretation for crop management. That is valuable enough on its own without the anthropomorphism.

One practical tradeoff: the sensors attach with stainless steel screws that must penetrate the cambium, the living layer under the bark. For orchards and forestry that's probably fine, but it's worth knowing the measurement isn't fully passive.

Mast's interest began with research on a houseplant wired to a motorized stand that moved toward light based on the plant's electrical signals. Electric Plant's seed round also included Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC and Dug Song.