An AI model now reads the electrical signals of 125 plant species

Justin Mast’s life was changed by a plant. A few years ago, while reading scientific literature, he came across research on a houseplant wired to a motorized stand. When the plant needed more light, electrical signals flowed through its living tissue; a control board sensed the current and wheeled the plant toward the lamp. “I’m like, what the heck?” Mast told TechCrunch. “That was the entry to the rabbit hole for me.”

Mast teamed up with Lee von Kraus, a neuroscientist who has worked with DARPA, the US defense research agency, to found Electric Plant. For about three years the company has collected data to build Bombadil, a large language model, the same family of AI that powers chatbots, trained to interpret the faint electrical signals moving through a plant’s stem. Bombadil now covers about 125 species, from ferns, redwoods and tomatoes to dahlias.

Before this, growers worked with coarse tools that react once a plant is already short of water, nutrients or light. Soil probes, handheld spectroscopes and satellite imagery “work well enough,” Mast says, but “most of that stuff still lags quite a bit behind the real-time needs of the plant.” Electric Plant’s sensors are about the size of an iPhone folded lengthwise. They are fixed to the plant with stainless steel screws that reach the cambium, the living layer just under the bark, to pick up microvolt-scale changes. The data travels over a low-power wireless network to a base station, then to the cloud, where Bombadil interprets it and sends back advice. The company says a couple of sensors per acre are enough, and that each costs tens of dollars, where existing tools generally cost thousands.

For a grower, the practical question is when and for whom. Farmers and foresters could use the results to adjust watering or fertilizer schedules. Mast claims such a system can boost yields while reducing water use by 20% or more; that is the founders’ figure, and no customer, price or launch date has been given. Electric Plant has not disclosed its first customers, though Mast says applications in agriculture and forestry are easy to imagine. Its ambition goes further: to let people “speak” to plants one day. “We’re listening more than we’re speaking, but listening’s the first step,” he said.

The company is financing that path with an $11.2 million seed round led by Onto Ventures and Climactic, with participation from Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC and Dug Song. Mast calls Bombadil “the most diverse model ever created on biological electrical signals,” adding that nothing equivalent has been done in the animal world. That is his assessment, and it will be tested as the sensors leave the lab and reach growers’ fields.