AI Deciphers Plant Electrical Signals to Optimize Crops

Overview

A new startup called Electric Plant is leveraging artificial intelligence to decode the subtle electrical signals plants emit, offering growers unprecedented visibility into crop health and needs. Founded by Justin Mast and neuroscientist Lee von Kraus, the company has developed a large language model named Bombadil that interprets bio-electrical data from plant stems.

The technology aims to bridge the gap between coarse, reactive agricultural tools and the real-time physiological requirements of living plants. By listening to these internal electrical communications, the system promises to optimize water usage and boost yields across diverse agricultural sectors.

The Science Behind Plant Communication

The concept of "talking" to plants originated from Mast’s encounter with research involving a cyborg houseplant. This experimental plant was wired to a motorized stand that responded to its electrical signals, moving closer to light sources when the plant’s current indicated a need for more illumination.

“I’m like, what the heck? That was the entry to the rabbit hole for me,” Mast recalled.

This discovery sparked a three-year effort to collect and analyze biological electrical data. Mast partnered with von Kraus, who brings extensive experience from working with DARPA, to build a robust dataset. Their goal was to create an AI model capable of translating these microvolt changes into actionable insights for growers.

Bombadil: The Plant Interpretation Model

At the core of Electric Plant’s solution is Bombadil, a large language model trained on electrical signals from over 125 different plant species. The diversity of the training data is significant, ranging from ferns and redwoods to tomatoes and dahlias.

Mast describes Bombadil as the most diverse model ever created for interpreting biological electrical signals. He notes that no equivalent system exists in the animal world, marking a novel application of AI in agricultural biotechnology.

The startup initially used traditional tools, such as soil probes and handheld spectroscopes, to generate baseline data for training. However, the resulting model is designed to surpass the capabilities of these legacy devices, which often lag behind the actual, immediate needs of the plant.

Hardware and Deployment

Electric Plant’s hardware consists of sensors roughly the size of an iPhone folded lengthwise. These devices are attached to trees or crops using stainless steel screws that penetrate the cambium, the layer of living tissue just beneath the bark. This connection allows the sensor to detect microvolt fluctuations occurring within the plant.

Data is transmitted from the sensors to a base station via a low-power wireless network. From there, the information is sent to the cloud, where Bombadil processes it and sends results back to the user. The system requires only a couple of sensors per acre to achieve accurate readings, making it scalable for large agricultural operations.

Market Position and Funding

To bring this technology to market, Electric Plant recently secured a $11.2 million seed round. The funding was led by Onto Ventures and Climactic, with participation from Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC, and Dug Song.

One of the primary value propositions is cost. Existing sophisticated agricultural tools typically cost thousands of dollars, whereas Electric Plant’s sensors are priced in the tens of dollars. This affordability makes high-precision plant monitoring accessible to a wider range of growers.

Mast emphasized that even in highly controlled greenhouse environments, farmers often lack full visibility into their plants' internal states. Traditional methods react to demands for water, nutrients, or light rather than predicting them.

“That technology is sophisticated, but its ability to get at the inside of the plant isn’t very sophisticated. Most of that stuff still lags quite a bit behind the real-time needs of the plant,” Mast explained.

Impact on Agriculture

The practical applications of this technology are broad, covering both agriculture and forestry. By interpreting plant signals, growers can adjust watering and fertilizer schedules with greater precision. Mast claims that such optimization can lead to increased yields while reducing water consumption by 20% or more.

Although the company has not disclosed its initial customers, the potential for impact is significant. The ability to "listen" to plants represents a shift toward more responsive and sustainable farming practices.

“We’re listening more than we’re speaking, but listening’s the first step,” Mast said.

As the startup moves forward, Electric Plant aims to refine Bombadil’s accuracy and expand its species coverage. The combination of affordable hardware and advanced AI interpretation could redefine how humans interact with and manage plant life.