Opportunity assessment: high. Target: the inside of a plant. Status: funded.
Electric Plant, a three-year-old startup, has raised an $11.2 million seed round to bring an AI model called Bombadil to market. The model reads the faint electrical signals running through a plant’s stem and turns them into practical guidance for growers. TechCrunch AI reports that the round was led by Onto Ventures and Climactic. Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC and Dug Song also took part.
The company says Bombadil can already interpret signals from about 125 plant species. Those range from ferns and redwoods to tomatoes and dahlias.
🎯 Situation Report
The origin story is unusual. CEO Justin Mast came across research on a “cyborg” houseplant mounted on a motorized stand. When the plant needed more light, electrical signals moved through its tissue, and a control board picked them up and rolled the plant toward the light.
“I’m like, what the heck?” Mast told TechCrunch. “That was the entry to the rabbit hole for me.”
He co-founded Electric Plant with Lee von Kraus, a neuroscientist who has worked with DARPA. They’ve spent the last three years collecting the data behind Bombadil.
📡 How the System Works
- Sensor attachment. Each sensor is roughly the size of an iPhone folded lengthwise. Stainless steel screws hold it to the plant and reach into the cambium, the living layer just under the bark, where microvolt changes happen.
- Transmission. Readings travel over a low-power wireless network to a base station.
- Density. Mast says a couple of sensors per acre is enough for accurate results.
- Interpretation. Bombadil runs in the cloud, decodes the signals, and sends recommendations back to the customer.
- Action. Growers use those recommendations to adjust watering, fertilizer and other inputs.
One caveat on wording: the company calls Bombadil a large language model. It’s better understood as a foundation model trained on bioelectrical signals, not on text. The “language” here is voltage.
⚖️ Why This Matters
Farmers and foresters already have soil probes, handheld spectroscopes and satellite imagery. Mast’s argument is that all of these look at a plant from the outside and react too slowly.
“That technology is sophisticated, but its ability to get at the inside of the plant isn’t very sophisticated,” he said. “Most of that stuff still lags quite a bit behind the real-time needs of the plant.”
I think three things make this one worth watching:
- Cost. Existing precision tools usually cost thousands of dollars. Mast says Electric Plant’s sensors cost “tens of dollars.” A price drop that big opens markets that could never pay for precision agriculture before.
- Resource efficiency. Mast says the system can raise yields and cut water use by 20% or more. If that holds up at scale, it matters a lot in drought-prone regions.
- Model breadth. “This is the most diverse model ever created on biological electrical signals,” Mast said. “No one’s ever done anything equivalent to this in the animal world.”
🔍 Points of Caution
- Undisclosed customers. Mast wouldn’t name the initial customers. Agriculture and forestry are the obvious targets, but there’s no public deployment data yet.
- Bootstrapped ground truth. Bombadil was trained using the same legacy tools it says it will outperform. Mast says “we know the model will surpass a lot of that equipment.” For now, that’s a claim, not a result.
- Self-reported numbers. The 20% water savings figure comes from the company. Independent field trials will settle it.
🧭 Strategic Context
This fits a wider pattern. Foundation models are moving past text and images into raw sensor data from biology, industry and the environment. The approach is the same in each case: gather a large, varied dataset of signals, train one model on it, and sell the interpretation as a service.
Agriculture suits this well. Margins are thin, water is getting scarcer, and growers have long made decisions by eye and by schedule rather than from live data.
📋 What Comes Next
The seed money goes toward getting the sensors and Bombadil to market. Expect more species coverage, the first named customers and, ideally, outside yield data.
The founders’ long-term idea is a tool that lets people “speak” to plants. For now, Mast keeps it modest: “We’re listening more than we’re speaking, but listening’s the first step.”
The full story is at TechCrunch AI.