Electric Plant Raises $11.2 Million to Launch Bombadil AI Model for Plant Health

San Francisco-based Electric Plant Company has raised $11.2 million in seed funding and introduced Bombadil, an artificial intelligence (AI) foundation model designed to monitor plant health by interpreting electrical signals from living plants. Announced on 20 October 2026, the launch follows three years of technology development and data collection. The company says its model was trained on more than four million hours of electrical recordings from over 1,400 plants across 100 species, allowing it to identify physiological stress without relying exclusively on external environmental measurements.

The funding round was led by Onto Ventures and Climactic, with participation from Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC and Dug Song. Founded in 2023 by Justin Mast, former founder of Bloomscape, and neuroscientist Lee von Kraus, Electric Plant Company is developing technology that could help growers, agricultural businesses and researchers monitor plant conditions continuously and respond to emerging stress before visible damage appears.

 

AI Model Uses Electrical Signals to Assess Plant Health

 

Bombadil works with a sensor called Bud, which attaches directly to living plants and records electrical activity within their tissues. These signals reflect physiological processes and responses to changing environmental conditions. The information is transmitted to Bombadil, which analyses the electrical readings and estimates seven physiological stress indicators, including photosystem II efficiency, leaf temperature relative to surrounding air and the Crop Water Stress Index (CWSI), a recognised measure of plant water stress.

Traditional plant monitoring commonly relies on soil moisture sensors, weather information, satellite imagery and visual inspection. These methods provide useful information about growing conditions but may not directly capture changes occurring inside plant tissues. Electric Plant's approach aims to provide an additional source of information by measuring plant responses themselves. "Bombadil makes that kind of intimate knowledge possible at scale, using data derived from inside the plant itself," said Justin Mast, Co-founder and CEO of Electric Plant Company.

 

Read more: NVIDIA Partners with Earthshot Prize to Scale Climate and Environmental Solutions Using AI

 

Foundation Model Trained on Four Million Hours of Plant Data

 

Electric Plant developed Bombadil using self-supervised learning, an AI training approach that allows models to identify patterns in large datasets without requiring every observation to be manually labelled. The initial model was trained on electrical recordings collected across different plant species and environmental conditions. According to the company, its biological signal database has since expanded to seven million hours, providing additional information for future model development.

A central objective is to allow Bombadil to interpret electrical activity from plants that were not included in its original training dataset. Earlier plant electrophysiology models often required separate development for individual species or growing environments, limiting their wider application. James Kirk, Head of AI and Machine Learning at Electric Plant Company, explained, "Bombadil learns what is common across species, which is why it can read a plant it never saw in training." The company has released a technical whitepaper describing the model, although broader commercial performance across different crops and environmental conditions remains to be established.

 

Applications in Agriculture and Environmental Monitoring

 

Electric Plant believes the technology could support more precise agricultural management by helping growers identify water stress and other physiological changes directly from plant signals. Earlier identification of stress could allow farmers to adjust irrigation, nutrient application and other growing conditions before problems become visible. Potential applications also extend to forestry, greenhouse production and environmental monitoring, where continuous information about plant health could support management decisions.

The company intends to expand Bombadil's capabilities beyond its current physiological indicators. Future applications could include identifying specific plant diseases, predicting flowering periods and monitoring other aspects of plant development. Its foundation model architecture is designed to accommodate additional analytical tasks without requiring a completely new model for each application. However, the announcement does not provide independently verified figures for improvements in crop yields, water savings or reductions in agricultural inputs resulting from commercial deployment.

 

Explore OneStop ESG Marketplace: Regenerative & precision agriculture

 

Funding Will Support Commercial Expansion

 

Electric Plant Company has already begun deploying its technology at commercial sites, with hundreds of Bud sensors operating in the field. Rather than replacing existing agricultural service providers, the company plans to license its technology to businesses already responsible for plant management. This approach could allow established operators to integrate electrical signal monitoring into their existing services while expanding the collection of plant health data.

Tom McQuillen, Partner at Onto Ventures, highlighted the potential for AI to identify patterns shared across plant species and improve understanding of plant physiology. The new funding will support Electric Plant's commercial expansion as it introduces its technology across additional crops and customers. The company's longer-term opportunity lies in demonstrating that plant electrophysiology can provide reliable, actionable information across different agricultural settings. Its environmental impact will depend on whether those insights translate into measurable improvements in resource efficiency, crop resilience and plant health management.