Editorial illustration for Startup Builds AI Model to Decipher Plants' Electrical Language
AI Decodes Plants' Electrical Signals for Movement
Startup Builds AI Model to Decipher Plants' Electrical Language
Justin Mast found the research almost by accident: a houseplant wired to a motorized stand, rolling itself toward a light source whenever its internal electrical signals spiked. A control board was reading the current and acting on it. Mast couldn't shake the idea. That stray discovery sent him down a path that ended with a company called Electric Plant, co-founded with neuroscientist Lee von Kraus, who has worked with DARPA on biological signal processing.
For the past three years, the two have been collecting data on the faint electrical activity that runs through plant stems, the kind of signaling that plants use internally to respond to light, water, and stress. That data feeds Bombadil, a large language model built to translate those signals into something a grower can actually use. Electric Plant just closed an $11.2 million seed round, led by Onto Ventures and Climactic, with Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC, and Dug Song also putting in money.
Mast's pitch starts with a gap he says exists even in high-end greenhouses: growers still can't see what their plants actually need.
Mast wound up partnering with Lee von Kraus, a neuroscientist who has worked with DARPA, to found Electric Plant, which has been spending the past three years collecting data to build a large language model, Bombadil, capable of interpreting the subtle electrical signals coursing through a plant’s stem and relaying details about the plant’s needs to growers.
Why this matters
Electric Plant's pitch is a reminder that AI's next frontier isn't just language or code, it's signal processing applied to biology we never bothered to listen to. Mast and Lee are building a model to read electrical patterns in plant tissue the way an NLP model reads text, and that's a genuinely different training problem: no labeled sentences, no human feedback loop, just raw voltage fluctuations that may or may not carry intent. For researchers, the interesting part is what counts as "data" here and how you validate a model's output when the subject can't confirm you got it right.
For founders, the funding round signals investors are willing to bet on sensor-plus-AI products outside the usual software stack. We're skeptical of the "speak to plants" framing, Mast himself admits they're still just listening. That's the honest part.
Worth watching: whether Bombadil's interpretations hold up against controlled experiments, or whether this becomes another case of pattern-matching dressed up as translation.
Related Reading
Open Source D&B rebuilds 642 million‑business database after AI agents hit limits
Open Source Box CEO Aaron Levie Says AI Will Drive "Biggest Cybersecurity Upgrade
Open Source Study: 6.7% of Deepfake Requests on Hugging Face Targeted Children
From Archives Cloudflare Adds Clairic System One Model to Workers AI for Data Extraction
From Archives Multimodal AI Model Clef, Fine-Tuned From Qwen, Comes to Ollama's Cloud
Common Questions Answered
What is Bombadil and how does it interpret plant electrical signals?
Bombadil is a large language model developed by Electric Plant that has been trained over three years to interpret the subtle electrical signals coursing through a plant's stem. The model translates these electrical patterns into actionable information about the plant's needs, allowing growers to better understand and respond to what their plants require for optimal growth.
How did Justin Mast come up with the idea for Electric Plant?
Justin Mast discovered the concept almost by accident when he observed a houseplant wired to a motorized stand that rolled itself toward a light source whenever its internal electrical signals spiked. A control board was reading the plant's electrical current and acting on it, which sparked the idea that led him to co-found Electric Plant with neuroscientist Lee von Kraus.
What is Lee von Kraus's background and why was he chosen as co-founder?
Lee von Kraus is a neuroscientist who has previously worked with DARPA on biological signal processing, making him an ideal partner for Electric Plant. His expertise in interpreting biological signals was crucial for developing the technology to read and decode the electrical language of plants.
How is training Bombadil different from training traditional NLP models?
Unlike traditional natural language processing models that rely on labeled sentences and human feedback loops, Bombadil's training involves raw voltage fluctuations from plant tissue with no predetermined labels or human guidance. This represents a genuinely different training problem that requires novel approaches to signal processing applied to biological systems.
Why does Electric Plant's work represent AI's next frontier according to the article?
Electric Plant's pitch highlights that AI's next frontier extends beyond language and code into signal processing applied to unexplored biological systems. By building a model to read electrical patterns in plant tissue the way NLP models read text, the company is opening entirely new possibilities for understanding and communicating with the natural world.
Further Reading
- Plants can speak using electricity, so this startup built an AI model to decipher their language - TechCrunch
- When Plants Respond: Electrophysiology and Machine Learning for Green Monitoring Systems - Springer
- Now scientists can hear when plants are saying ‘Ouch!’ - The Times
- PlantLeaf: Detect and analyse acoustic and bioelectric signals emitted by plants - GitHub
