Electric Plant, a two-year-old San Francisco startup, announced October 20, 2026, it has raised $11.2 million in seed funding to scale a sensor that claims to read electrical signals from inside plants and translate them into actionable stress data. Onto Ventures and Climactic led the round, with backing from Ponderosa, Endeavor Catalyst, Ludlow, Michigan Rise, The Fund, Schox VC and Dug Song, the angel investor and security entrepreneur.
The pitch sounds almost mystical: tiny devices screwed into tree bark, streaming biometric data to an artificial-intelligence model that founder Justin Mast says can decode what 100 species are experiencing in real time. But the company insists its technology addresses a prosaic problem in commercial agriculture. Growers spend hundreds or thousands of dollars per acre on soil probes, weather stations and remote sensing to infer when crops are thirsty or stressed. Electric Plant's hardware costs "tens of dollars" per unit, Mast told TechCrunch Monday, and purports to measure stress from within the organism itself.
The startup has deployed hundreds of sensors at commercial sites, according to a Business Wire release, though it declined to name customers or specify which crops or geographies are involved.
Reading the plant's own language
Electric Plant's sensor, which the company calls Bud, uses stainless-steel screws that penetrate the cambium layer of woody plants to capture electrical signals. Those signals travel wirelessly to a base station. The company pairs the hardware with Bombadil, an AI model the company describes as the first foundation model for plant electrophysiology.
Bombadil was pretrained on more than four million electrode-hours of data collected across 100 species, according to a white paper the startup published in mid-October. The model estimates seven stress markers that agronomists typically measure with separate instruments, including fluorescence and thermal indicators.
The economic contrast with incumbents is stark. Soil-moisture probes and canopy monitors from companies like CropX and Arable can run hundreds to roughly $1,200 per unit. Mast argues his approach is more practical: a couple of sensors per acre, rather than blanketing fields with external monitors. He claims the system could boost yields while cutting water use by 20 percent or more, though the company has not released independent agronomic trials to validate those figures.
A team of horticulturists and neuroscientists
Mast is a fifth-generation horticulturist who previously founded Bloomscape, a direct-to-consumer houseplant business. His co-founder, Lee von Kraus, holds a doctorate in neuroscience and has worked with DARPA and the EPA. Von Kraus also co-founded Halo Neuroscience, a brain-stimulation startup. James Kirk, the company's head of AI and machine learning, authored the Bombadil technical paper alongside Mast and von Kraus.
Electric Plant's LinkedIn page listed the team at somewhere between two and 10 employees as of October 21, 2026.
In Monday's announcement, Mast framed the technology in almost artisanal terms. "Master gardeners spend a lifetime learning a single plot, from the soil moisture level to the way the sunlight hits a tree in a specific corner," he said. "Bombadil makes that kind of intimate knowledge possible at scale, using data derived from inside the plant itself."
'Inside-out' versus the status quo
The company positions itself as practicing "inside-out" sensing, in contrast to the "outside-in" proxies that dominate precision agriculture today. Switzerland-based Vivent Biosignals also sells commercial plant-electrophysiology systems, using non-invasive electrodes to monitor stress in real time. Electric Plant is betting that its invasive approach and cross-species AI model offer an edge.
Tom McQuillen, a partner at Onto Ventures, said in the release that "what excited us about Electric Plant is the possibility that all plants share an underlying electrical language that can be learned across species."
Whether that linguistic metaphor holds up under field conditions is another question. The Bombadil white paper, released the day before the funding announcement, reports that the model tracks within-plant changes under water stress. It also notes that three of the seven physiological markers performed worse than a "no-skill" baseline on one position metric, attributed to per-sensor offsets. The authors flag calibration and multi-sensor blending as areas for future work.
Kirk acknowledged the challenge in the announcement: "Most earlier models of plant electrical signals worked for one species in one setting, so a new plant or new environment meant starting over."
What comes next
Electric Plant's website lists use cases that stretch beyond row crops into orchards, forestry health monitoring, wildfire-risk assessment, water-quality sensing and real-time carbon accounting. The company declined to share how it plans to spend the new capital or whether it intends to expand the team significantly.
Michigan Rise, a state-backed pre-seed fund, listed Electric Plant in its portfolio in mid-2025, indicating the startup had raised earlier support before Monday's seed round.
For now, the company is pitching a future in which plants broadcast their needs in a language growers can finally understand. Whether the technology proves reliable enough to displace entrenched competitors, and whether the economics hold at commercial scale, will become clearer as the startup moves from initial deployments to broader agronomic trials.
