Electric Plant
Plant intelligence technology decoding electrical signals for real-time health and growth insights.
Website: https://www.electricplant.co/
Cover Block
Open sources
| Field | Value |
|---|---|
| Name | Electric Plant |
| Tagline | Plant intelligence technology decoding electrical signals for real-time health and growth insights. [Electric Plant, October 2026] |
| Headquarters | San Francisco, US [Electric Plant, October 2026] |
| Founded | 2023 [Crunchbase] |
| Stage | Seed [Simplify, October 2026] |
| Business model | Hardware + Software |
| Industry | Agtech |
| Technology | AI / Machine Learning |
| Geography | North America |
| Growth profile | Venture Scale |
| Founding team | Co-Founders (2), Justin Mast and Lee von Kraus [FinancialContent, October 2026] [LinkedIn] |
| Funding label | Seed |
| Total disclosed funding | $11.2 million [Simplify, October 2026] |
Links
Open sources
- Website: https://www.angel.co/
What an Investor Needs First
PUBLIC Electric Plant is building a hardware-plus-software system that reads electrical signals from plants and trees, then interprets them into real-time health, growth, and stress signals, a proposition that merits attention now because the company appears to have moved from research positioning into first commercial deployment and active hiring after a seed financing [Electric Plant, October 2026] [Built In, August 2026] [Simplify, October 2026]. Founded in 2023 and based in San Francisco, the company sits at the intersection of agtech, sensing hardware, and machine learning, with its early narrative centered on a practical question rather than a broad climate thesis: whether plant electrophysiology can become an operational data layer for growers and tree-care operators [Simplify, October 2026] [arxiv.org, October 2026].
The core product combines plant-sensing hardware, environmental sensors, cloud software, and a foundation model called Bombadil; the differentiating claim is non-invasive, continuous measurement of signals generated inside living plants, rather than reliance solely on soil, weather, or image data [Electric Plant, October 2026] [Everywhere Ventures job board, September 2026] [arxiv.org, October 2026]. Public technical evidence is still early, but the strongest signal is the October 2026 Bombadil preprint, which states the model was pretrained on more than 4 million hours of electrical signals from 100 plant species across ten environments and could track seven physiological indicators of stress, including cases without labels in training [arxiv.org, October 2026].
The founding team is legible enough for seed investors. Justin Mast is publicly associated with Bloomscape as founder and CEO, and prior coverage ties him to a multigenerational greenhouse family background; Lee von Kraus is publicly described as a neuroscientist with DARPA-linked work and is listed on LinkedIn as building Electric Plant [General Assembly] [Crunchbase] [TechCrunch, August 2019] [Jingletree] [LinkedIn].
On capitalization, the clearest disclosed figure is an $11.2 million seed round reported as closing in November 2025, with Onto Ventures and Climactic identified as leads and a broader syndicate that includes Everywhere Ventures, ReGen Ventures, and Michigan Rise [Simplify, October 2026]. The business model is best understood as hardware plus recurring software, although public materials stop short of giving pricing, contract structure, or retention data [Simplify, October 2026] [Electric Plant, October 2026].
Over the next 12 to 18 months, the key watch items are straightforward: whether the first orchard deployment converts into repeatable field performance, whether commercial users beyond an initial site are named publicly, and whether the company can translate intriguing model research into evidence of measurable operational value in plant and tree care [Everywhere Ventures job board, September 2026] [Built In, August 2026] [arxiv.org, October 2026]. The setup is promising, but the public record still leans heavily on company-linked materials and an early technical paper, so the central diligence question is execution, not imagination.
Claim stands unchecked -- This section relies materially on company-linked recruiting pages and company descriptions, with partial corroboration from Simplify, Built In, LinkedIn, and an arXiv preprint.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Business Model | Hardware + Software |
| Industry / Vertical | Agtech |
| Technology Type | AI / Machine Learning |
| Geography | North America |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding | $11.2 million seed disclosed [Simplify, October 2026] |
Inside the Company
Open sources Electric Plant appears to be an early-stage agtech company organized around a narrower proposition than most crop software startups attempt at seed: read electrical signals inside living plants, then turn those signals into operating insight for growers and tree-care users [Crunchbase] [Electric Plant, October 2026]. Public records place the company in San Francisco and date its founding to 2023, with Justin Mast and Lee von Kraus identified as founders in company materials and profile databases [Crunchbase] [Electric Plant, October 2026].
The public chronology is still sparse, but the sequence is reasonably clear. Crunchbase lists the company as founded in 2023 [Crunchbase]. Company materials available through its hiring presence indicate that by October 2026 Electric Plant was recruiting for machine learning, applied AI, and technical program roles, while describing a system that translates plant electrical signals into actionable insights [Electric Plant, October 2026].
What can be said with confidence is that the company is past the idea stage and presenting itself as a hardware-plus-software platform rather than a research project alone [Crunchbase] [Electric Plant, October 2026]. What cannot yet be established from the permitted public sources for this section is the legal entity name beyond the company branding used in profiles and hiring pages, or a fuller milestone record between founding and the 2026 hiring push.
Partially corroborated -- Confirmed in part by Crunchbase and company website materials; founding year, headquarters, and founder identities are directionally corroborated, but the milestone record remains limited in public filings and company-controlled sources.
Under the Hood
MIXED
The product claim is straightforward, even if the public evidence base is still thin. Electric Plant says it is building plant-intelligence technology that reads electrical signals from plants and trees, combines those signals with environmental data, and converts them into real-time information about plant health, growth, and stress [Electric Plant, October 2026] [arxiv.org, October 2026]. Public materials describe the system as a combination of plant-sensing hardware [PUBLIC], environmental sensors [PUBLIC], cloud software [PUBLIC], and a foundation model called Bombadil [PUBLIC], with the operating premise that plant electrophysiology can provide an earlier or more direct signal than soil, weather, or visual readings alone [Electric Plant, October 2026] [Everywhere Ventures job board, September 2026] [arxiv.org, October 2026].
The most concrete technical detail comes from the Bombadil research record rather than the company website. According to an October 2026 arXiv paper, Bombadil was pretrained without labels on more than 4 million hours of electrical signals collected from 100 plant species across ten environments, and the model tracked changes in seven physiological stress indicators, including on species and stressors without labels in training [arxiv.org, October 2026]. Hiring pages also indicate a live deployment context rather than a lab-only system: one job posting says the company was working with its first real customer in an orchard, while career listings in October 2026 show active hiring for machine learning, applied AI, and technical program management roles [Everywhere Ventures job board, September 2026] [Electric Plant, October 2026]. The likely stack beyond that, including agent systems and production ML orchestration, is best treated as provisional and should be read only as inferred from job postings (inferred from job postings) [Electric Plant, October 2026] [freehire.me].
Partially corroborated -- Product architecture is supported by company materials and job postings, with Bombadil details corroborated by arXiv. Several operating claims still rely on company-linked sources rather than independent product validation.
Market Research
PUBLIC
The market matters now because Electric Plant is trying to turn a biological signal that has mostly lived in research settings into an operational input for growers and tree-care operators, at a time when agriculture is under pressure to make earlier, more precise decisions with less labor and tighter environmental constraints [arxiv.org, October 2026] [TechCrunch, October 2026].
The public record does not support a direct TAM, SAM, or SOM for plant electrophysiology, so the cleaner approach is to work outward from adjacent markets rather than imply precision that is not there. Electric Plant's own positioning points first to plant and tree care, then to agriculture, environmental monitoring, and research [Electric Plant, October 2026]. TechCrunch's October 2026 coverage and the Bombadil preprint both frame the product around interpreting electrical signals from living plants into health, growth, and stress indicators, which places the company at the intersection of agtech sensing, crop intelligence software, and plant-science instrumentation rather than in a single established category [TechCrunch, October 2026] [arxiv.org, October 2026].
Demand, if it forms, is likely to come from use cases where visual inspection or ambient data arrives too late. The company's core claim is that internal electrical activity can reveal plant stress and growth dynamics continuously, while Bombadil was pretrained on more than 4 million hours of signals from 100 plant species across ten environments and is reported to track seven physiological stress indicators, including cases without labels in training [arxiv.org, October 2026]. That does not establish commercial performance, but it does clarify the wedge: earlier biological detection, if it generalizes outside research and pilot settings, could sit upstream of irrigation, crop protection, arborist intervention, and greenhouse operations [TechCrunch, October 2026] [Electric Plant, October 2026].
The nearest substitutes are already familiar to buyers. Growers and plant operators can infer stress from soil sensors, weather feeds, imaging, manual scouting, and standard environmental controls, while research institutions can buy plant-science instrumentation without adopting a vertically integrated hardware and model stack [Electric Plant, October 2026] [arxiv.org, October 2026]. Electric Plant therefore does not need to create demand for plant monitoring from scratch, but it does need to prove that electrophysiology adds signal quality that existing sensor categories do not already capture at lower cost and lower workflow friction.
Macro and regulatory forces cut in a broadly favorable direction, even if the evidence here is indirect. More volatile weather, water management pressure, and the rising cost of field labor all increase interest in tools that can identify stress earlier and help operators target interventions more precisely, while urban forestry and environmental monitoring create a second path beyond row-crop agriculture [TechCrunch, October 2026] [Electric Plant, October 2026]. The caution is that no public source in this record shows procurement cycles, reimbursement-like support, or regulation-driven adoption, so timing still depends on whether customers view the product as mission-critical instrumentation rather than an experimental sensing layer.
| Segment claim | Evidence |
|---|---|
| Core commercial wedge: plant and tree care | Company positioning identifies plant and tree care as the initial market [Electric Plant, October 2026] |
| Adjacent expansion path: agriculture | Company materials cite broader ambitions in agriculture [Electric Plant, October 2026] |
| Adjacent expansion path: environmental monitoring | Company materials cite environmental monitoring as a target category [Electric Plant, October 2026] |
| Research market relevance | Bombadil is presented in a scientific preprint grounded in multi-species electrophysiology data [arxiv.org, October 2026] |
The table underscores the current state of the opportunity: there is a visible category map, but not yet a publicly sourced market-sizing stack. For now, the more useful question is not absolute TAM, but whether Electric Plant can establish one beachhead where biological signal quality changes operating decisions often enough to justify hardware deployment.
Claim stands unchecked -- This section relies mainly on company positioning and a single research preprint, with no named third-party market-sizing report in the provided public sources.
Competition and Substitutes
MIXED Electric Plant appears to sit at the edge of several existing markets rather than inside a settled category: plant and tree care instrumentation, broader ag sensing, and applied AI for biological signal interpretation [Electric Plant, October 2026] [arxiv.org, October 2026].
The immediate map is easier to describe by substitute than by direct rival, because the source set does not name a single company pursuing the same plant-electrophysiology stack [Simplify, October 2026]. In practice, Electric Plant is competing first against incumbent workflows that infer plant health from soil probes, weather data, visual inspection, and periodic lab testing, and second against challengers building more software-heavy crop intelligence products that do not own a proprietary signal layer [Electric Plant, October 2026] [arxiv.org, October 2026]. That matters because buyers in orchards and commercial plant care do not need to agree that plants "speak" electrically to reject the product; they only need a cheaper or simpler way to answer irrigation, stress, or growth questions with acceptable accuracy [Everywhere Ventures job board, September 2026].
The company does have one credible edge on paper today, and it is data rather than channel. Bombadil was reportedly pretrained without labels on more than 4 million hours of electrical signals from 100 plant species across ten environments, which, if reproducible in field performance, suggests a dataset moat that would be difficult for a new entrant to assemble quickly [arxiv.org, October 2026]. The founding mix also fits the problem unusually well: Justin Mast brings commercial plant domain exposure through Bloomscape, while Lee von Kraus's public record points to biosignal and neuroengineering work that is directionally relevant to interpreting weak biological signals [Crunchbase] [General Assembly] [Jingletree]. The durability of that edge is still unproven. Dataset advantages in applied AI tend to persist only if the company keeps collecting production data from real deployments and can tie those signals to outcomes buyers care about, rather than to model novelty alone [Everywhere Ventures job board, September 2026] [arxiv.org, October 2026].
Exposure is just as clear. Electric Plant does not appear, from the public material reviewed, to own a large installed distribution channel in commercial agriculture, and the first customer context cited publicly is an orchard rather than a broad multi-crop network [Everywhere Ventures job board, September 2026]. That leaves the company vulnerable to adjacent incumbents with established agronomy relationships, even if those incumbents have a technically thinner signal stack. It is also exposed to category disbelief: non-invasive, continuous plant electrophysiology is an intellectually interesting wedge, but buyers may benchmark it against simpler sensor packages or crop-advisory software that can be trialed with less operational change [Electric Plant, October 2026]. Put differently, the main competitive risk may not be a named startup outrunning Electric Plant. It may be the market choosing "good enough" proxies over a deeper but newer modality.
The most plausible 18-month scenario is a split between technical validation and commercial concentration. Electric Plant is the likely winner if the orchard deployment turns into repeatable evidence that electrical signals improve operating decisions in ways growers can measure, because that would strengthen both its proprietary dataset and its sales narrative at the same time [Everywhere Ventures job board, September 2026] [Built In, August 2026]. The likely loser if that does not happen is Electric Plant itself, not because the science is disproven, but because adjacent sensing and farm-software vendors can continue selling established workflows while the company carries the burden of educating the market [Electric Plant, October 2026] [Simplify, October 2026]. With no named direct competitor confirmed in the source set, the competitive question remains less about share loss to a specific rival and more about whether Electric Plant can define a category before substitute products close the buyer's problem well enough.
Reasoned from indirect evidence -- This section relies on public sources for Electric Plant's product, hiring, and technical claims, but the source set names no direct competitors and much of the market mapping is analytical inference from company, recruiting, and arXiv materials.
Opportunity
PUBLIC
The prize here is not a better plant sensor, it is the possibility that plant electrophysiology becomes a new data layer for commercial growing and tree care, with Electric Plant positioned as one of the first companies trying to turn that layer into operating software and model infrastructure [arXiv, October 2026] [Electric Plant, October 2026] [Simplify, October 2026].
The headline opportunity is straightforward: if the company can show that electrical signals inside plants predict stress, growth, and health earlier or more reliably than visible inspection or standard environmental monitoring, it could become the system of record for plant operations where asset health is directly tied to yield, survival, or labor cost [arXiv, October 2026] [Electric Plant, October 2026]. That outcome is still early, but it is not purely aspirational in the sources. The public record shows three things that matter: a product architecture that combines hardware, environmental sensing, cloud software, and a model called Bombadil [Electric Plant, October 2026]; a research claim that Bombadil was pretrained on more than 4 million hours of plant electrical signals across 100 species and 10 environments [arXiv, October 2026]; and evidence that the company had moved from research framing to a first commercial deployment by 2026, with orchard use cited in recruiting material [Simplify, October 2026] [Everywhere Ventures job board, September 2026]. For a company founded in 2023, that sequence suggests a real attempt to build category infrastructure, not only a lab project [Crunchbase] [Simplify, October 2026].
The upside paths are narrower than the rhetoric around AI-for-biology might imply, but they are legible enough to map.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Orchard operations system | Electric Plant proves value in permanent crops, then expands from one deployment into a standard monitoring layer for orchards and managed tree assets | A reference deployment in an orchard shows measurable operating value and drives repeat sales into adjacent growers or service providers [Everywhere Ventures job board, September 2026] [Simplify, October 2026] | Public hiring materials place the first real customer in an orchard and describe the system as operational, not purely experimental [Everywhere Ventures job board, September 2026] [Electric Plant, October 2026] |
| Foundation-model layer for plant health | Bombadil becomes the model layer other hardware, research, or ag software groups build on for plant stress interpretation | Publication and external validation of the model's cross-species signal decoding creates technical credibility and integration demand [arXiv, October 2026] | The arXiv paper claims unlabeled pretraining at unusual scale and tracking across seven physiological stress indicators, which, if reproduced in the field, would support a software and API-style business in addition to device revenue [arXiv, October 2026] |
| Category-defining tree and landscape intelligence platform | The company starts in commercial plant and tree care, then broadens into urban forestry, environmental monitoring, and agriculture as the same signal stack is reused across use cases | One early vertical funds deployment learning, while the same sensing-plus-model stack is adapted to adjacent plant-health workflows [Electric Plant, October 2026] [Simplify, October 2026] | The company's own materials consistently frame the initial wedge as plant and tree care, with broader ambitions across agriculture, research, and environmental monitoring [Electric Plant, October 2026] [Simplify, October 2026] |
What compounding could look like is a data and deployment flywheel. Each installed system adds more electrophysiology and environmental data, more species coverage, more operating contexts, and more labeled or weakly labeled outcomes around stress and growth response, which in turn should improve model performance and make the next sale easier if the outputs are trusted by growers and operators [arXiv, October 2026] [Electric Plant, October 2026]. The early signs of that loop are visible, if only faintly, in the sources: the company is hiring across machine learning, applied AI, and technical program management after closing its seed round, which is consistent with a team moving from prototype to repeatable deployment and model iteration [Electric Plant, October 2026] [Simplify, October 2026]. Because the stack includes both hardware and software, there is also a plausible lock-in effect if historical plant signal baselines become useful for interpreting future anomalies at a given site [Electric Plant, October 2026].
The size of the win is harder to bound with public comparables because the company sits between agtech, sensing hardware, and applied AI, and no direct peer is named in the source set. The cleanest public anchor available here is its own reported $39 million valuation in 2026, which is only an early pricing point, not evidence of market size [PremierAlts, 2026]. If the orchard-operations or model-layer scenarios above were to play out, the company would need to grow from a seed-stage valuation into a business worth several hundred million dollars at minimum to count as category infrastructure, and potentially more if it became the default data layer for plant health across multiple end markets (scenario, not a forecast) [PremierAlts, 2026] [arXiv, October 2026] [Simplify, October 2026]. That is an ambitious outcome, but the public evidence does support the core reason to pay attention: Electric Plant is trying to industrialize a novel biological signal source with enough capital, technical specificity, and early commercial motion to make the upside legible.
Reasoned from indirect evidence -- This section relies on a mix of company materials, recruiting pages, one arXiv paper, and a single-source valuation datapoint from PremierAlts.
Sources
Open sources
[Electric Plant, October 2026] Careers | https://angel.co/
[Crunchbase] Electric Plant Company - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/electric-plant-company
[FinancialContent, October 2026] Electric Plant founder profile | https://www.financialcontent.com
[LinkedIn] Lee von Kraus | https://www.linkedin.com/
[Simplify, October 2026] Jobs and Employment at The Electric Plant Company | https://wellfound-1.angel.co/recruit/pricing
[Built In, August 2026] Staff Electrical Engineer | https://angel.co/rydar
[arxiv.org, October 2026] Bombadil preprint | https://arxiv.org/
[General Assembly] Justin Mast | https://generalassemb.ly/
[TechCrunch, August 2019] Bloomscape raises $7.5M to sell you plants of all sizes | https://techcrunch.com/2019/08/01/bloomscape-series-a/
[Jingletree] Lee von Kraus profile | https://jingletree.com/
[Everywhere Ventures job board, September 2026] Applied AI Engineer, Agent Systems | https://angel.co/company/helloskip/jobs/2169653-product-manager-for-mission-driven-startup
[TechCrunch, October 2026] Exclusive: Plants can speak using electricity, so this startup built an AI model to decipher their language | https://techcrunch.com/2026/10/20/plants-can-speak-using-electricity-so-this-startup-built-an-ai-model-to-decipher-their-language/
[freehire.me] Applied AI Engineer, Agent Systems | https://freehire.me/jobs/applied-ai-engineer-agent-systems-the-electric-plant-nat3c5yd
[PremierAlts, 2026] Electric Plant Co Private Stock Price & Valuation ($39M) | https://www.premieralts.com/companies/electric-plant-co
Articles about Electric Plant
- After 4 Million Hours of Plant Signals, a Startup Decodes Orchard Stress — The San Francisco startup, backed by $11.2 million, is deploying its first commercial system to interpret the language of plant electrophysiology.