Earth AI

AI-powered mineral exploration company that discovers, develops, and owns critical mineral projects.

Cover Block

PUBLIC

Field	Detail
Company	Earth AI
Tagline	AI-powered mineral exploration company that discovers, develops, and owns critical mineral projects.
Headquarters	San Mateo, United States
Founded	2017
Stage	Series B
Business Model	Other (Project Development & Sale)
Industry	Cleantech / Climatetech
Technology	AI / Machine Learning
Geography	North America (Operations in Australia)
Growth Profile	Venture Scale
Founding Team	Solo Founder (Roman Teslyuk)
Funding Label	Series B
Total Disclosed Funding	~$40.98 million [CB Insights, 2026]

Executive Summary

PUBLIC Earth AI operates a vertically integrated, AI-powered mineral exploration business that identifies and develops critical mineral deposits, a model that directly addresses the extreme inefficiency and high cost of traditional discovery methods [Latitude Media, Jan 2025]. Founded in 2017 by Roman Teslyuk, a geoscientist pursuing a doctorate at the University of Sydney, the company emerged from Y Combinator with a thesis that software could drastically improve exploration outcomes [TechCrunch, 2019]. Its core wedge combines a proprietary Mineral Targeting Platform, which analyzes public data to locate prospects, with owned drilling rigs to validate those targets, creating a closed-loop discovery engine [Latitude Media, Jan 2025].

Teslyuk's eight years of mineral exploration and academic experience anchor the technical credibility of the operation, though the company remains lean with an estimated dozen employees [TechCrunch, 2019] [Y Combinator]. The business model involves discovering and proving deposits, then selling the validated assets to large mining companies, a path supported by a recent $20 million Series B led by Tamarack Global and Cantos Ventures, bringing cumulative funding to approximately $41 million [PR Newswire, Jan 2025] [CB Insights, 2026]. Over the coming 12-18 months, the critical watchpoints are the translation of its claimed 75% discovery success rate into commercial sales with named off-takers and the scaling of its Australian field operations beyond the six confirmed prospects [miningstockeducation.com, 2025].

Data Accuracy: YELLOW -- Core funding and founder details are well-sourced; key performance metrics are company-provided.

Taxonomy Snapshot

Axis	Value
Stage	Series B
Business Model	Other
Industry / Vertical	Cleantech / Climatetech
Technology Type	AI / Machine Learning
Geography	North America
Growth Profile	Venture Scale
Founding Team	Solo Founder
Funding	~$40,980,000

Company Overview

PUBLIC Earth AI is a San Mateo-based mineral exploration firm founded in 2017 by geoscientist Roman Teslyuk, who was pursuing a doctorate at the University of Sydney at the time [TechCrunch, 2019]. The company's formation was driven by Teslyuk's academic and field experience, aiming to apply data science to the historically inefficient process of finding mineral deposits. Its early development was supported by Y Combinator, which the company joined in 2019, providing initial capital and validation for its AI-driven exploration model [TechCrunch, 2019].

Key operational milestones have centered on proving its technology in the field. The company has confirmed the discovery of six new mineralized prospects in Australia, containing tungsten, cobalt, and gold, using its predictive platform [miningstockeducation.com, 2025]. A significant inflection point was the January 2025 closing of an oversubscribed $20 million Series B round led by Tamarack Global and Cantos Ventures, bringing its total disclosed funding to approximately $40.98 million [PR Newswire, Jan 2025] [CB Insights, 2026]. While headquartered in the United States, its primary exploration activities are concentrated in New South Wales, Australia [Latitude Media, Jan 2025].

Data Accuracy: GREEN -- Founding details confirmed by TechCrunch and Y Combinator. Funding totals corroborated by PR Newswire and CB Insights. Operational claims cited from mining industry publications.

Product and Technology

MIXED The company's core proposition is a vertically integrated workflow that begins with AI-driven targeting and ends with physical proof of concept, a structure designed to compress the timeline and cost of mineral discovery. Earth AI uses a proprietary Mineral Targeting Platform to analyze public geological data, identifying prospective deposits of critical minerals like tungsten, cobalt, and gold [Latitude Media, Jan 2025]. This software-based analysis is then followed by what the company calls its "wedge": deploying its own, proprietary mobile low-disturbance drilling rigs to test and develop the identified sites [mining-technology.com]. The business model is to own these validated mineral projects and sell them to large mining companies, a path cited as targeting firms like BHP or Rio Tinto [Latitude Media, Jan 2025].

Performance claims center on dramatically improved discovery economics. The company states its technology has improved discovery success rates from an industry standard of 0.5% to 66% [earth-ai.com/technology, 2026]. External reports and investor communications cite a more specific 75% discovery success rate while reducing exploration costs by up to 80% compared to traditional methods [LinkedIn, Ben Werdegar] [aimresearch.co]. The primary evidence for this performance is the confirmation of six new mineralized prospects in Australia [theaiinsider.tech, 2025]. The technology stack [PUBLIC] is described as an AI/ML platform, but specific model architectures or data partnerships are not detailed in public sources. A review of open roles suggests ongoing hiring for software engineering and IT support, indicating a continued build-out of the core platform (inferred from job postings) [LinkedIn, 2026].

Data Accuracy: YELLOW -- Key performance metrics are company-sourced or cited by investors; the discovery of six prospects is corroborated by multiple outlets.

Market Research

PUBLIC The market for critical minerals is defined less by traditional software TAM and more by the immense capital expenditure and strategic urgency of securing supply chains for the energy transition.

Third-party market sizing specific to AI-driven mineral exploration is not available in the captured sources. However, the context for Earth AI's business is framed by the broader market for critical minerals essential to clean energy technologies. The International Energy Agency (IEA) has projected that demand for minerals like lithium, cobalt, nickel, and copper could grow by as much as 400% to 600% by 2040 under net-zero scenarios, a figure often cited in industry analysis [TechCrunch, March 2025]. This demand is driven by the global build-out of electric vehicles, grid-scale battery storage, and renewable energy infrastructure, creating a multi-trillion-dollar upstream opportunity for resource discovery and development.

Demand tailwinds are structural and policy-driven. National security concerns over concentrated supply, particularly from China, have led to legislation like the U.S. Inflation Reduction Act, which includes incentives for domestically sourced or processed critical minerals [TechCrunch, March 2025]. This regulatory push, combined with corporate net-zero commitments from major automakers and tech companies, is forcing traditional mining giants and new entrants to accelerate exploration. The primary substitute for new discovery is recycling, but analysts note recycled material volumes will be insufficient to meet projected demand for decades, locking in the need for new primary supply.

Key adjacent markets include traditional mineral exploration services, dominated by large engineering firms, and the commodity trading of the minerals themselves. Earth AI's wedge targets the exploration phase, a segment historically characterized by high capital intensity and low success rates. A shift in this segment's economics, even on a small number of projects, could command significant value. Macro forces include geopolitical tensions affecting resource nationalism, environmental permitting timelines, and the volatility of commodity prices, which directly influence the capital available for greenfield exploration projects.

Industry Avg. Discovery Success | 0.5 | %
Earth AI Claimed Discovery Success | 75 | %
Earth AI Claimed Cost Reduction | 80 | %

The visual underscores the core inefficiency Earth AI aims to address. While the company's claimed metrics are impressive, they represent a potential order-of-magnitude improvement over an industry baseline that has remained stagnant for decades, highlighting the scale of the opportunity if the technology proves scalable.

Data Accuracy: YELLOW -- Market context is supported by cited industry reports, but specific TAM for the company's niche is not publicly quantified.

Competitive Landscape

MIXED Earth AI operates in a niche defined by the application of data science to mineral exploration, a space where direct, like-for-like competitors are rare but where competitive pressure comes from established industry incumbents and adjacent technology providers.

Given the absence of named, direct competitors in the structured sources, a formal comparison table is omitted. The competitive analysis proceeds with a mapping of the broader field.

Traditional mineral exploration is dominated by large, integrated mining companies (e.g., BHP, Rio Tinto) and junior exploration firms. Their primary competitive advantage is scale, geological expertise, and ownership of vast land positions. However, their exploration methodology is often characterized by high-cost, low-success-rate campaigns based on conventional geological models. Earth AI's wedge is not to compete for land but to introduce a fundamentally different, software-driven discovery process that aims to be both cheaper and more accurate. The most direct competitive threat in this segment would be if a major miner developed or acquired equivalent AI capabilities in-house, though such efforts have historically been slow to materialize at the core of exploration workflows.

Adjacent competition comes from geoscience software and data providers, such as Seequent (a Bentley Systems company) or mining-focused divisions of larger GIS firms. These companies sell tools for geological modeling and data visualization but typically stop at the software layer; they do not engage in physical drilling or claim ownership of discovered resources. Earth AI's vertical integration,combining targeting software with proprietary drilling and project development,places it in a different business model category. Its exposure here is that these established software vendors have deep customer relationships and could theoretically expand their offerings downstream, though moving into capital-intensive field operations represents a significant strategic leap.

Where Earth AI appears to have a defensible edge today is in its proprietary dataset and integrated feedback loop. The company's Mineral Targeting Platform is trained not just on public geological data but, crucially, on the results of its own drilling campaigns. This creates a closed-loop system where each drilled prospect, whether a success or failure, improves the AI model. A traditional software vendor lacks this direct, high-fidelity field validation data. This edge is durable as long as Earth AI continues to execute drilling programs and retains exclusive rights to the resulting data. However, it is perishable if the underlying algorithms become commoditized or if a well-funded competitor replicates the data-acquisition cycle.

The company is most exposed in the capital and execution domain. While it has raised significant venture funding, its model requires continuous capital deployment for drilling to both prove resources and generate training data. A downturn in funding availability or a series of unsuccessful drills could stall this flywheel. Furthermore, its go-to-market strategy for selling validated deposits relies on relationships with the same large miners who are its potential competitors. A lack of a proven track record of commercial sales to these entities, beyond the cited archetype, remains a key vulnerability.

The most plausible 18-month competitive scenario hinges on proof of commercial scale. If Earth AI can successfully transition one of its six discovered Australian prospects into a monetizable asset sale to a major like BHP or Rio Tinto, it would validate the entire model and likely attract further capital and partnership interest. In this scenario, Earth AI becomes the winner, positioned as a new category leader in tech-enabled resource generation. Conversely, if commercial deals fail to materialize and the capital-intensive drilling cycle slows, the loser would be Earth AI's current venture-scale thesis, potentially forcing a pivot to a capital-light software licensing model and ceding the integrated advantage to any player with deeper pockets and patience.

Data Accuracy: YELLOW -- Competitive mapping is inferred from industry structure and company positioning; no direct competitor profiles are publicly cited.

Opportunity

PUBLIC If Earth AI can scale its vertically integrated model, the prize is a fundamental reordering of the capital-intensive, high-risk business of discovering the critical minerals required for the energy transition.

The headline opportunity is for Earth AI to become the first technology-led owner-operator of a portfolio of proven mineral assets, effectively acting as a discovery engine and project incubator for the mining industry. This outcome is reachable because the company's model bypasses the traditional exploration risk that deters capital. Instead of selling software or services, Earth AI uses its proprietary Mineral Targeting Platform to identify high-probability targets and its own drilling capability to prove them, aiming to sell validated deposits to large miners [Latitude Media, Jan 2025]. The evidence of six confirmed new mineralized prospects in Australia, including tungsten, cobalt, and gold, demonstrates the model can produce tangible, saleable assets [miningstockeducation.com, 2025]. By owning the discovery and development process end-to-end, the company captures value across the entire early-stage risk curve, a position no pure software or traditional junior explorer currently holds.

Growth could follow several concrete paths, each with identifiable catalysts.

Scenario	What happens	Catalyst	Why it's plausible
Project Portfolio Scale	Earth AI systematically discovers and proves a pipeline of 50+ mineral prospects across multiple jurisdictions, becoming a preferred source of “turn-key” projects for majors.	A first major asset sale to a named miner like BHP or Rio Tinto, validating the commercial model and attracting project financing.	The company is already operating in mineral-rich Australia and has confirmed six prospects; the Series B capital provides runway to expand drilling campaigns [Latitude Media, Jan 2025] [PR Newswire, Jan 2025].
Technology Licensing	The Mineral Targeting Platform is licensed to other junior explorers or mining majors, creating a high-margin software revenue stream alongside the core project business.	Publication of a third-party, peer-reviewed study validating the AI platform's predictive accuracy against historical exploration data.	The company's claimed discovery success rate of 75% (versus a 0.5% industry average) is a powerful marketing tool, even if self-reported, suggesting underlying technical efficacy [LinkedIn, Ben Werdegar].
Geographic Expansion	The integrated model is replicated in other geopolitically stable regions with high mineral potential but under-explored geology, such as Canada or Scandinavia.	Securing a strategic investment or joint venture with a regional mining champion or sovereign wealth fund.	The model is geology-agnostic and driven by public data; successful proof in Australia provides a blueprint for new frontiers [TechCrunch, Mar 2025].

Compounding for Earth AI is a data and capital flywheel. Each drilling campaign, whether successful or not, generates proprietary subsurface data that refines the AI targeting models, improving future success rates and lowering costs per discovery. This creates a data moat that widens with each project. Concurrently, a track record of successful asset sales or partnerships would lower the cost of capital for future exploration campaigns, enabling larger, more ambitious projects. There is early, though indirect, evidence this flywheel is starting: the company's ability to raise an oversubscribed $20 million Series B from institutional investors like Tamarack Global suggests confidence in its ability to deploy capital effectively [PR Newswire, Jan 2025].

The size of the win, in a successful project portfolio scenario, can be framed by looking at the value of proven mineral resources. While direct comparables are scarce, junior mining companies with a single advanced-stage project can achieve market capitalizations in the hundreds of millions of dollars. Earth AI's ambition to build a portfolio of such assets, discovered at a fraction of the traditional cost, points to a potential enterprise value that could reach the low billions if it can consistently produce and monetize high-quality projects. This is a scenario-based outcome, not a forecast, but it illustrates the magnitude of the opportunity inherent in improving the economics of mineral discovery.

Data Accuracy: YELLOW -- The core opportunity thesis is built on company claims of technical efficacy and a confirmed operational footprint, but key commercial milestones (asset sales, licensed software revenue) remain uncorroborated by independent sources.

Sources