There is a place where the energy transition is measured in meters of rock, not megawatts of capacity. It is dark, damp, and historically documented with a clipboard, a pencil, and a geologist’s best guess. This is the wedge for GroundedAI, a Toronto-based startup that has spent eight years quietly building a full-stack platform to map the underground in 3D. The goal is simple: turn geological complexity into actionable data, replacing subjective sketches with standardized, AI-assisted models [GroundedAI.com].
The Wedge: From Clipboards to Cloud
GroundedAI’s product suite is a two-part system designed for the field and the office. In the tunnel, a geological engineer uses the Lithos app on an iPad to capture high-fidelity 2D and 3D data, essentially a specialized LiDAR scanner for rock faces [rockmasstech.com/lithos]. That data flows into the Geographer web platform, which aggregates it, generates reports, and integrates with downstream industry-standard modeling software like Deswik and Leapfrog [GroundedAI.com]. The company is developing automatic feature detection for elements like rock bolts and fractures, aiming to streamline the entire workflow from data collection to decision-making.
A Long Bet on a Hard Problem
Founded in 2016 by geological engineer Shelby Yee and Matthew Gubasta, both Queen’s University students at the time, GroundedAI represents a patient, capital-efficient bet [Forbes]. The company has raised a total of $2.21 million across 15 disclosed rounds, a funding pattern that suggests a steady grind of grants, non-dilutive capital, and small equity rounds [CBInsights]. Backers include the Mining Innovation Commercialization Accelerator (MICA), BDC Capital, and SOSV [CBInsights]. CEO Shelby Yee’s profile has risen in mining circles, evidenced by a 2022 award from Young Mining Professionals [Bloomberg, 2022]. The company reports 372 engineers on its early-access waitlist and is running an alpha/beta program with testers at unnamed underground operations [GroundedAI.com].
| Founder | Role | Background Note |
|---|---|---|
| Shelby Yee | Co-Founder & CEO | Geological engineer; 2022 YMP award winner [Forbes, Bloomberg]. |
| Matthew Gubasta | Co-Founder & President | Fellow Queen’s University alumnus [Forbes]. |
Where the Ground Could Shift
The bet is compelling but comes with the inherent friction of selling into conservative, safety-first industries. The risks are not about the technology's promise, but its path to widespread adoption.
- Sales cycle gravity. Selling enterprise SaaS to mining majors is a marathon. Pilots can be long, procurement cycles glacial, and the cost of being wrong is catastrophically high. GroundedAI must prove its platform reduces more risk than it introduces.
- The integration maze. The value proposition hinges on smooth integration with legacy modeling tools. Any friction in data handoff between Geographer and a mine planner’s Deswik suite could relegate the tool to a siloed novelty.
- The quiet traction. While a waitlist of 372 is a positive signal, the absence of named customer deployments or detailed case studies in the public record makes it difficult to gauge real-world efficacy and renewal motion.
The Unit Economics of Clarity
The ultimate metric here isn't just software adoption, it's the carbon and cost saved by reducing over-excavation, preventing collapses, and optimizing resource extraction. If a typical tunnelling project overalls by just 5% due to poor geological data, that's thousands of cubic meters of unnecessary concrete and steel, each with a massive embedded carbon footprint. A platform that shaves that margin down represents a direct decarbonization lever, turning geological uncertainty into a quantifiable climate win.
GroundedAI's real competition isn't another startup. It's the entrenched incumbent of habit: the clipboard, the paper map, and the institutional reluctance to change a process that has "worked" for decades. Their task is to prove that in the dark, data is safer than tradition.