Visualizing the Invisible: Using AI for 3D Modeling of Carbonatite Pipes for REE Discovery

Executive Summary

• Mining tech providers and REE companies announced late-2025 updates and pilots using AI to model carbonatite pipes, improving target definition and drilling efficiency (Seequent; Veracio; KorrAI).
• Government support accelerated, with new December funding calls and guidance intended to scale critical mineral discovery and processing in the U.S. and EU (U.S. DOE; European Commission).
• Recent research introduces neural implicit models and graph-based inversions that enhance 3D geological reconstructions from sparse data (arXiv; IEEE Xplore).
• Industry sources suggest AI-enabled 3D modeling can cut early-stage exploration cycles by 20-30%, improving capital efficiency for REE programs (McKinsey Metals & Mining analysis).

AI-Powered 3D Modeling Moves Into REE Exploration Workflows Over the past six weeks, mining technology vendors and rare earth producers have advanced AI-enabled 3D modeling to delineate carbonatite pipes—key hosts of REEs—by fusing geophysics, geochemistry, and drill-core imagery. Software and hardware updates are designed to build higher-fidelity subsurface models from sparse datasets, reducing uncertainty before expensive drilling campaigns. Announcements from platforms such as Seequent and Veracio, combined with pilots reported by exploration groups, underscore momentum as 2026 begins. Exploration specialists describe material gains from combining implicit geological modeling with machine learning at carbonatite prospects. Firms including AI-first explorer KoBold Metals and REE operators such as MP Materials and Lynas Rare Earths are expanding digital toolkits to integrate hyperspectral core imaging, magnetics, gravity, and geochemical signatures into unified 3D views that sharpen pipe boundaries and feeder zones, according to company materials and analyst coverage (Reuters mining coverage). New Releases and Pilots Announced in Late 2025 In December, geoscience software updates highlighted practical steps to bring AI into the loop for 3D modeling. Industry sources point to Seequent’s Leapfrog updates enhancing structural interpretation workflows and data fusion across geophysics and geochemistry, which are increasingly applied to REE-hosting carbonatites (Seequent Leapfrog). Hardware-enabled AI in scanning and logging also progressed, as Veracio expanded automated core imaging and analytics designed to accelerate orebody characterization and reduce sampling bias (Veracio news). On the exploration side, geospatial AI startup KorrAI has been active in late-2025 partnerships focused on critical minerals target generation, leveraging satellite, UAV, and ground datasets for model-ready inputs. Projects with REE developers such as Aclara Resources are centered on anomaly mapping and prospect-scale 3D modeling, according to company disclosures and investor presentations (Aclara investor materials). Subsurface modeling integrators like Mira Geoscience also reported late-year updates to machine learning workflows for geophysical inversion, which are increasingly referenced in REE targeting (Mira Geoscience news). Regulatory Tailwinds and Funding Signals Policy and funding signals in the last 45 days are reinforcing AI-driven exploration priorities for critical minerals. In December, U.S. federal programs issued calls and guidance aimed at strengthening the domestic supply chain for rare earths, highlighting upstream exploration and advanced processing—areas where AI-enabled modeling can compress timelines (U.S. DOE Technology Transitions; DOE Science & Innovation). The European Commission continued implementing the Critical Raw Materials Act, with late-2025 communications detailing support mechanisms for strategic projects and permitting streamlining across member states (EU Critical Raw Materials Act). These policy updates are reshaping exploration economics by potentially reducing permitting friction and providing co-funding for technology-led initiatives. Industry analysts note that targeted funding for data acquisition and AI modeling can raise discovery probabilities while lowering cost per target—particularly in complex geological settings like carbonatite pipe systems (McKinsey Metals & Mining analysis). This builds on broader Mining trends where digital twins and AI-assisted geoscience are increasingly embedded in feasibility and development workflows. Research Highlights: Neural Implicits and Graph-Based Inversions Recent peer-reviewed and preprint literature published in November–December 2025 introduces methods relevant to carbonatite modeling. Neural implicit representations and diffusion-based approaches show promise for reconstructing 3D geology from limited observations, while graph neural networks improve inversion of geophysical datasets (magnetics, gravity) frequently used to delineate carbonatite pipes (arXiv geological modeling; IEEE Xplore). These techniques help resolve pipe geometries, fenite aureoles, and feeder structures—critical for prioritizing drill targets. Applied case studies in late-2025 conference proceedings and technical notes spotlight multi-modal workflows: combining hyperspectral core imaging with ML classification to flag REE-bearing intervals; joint inversion of gravity and magnetics using deep learning to sharpen contrasts at depth; and probabilistic structural modeling to quantify uncertainty. Vendors including Seequent, Veracio, and Mira Geoscience reference these methods in product notes and training materials, aligning with the needs of REE producers such as MP Materials and Lynas for faster, data-driven decision-making (Seequent Learning; Veracio solutions). Company and Program Updates Snapshot

Organization	Announcement (Nov–Dec 2025)	Focus Area	Source
Seequent	Late-2025 Leapfrog enhancements	AI-assisted structural and data fusion	Product page
Veracio	December core-scanning analytics updates	AI imaging and ore characterization	Newsroom
KorrAI & Aclara Resources	Late-2025 pilot collaborations	Geospatial AI for REE target generation	Investor materials
Mira Geoscience	Year-end ML inversion updates	Geophysical joint inversion workflows	Company news
U.S. DOE	December calls/guidance on critical minerals	Funding and technology adoption	DOE TT
European Commission	Late-2025 CRM Act implementation updates	Strategic project support and permitting	CRM Act

{{INFOGRAPHIC_IMAGE}}

Commercial Implications and Outlook For REE producers and developers, AI-centered 3D models of carbonatite pipes promise faster cycles from anomaly to drill-ready targets and tighter capital discipline. Industry estimates suggest early-stage exploration timelines can be reduced by 20-30% when multi-modal datasets are ingested and reconciled via ML-driven implicit modeling, lowering false positives and improving hit rates (McKinsey Metals & Mining). This aligns with latest Mining innovations where data platforms, edge devices, and cloud compute converge in operational workflows. Expect 2026 to bring more integrated stacks: geophysical inversion tools paired with core-scanning AI; geospatial models that continuously update as new data arrives; and probabilistic uncertainty quantification embedded in investor reporting. Leading companies—software providers like Seequent and exploration groups such as KoBold Metals, alongside REE operators including MP Materials and Lynas—are positioned to operationalize these workflows based on this season’s announcements and pilots (Veracio news; Mira Geoscience updates). FAQs { "question": "What makes carbonatite pipes prime targets for rare earth exploration?", "answer": "Carbonatite pipes host elevated concentrations of rare earth elements due to mantle-derived magmas enriched in carbonates and incompatible elements. Their vertical, pipe-like geometry and associated fenite aureoles create distinct geophysical signatures, including gravity and magnetic contrasts. AI-driven 3D modeling integrates geophysics, geochemistry, and core imagery to better resolve pipe boundaries and feeder zones, improving target fidelity and drill planning. This approach is increasingly reported by vendors and REE operators in late-2025 updates and pilots (Seequent, Veracio, Mira Geoscience)." } { "question": "How did AI-related mining technology announcements change in the last 45 days?", "answer": "Late-2025 releases emphasized AI-assisted structural modeling, automated core-scanning analytics, and ML-driven geophysical inversion workflows. Seequent highlighted Leapfrog enhancements for data fusion; Veracio expanded imaging and classification features; and Mira Geoscience reported machine learning updates for joint inversion. Exploration-focused collaborations from KorrAI and Aclara Resources centered on geospatial AI to prioritize REE anomalies. These developments aim to reduce uncertainty and shorten pre-drilling timelines, according to company materials and analyst commentary." } { "question": "What are the commercial benefits of AI-enabled 3D modeling for REE projects?", "answer": "AI-enabled 3D modeling can compress early exploration cycles by 20–30%, industry sources suggest, by reconciling multi-modal datasets and quantifying uncertainty before drilling. For REE operators, this translates to fewer drill holes per discovery, lower cost per target, and faster movement from anomaly to resource definition. The approach also supports better stakeholder communications, as probabilistic models provide transparency about confidence intervals, with tools from Seequent, Veracio, and Mira Geoscience referenced in recent product notes." } { "question": "What policy and funding signals are relevant to AI adoption in REE exploration?", "answer": "December communications from the U.S. Department of Energy and the European Commission emphasized strengthening critical mineral supply chains, including upstream exploration. Funding calls and CRM Act implementation guidance aim to streamline permitting and support strategic projects, creating tailwinds for technology-enabled workflows. These signals encourage co-funding for data acquisition and AI modeling, potentially improving discovery probabilities and capital efficiency across REE portfolios in 2026." } { "question": "Which AI techniques are most promising for modeling carbonatite pipes?", "answer": "Neural implicit representations and diffusion-based models are promising for reconstructing complex geology from sparse inputs, while graph neural networks enhance inversion of magnetics and gravity—key for delineating carbonatite pipes. Hyperspectral core imaging combined with ML classification can quickly tag REE-bearing intervals, feeding more accurate models. Recent late-2025 papers and technical notes on arXiv and IEEE Xplore highlight these approaches, and vendors are integrating them into commercial workflows for REE targeting." } References

• Leapfrog Geological Modeling - Seequent, December 2025
• Product and Program Updates - Veracio, December 2025
• ML and Inversion Workflow Updates - Mira Geoscience, December 2025
• Geospatial AI for Exploration - KorrAI, December 2025
• Investor Materials - Aclara Resources, December 2025
• Critical Materials Technology Transitions - U.S. Department of Energy, December 2025
• Critical Raw Materials Act - European Commission, December 2025
• Implicit Neural Geological Modeling (search) - arXiv, November–December 2025
• Graph-Based Geophysical Inversion (search) - IEEE Xplore, November–December 2025
• Metals & Mining Perspectives - McKinsey & Company, December 2025