Exploration has several clocks running at once: compiling records, surveying, interpreting results, securing access, drilling and waiting for assays. Technology can shorten individual steps and improve the decisions between them. Six companies illustrate the opportunity. Some supply measurements or analytical services; others explore and develop projects themselves.
What the technologies do
Octavia: linking geophysics to drill planning
Octavia describes an early-stage platform combining advanced geophysics and proprietary AI for discovery, expansion and drill planning, with a focus on gold, copper and uranium.
Its proposed role is to turn geological and geophysical information into targets and drilling plans, whether testing new prospects or extensions around known deposits. The potential saving is a shorter cycle between new information and a revised programme, with holes chosen to resolve the most important uncertainties.
Esper: screening large areas with hyperspectral imagery
Esper supplies mineral maps derived from hyperspectral satellite imagery. Hyperspectral sensors measure reflected light across many narrow wavelength bands. Different surface materials have characteristic absorption patterns, allowing analysts to map exposed minerals and alteration associated with mineral systems. USGS explains the underlying method.
Where surface exposure is informative, this can screen large tenement packages and focus fieldwork before teams mobilise. Deeper targets are inferred by combining surface observations with geology and other surveys. Image quality, ground cover and the minerals sought determine its usefulness.
Atomionics: gravity measurements and a three-dimensional subsurface model
Atomionics combines GRAVIO, a cold-atom gravity sensor, with its ORE-O interpretation system. Cold-atom sensing uses the behaviour of atoms to measure very small changes in gravity. The company describes mobile data collection and AI-assisted conversion of those measurements into three-dimensional geological models.
Gravity is useful where underground rocks have contrasting densities. It can help constrain structures, intrusions and concealed geological bodies, including beneath cover. The interpretation needs geological constraints because different underground density distributions can produce similar measurements. USGS research illustrates that modelling problem.
Faster acquisition and interpretation could narrow the search area and improve hole placement. A gravity model guides that test; assays establish the metal content.
ALS Geoanalytics: making old and new data work together
ALS renamed its GoldSpot services business Geoanalytics in 2024. The business combines geoscience, data science and machine learning.
Bringing fragmented maps, drilling, geochemistry and surveys together can reveal patterns, rank prospects and identify where limited new fieldwork would resolve an important gap. That can shorten interpretation and avoid duplicating earlier work.
There is a concrete historical success at New Found Gold’s Queensway project. In its 29 January 2020 release, GoldSpot linked combined historical-data compilation, new field mapping and machine-learning targeting to hole NFGC-19-01, which returned 19 metres at 92.86 grams of gold per tonne. The reported interval is a downhole length. New Found Gold’s COO credited the combined traditional and machine-learning work; GoldSpot also disclosed its financial interest in the company.
The reported result supports the combined programme of archived data, fresh observations and machine learning. It does not isolate the algorithm’s contribution from the geologists’ work or establish an economic mine.
KoBold Metals: a continuously updated exploration system
KoBold describes a system that organises historical geoscience records, collects new measurements and produces predictions with uncertainty attached. Field results feed back into the models. Its own sensors include aerial imaging equipment and drill-core photography.
From target generation through deposit definition, this approach can help identify what to measure next, make each drilling round more informative and reduce delays between evidence and action.
Mingomba in Zambia provides a substantial project milestone. Partner ZCCM-IH reported the launch of shaft sinking on 29 April 2026. Its 2023 annual-report extract records Mingomba’s origins in the Lubambe Extension Project. The achievement is the advancement of a major copper project using an integrated technical approach, building on an existing geological history.
EARTH AI: connecting prediction with field validation
EARTH AI combines target prediction, geological fieldwork and its own drilling capability. Keeping those activities connected is intended to shorten the feedback loop from a promising model to a physical test and then to the next decision.
Legacy Minerals’ 25 November 2025 ASX announcement documents field validation at Fontenoy. It reported 120 metres at 0.30 grams per tonne combined palladium, platinum and gold from 298 metres downhole in EFO7D, including 10 metres at 1.2 grams per tonne from 388 metres. The broad interval used no grade cut-off. These are exploration results establishing mineralisation, rather than a resource or an economic assessment.
How to stack the tools in a programme
An illustrative copper programme with historical records, exposed ground and younger cover could follow this sequence. It combines capabilities selectively; no vendor partnership or need to use every tool is implied.
-
Build one reliable starting dataset. Compile drilling, maps, assays and geophysics, checking coordinates, units and sources. Geoanalytics-style integration establishes the geological questions. The first gate is whether the evidence supports a coherent target concept.
-
Screen the region, then visit the priorities. Esper-type hyperspectral imagery, combined with existing geology, could focus mapping and sampling where surface exposure is useful. Field observations test the interpretation. Advance targets with supporting evidence; revise the others.
-
Collect the missing subsurface measurement. Choose the survey for the question. Atomionics-type gravity may constrain a concealed body with a density contrast; other targets need electrical, electromagnetic or magnetic data. Proceed when the combined model defines a worthwhile drilling test.
-
Drill in stages and update the model. Record what each hole should test. Use results, including misses, to revise subsequent locations and depths. Octavia’s planning, KoBold’s iterative approach and EARTH AI’s connected targeting and drilling illustrate different ways to organise this step.
-
Release the next tranche of capital against what was learned. Continue, change the test or stop as the evidence warrants. Bring metallurgy, infrastructure and development constraints into decisions as mineralisation becomes better defined.
The capital-allocation benefit is earlier concentration on better-supported targets. Identifying a weak prospect quickly can also save a much larger commitment later.
Run environmental baseline work alongside exploration
Some of the same measurement and data-management capabilities can support environmental work. USGS has used imaging spectroscopy for environmental site screening and watershed evaluation in mining districts. Its research also describes hyperspectral applications for monitoring extraction, remediation and reclamation.
Suitable imagery could establish surface-condition maps, focus field studies and support repeat monitoring. Shared spatial data could help specialists compare access routes, drill pads and sensitive areas earlier. These are possible uses of the capabilities, rather than services attributed to every vendor.
Parallel baseline work could identify missing evidence sooner and reduce redesign or repeat mobilisation. Qualified specialists need to design and validate the studies. Western Australia’s flora and vegetation guidance, for example, addresses survey preparation, design, sampling, analysis and reporting for environmental assessment.
Earlier, better evidence may support more complete applications and quicker responses to questions. Approval timing also depends on assessment, consultation, seasonal evidence requirements and project impacts. Faster evidence collection cannot guarantee a faster regulatory decision.
Why this matters for mineral supply
The IEA’s 2026 Critical Minerals Outlook projects a copper supply gap of 25% in 2035 on its assessment of the project pipeline, narrower than the previous year’s estimate as projects advance. That is a conditional outlook, which can change with demand, investment and delivery.
Technology could help by finding viable deposits earlier, improving extensions around known deposits and letting teams test more credible ideas within the same budget. Those gains contribute to supply as projects progress through resource definition, engineering, approvals, financing and construction.
Judge each tool by the decision it improves. Combining better measurements, existing data and disciplined field testing could produce more viable projects per exploration dollar and move them forward sooner. That is a practical contribution to mitigating future mineral deficits.
Updated 13 October 2026: expanded the technology explanations, programme design and environmental applications. Company capabilities and project results are attributed to their sources; proposed programme benefits are the desk’s analysis.