Ground intelligence
Raise confidence in what the ground holds and the reserve base grows without a metre of new excavation. Marginal deposits become viable. Waste stops being moved. Ore stops being lost to dilution. That is the cheapest tonne in mining, and it is won with interpretation, not equipment.
Mimir is the layer that raises it — hardware-agnostic, running on data from any rig, sampler or sensor stack already on site.
01 — The opening
The EU Critical Raw Materials Act, US IRA and DPA allocations, and accelerated Nordic and Commonwealth permitting all set the same trap: public capital and political will are committed, but every target eventually resolves to a physical question — which deposits are viable, at what grade, under what ground conditions, and at what cost to the people nearest them. Nobody can answer it at the speed the policy assumes.
Mining already has the autonomous fleets, the automated rigs and the sensor coverage. The binding constraint was never hardware. It is that nobody trusts the interpretation enough to act on it in the shift, and so the decision waits for the assay lab.
02 — Evidence
Grade, ground, structure and reporting — the questions an operator has to answer before a shift, a permit or a resource statement. Performance is discussed under NDA against your own ground, not published here.
Ore grade estimation
Mining
Rock classification from imagery
Mining
Fracture zone mapping
Mining · Tunnelling
Underground development optimisation
Tunnelling
Ground bearing capacity
Construction & Geotechnics
Regulatory and ESG reporting
Mining · Construction & Geotechnics
Validation status, methodology and dataset detail differ by algorithm and are shared under NDA as part of an evaluation, against your ground rather than ours. We do not publish portfolio-level accuracy figures.
On evidence
We do not publish accuracy figures, benchmark comparisons or technology readiness levels for this portfolio. Validation status differs by algorithm: some are supported by measured results on real data, others are in development and not yet independently validated. An internal review of the validation record is under way, and until it concludes we would rather say less here than assert a figure we cannot stand behind line by line. Per-algorithm evidence, methodology and datasets are shared under NDA and should be assessed against your own ground before informing any operational, financial or reporting decision.
03 — Coverage
Grade estimation is commodity-specific — a model tuned on porphyry copper does not transfer unchanged to a rare-earth deposit. Ours are trained and benchmarked across five base and precious metals.
Exploration and deposit targeting classify against the DOE, DOD and EU critical minerals lists — lithium, cobalt, nickel, graphite, the rare earths and the rest of the strategic set. That is targeting coverage today and the next set of benchmarked grade models tomorrow.
04 — Why it compounds
Not tied to a rig, a sampler or an OEM. Every site already instrumented is a site we can run on, which makes distribution a software problem rather than a capital one.
Grade models improve with assay diversity, and ground models with condition diversity. Coverage compounds into accuracy that a single-site competitor cannot reach.
Spectral composition and ground-condition inference work the same way when the orebody is regolith. Research-stage, and a second buyer universe that costs us nothing to hold.
05 — The standard
Supply security is usually argued as volume. The harder and more durable version is precision: higher grade confidence means less waste moved, live ground and tailings monitoring means fewer catastrophic failures, and consultation risk understood early means communities are engaged before a permit is contested rather than after. As a public benefit corporation, that is the standard this is built to — supply security that does not transfer its costs onto the ground, the water, or the people nearest to it.
Investors
Portfolio detail, evidence base and the shape of the round are shared directly with named recipients.
Request accessOperators
Every deployment is assessed against your geology, drilling method, safety requirements and decision workflow before anything is promised.
Discuss a field pilot