Skip to content

Critical metals

Find it. Characterize it. Qualify it. Before anyone drills.

Active learning at every step. Lodestar tasks the hyperspectral capture and ranks the sites. Traverse samples the few that matter. Crucible designs and runs the chemistry and prices the route. The output is a Passport a buyer can underwrite.

Screening operatingMine loop in pilotSan Juan Mountains · real signatures · simulated targeting

The real bottleneck

The bottleneck is not finding the next deposit. It is process certainty.

Can an existing site produce qualified feedstock? What is actually in the material? What is the cheapest route to concentrate it? Those are data, modeling, and chemistry problems before they are mining problems.

The output

A Passport for every qualified source.

Composition, mineralogy, extractability, recommended route, and unit economics. Each field carries the evidence and the uncertainty behind it.

Orbit sets the prior. Field samples sharpen it. Lab results close it. Meridian stops spending on a field once it reaches target confidence, and moves the budget to the field that is still open.

Source Qualification Passport

Schematic · how evidence fills each field
At target
CompositionPending
MineralogyPending
ExtractabilityPending
RoutePending
Unit economicsPending
Orbit priorField sampleLab resultTarget
Problem
A program or operator holds an area with hundreds of historical workings, legacy assays that no longer match today's target commodity, and a sampling budget that covers a fraction of them.
Output
A ranked target list and a Passport per qualified source, with the evidence and uncertainty behind each field.
Method
Active learning chooses every measurement. Meridian tasks hyperspectral captures over the district, and Lodestar fuses them with multispectral imagery, lithology, terrain, and legacy assays. Traverse selects field samples by expected reduction in decision risk. Crucible characterizes them and screens extraction pathways in parallel.
Evidence
2,000+ sites screened across the western United States and 14 priority targets. Thorin Resources' permitted Camp Bird and Revenue-Virginius mines in Colorado's San Juan Mountains run the full loop, with physical assays checking the model.
Maturity
Screening is operating. The orbit-field-lab loop runs in pilot in one district. A new geology needs new validation.
Limits
Coactive does not mine, permit, or qualify material for a supply chain. It produces the evidence the people who do those things need.
Rock samples laid out on a tray for characterization
Play · 2:02
Summer 2026 · underground at the Thorin Resources mines
Thorin Resources partnershipRemote sensing over the Camp Bird and Revenue-Virginius mines, model-selected field samples, and onsite lab measurement in one feedback loop. One district, one commodity context. Results are not a basis for claims about other geologies until validated there.

Evidence

From a district of signatures to a route on foot.

2,000+
Sites screened
Western United States.
14
Priority targets
Ranked for field follow-up.
6
Permitted mines in the loop
Partner sites.
Days
Sample characterization
Days and weeks instead of months.
Sampling plan with targets ranked by information gain Orbit view of the district with critical-mineral hotspots ranked by signature strength Orbit view · signaturesSampling plan · information gain

Drag to compare. Left: goethite, jarosite, and hematite signatures ranked by strength and confidence across the district. Right: targets ranked by information gain, with the sites Traverse selected for the next campaign. Demonstration interface over real imagery.

Contact

Bring an area and a commodity.

We reply with whether the area is screenable from available imagery, what a screening would cover, and what it would not claim.