Russia · China · CIS

Digital control and predictive analytics for primary resources

MAGMA is a group of industry companies built around one platform. We automate material accounting from the face, the well and the cutting area through to the delivered lot, and add a machine-learning layer that forecasts recovery, losses and regime deviations before they become fact.

MAGMA UNIFIED ACCOUNTING CONTOUR INJECTION WELL RECOVERY · FLOW, GRADE INJECTION OBSERVATION PRODUCTIVE HORIZON · LEACHING CELL
An in-situ leaching cell: injection and recovery wells, the productive horizon and the monitoring contour.

One chain in every industry

Material is extracted, moved, weighed and delivered. Uranium, tungsten, timber, potash and coal differ in extraction technology but share the same accounting chain — and precision is lost in the same places: at handovers between sites, shifts and owners.

Source FACE · CUTTING AREA · WELL · FIELD Movement TRANSPORT · STORAGE · TRANSHIPMENT Intake WEIGHBRIDGE · LABORATORY · ACCEPTANCE Reporting STATE · BUYER METERING POINTMETERING POINTMETERING POINT scale tolerancelosses and short weightgrade and quality PHYSICAL MATERIAL MOVEMENT MAGMA OPERATES HERE
  1. SourceFACE · CUTTING AREA · WELL · FIELD
  2. METERING POINTscale tolerance
  3. MovementTRANSPORT · STORAGE · TRANSHIPMENT
  4. METERING POINTlosses and short weight
  5. IntakeWEIGHBRIDGE · LABORATORY · ACCEPTANCE
  6. METERING POINTgrade and quality
  7. ReportingSTATE · BUYER

We describe metering points

Every handover between sites, shifts and owners gets an instrument, a responsible person and an allowed tolerance. Before that, arguing about losses is pointless.

We reconcile the balance

Extracted, moved, received — three figures for one lot. Anything beyond tolerance is recorded as a discrepancy and broken down by shift, section and machine.

We forecast the deviation

A machine-learning layer trained on the asset's history returns the recovery profile, solution migration risk and shipment anomalies before they reach the monthly report.

Why no one else has assembled this set of work

Engineering contractors cover individual technology layers. IT integrators build the accounting contour. Research institutions take a method as far as a report and stop. We are not aware of another team on the Russian market that holds the whole set — geotechnological model, volume and weight balance, ML forecast and language interface — and carries it through to direct pilots and commercial contracts with major subsoil users.

The full stack of layers

From instruments and the hydrogeological model through to forecasting and statutory reporting — in one contour, without stitching three contractors together.

Direct pilots, not presentations

We come on site and work on the asset's own data: commercial operation in forestry and ore mining, a deployed commercial pilot in the nuclear industry.

Scientific grounding

Methods pass external review: a conference paper on uranium and rare earth metals, two articles ready for publication, a scientific adviser with forty years in the industry.

Industry and scientific grounding

Calculation methods are reviewed by industry specialists and taken to external venues — this settles the competence question before the technical discussion begins.

Official research partner
河南理工大学
Henan Polytechnic University
China's first mining university · 1909Jiaozuo, Henan Province, China

The memorandum was signed on 16 July 2026. The university works on the platform's calculation core alongside us and validates the methodology we calculate with.

Calculation architecture

Site hydrogeology, leaching kinetics, control of technological solutions

ML approach

Framing the forecasting problems, feature sets, model quality metrics

Methodology validation

Peer review of the calculation approaches and joint publications

How the research contour works

Scientific adviser

Alexander Boytsov — geologist, forty years in the uranium industry. Adviser to the First Deputy Director General of Techsnabexport, co-author of the OECD/IAEA Uranium 2024 report and the WNA Nuclear Fuel Report.

Conference in Almaty

A paper on machine learning in the management of in-situ leaching fields — the Uranium and Rare Earth Metals conference, 12–15 May 2026.

Publications

Two articles prepared for peer-reviewed publication: forecasting of technological solution excursions, and an AI assistant for the management of well fields.

Start with a survey of one asset

Describe the asset: what you extract or harvest, which instruments are installed, where accounting comes together and which systems you report into. We will respond with a metering point scheme and a pilot outline.