Russian Mining Giant Hands Over Metal Extraction to Neural Network

Norilsk Nickel has trained an AI system to control ore flotation—one of the most complex and vital stages in metal production.

In a major technological shift, Russian mining giant Norilsk Nickel is entrusting one of its most critical production stages to artificial intelligence. The company has deployed a neural network to manage flotation—the chemical process that extracts metals from ore and determines both the volume and quality of the resulting concentrate.

Previously, human operators manually adjusted parameters for flotation machines in a high-stakes, multi-variable environment. Now, the company's in-house 'Digital Optimizer' system processes massive datasets—tracking pulp mass and density, reagent levels, air flow intensity, and metal content—to automatically fine-tune machine performance in real time.

The AI model was trained for 18 months using historical production data. It not only runs the system for optimal metal recovery with minimal waste but can also predict operational shifts up to 15 minutes in advance.

Next up: computer vision. Developers plan to add visual feedback capabilities, allowing the system to 'see' the process in action and make even more precise adjustments.

Norilsk Nickel’s move reflects a broader trend in Russia’s industrial landscape, where neural networks are increasingly being integrated into high-stakes, high-complexity operations once managed entirely by humans.