AI Moves Onto the Heavy Industry Shop Floor
AI is turning Russia’s steel industry into a high-tech sector.
“Smart Vision” for Steel
An innovative “Smart Vision” system has been launched at the Amurstal steel plant in Komsomolsk-on-Amur. The hardware–software complex, built on artificial intelligence, is designed to control the quality of steel output. The technology analyzes a live video feed from cameras installed near an electric arc furnace, where temperatures reach 1,600 °C, and detects in real time the moment slag appears in the metal stream. As soon as the system recognizes early signs of slag, it immediately signals the operator to stop tapping. This prevents impurities from entering the ladle – a factor that directly affects both final product quality and equipment wear.
Previously, the decision to stop metal tapping relied on the subjective judgment of an experienced operator, inevitably carrying a risk of error. The introduction of an AI module fundamentally changes process control: intuition is replaced by objective data, and reaction by prediction and prevention. Thanks to machine learning, the system not only identifies slag but continuously refines its algorithms by learning from new data. As a result, quality control becomes not just more accurate, but more adaptive to changing production conditions.
Steel You Can Trust
AI here addresses a concrete industrial task with a measurable economic effect – reducing scrap rates, improving steel quality, and extending equipment service life.
The importance of the project goes far beyond a single shop floor. Amurstal produces steel used in the construction of residential and social infrastructure – from kindergartens to nuclear power plants – both in Russia and across Southeast Asia. In this context, quality is not merely a commercial metric, but a matter of safety. Digitizing key production stages allows the plant to meet the most demanding standards, strengthening trust among domestic and international customers alike.
Digitalization requires investment, skilled personnel, and close attention to information security. However, as metallurgists’ experience shows, these investments pay off. Russian companies are already demonstrating efficiency gains and product quality that meet international benchmarks.
Digital Solutions and Billions in Added Value
Amurstal’s project fits into a broader trend of digital transformation across Russia’s steel industry. In recent years, companies have actively deployed IT solutions ranging from process automation and the development of high-strength steels to the creation of digital twins for industrial equipment.
Severstal has implemented more than 60 AI solutions covering production, safety, environmental monitoring, and even corporate processes. In 2024, the economic impact of these projects exceeded 1 billion rubles (about $11 million). For example, reducing rolling pauses by just one second per slab on the 2000 hot rolling mill at the Cherepovets Steel Mill generated over 180 million rubles (around $2 million). An expert blast furnace system optimized iron ore preparation, adding another 97.6 million rubles (roughly $1.1 million) in profit.
The “Steel Vision” platform, based on computer vision, monitors product quality, worker safety, and equipment condition. It detects defects across multiple classes within 12 control systems. At the Yakovlevsky mining and processing plant, AI also monitors the safe operation of lifting equipment.
Magnitogorsk Iron and Steel Works is also investing heavily in digital technologies. In March 2025, the company launched a “smart” pipe accounting system capable of recognizing products with 99.6 percent accuracy. Automated inspection of sheet steel using machine vision and an AI-based temperature forecasting model are already improving efficiency and quality. Over five years, the economic effect of AI projects at the plant has exceeded 4.5 billion rubles (approximately $50 million).
EVRAZ NTMK is introducing intelligent billet heating control. A pilot project with CyberPhysics replaced manual regime adjustments for more than 120 combinations of steel grades and profiles with AI algorithms. This reduced natural gas consumption and delivered an annual effect of 300 million rubles (about $3.3 million) with investments of just 23.5 million rubles (around $260,000).
Metallurgy 4.0
Artificial intelligence has already become an integral part of modern Russian metallurgy. In the coming years, full automation of casting and rolling shops is expected, with AI managing the entire cycle – from raw material charging to product shipment – making Russian steelmaking more efficient, safer, and more competitive in the global market.
