Mining Needs AI Built for the Real World. IIT Kharagpur Is Testing It.

A new research Centre at India’s first institute of technology, funded by mining investor Vikram Sodhi, will use real mine data to develop AI tools for exploration, planning, maintenance, safety, and environmental monitoring.

Mining generates vast amounts of geological, operational, equipment and environmental data, but many of its most important decisions still depend on fragmented systems. Exploration, mine planning, equipment maintenance, processing and environmental monitoring often operate in separate data environments, even though the underlying risks are deeply connected underground and across the mine site.

Artificial intelligence could help close that gap, but only if it is built for the realities of geology and mining. Generic AI tools layered onto old systems are unlikely to be enough. The industry needs domain-specific AI — trained and tested against real mine data, field conditions and environmental constraints — if it is to make measurable improvements in safety, productivity and environmental performance. That premise is now being tested at a new IIT Kharagpur research Centre funded by Indian-origin mining investor Vikram Sodhi.

The Centre at the Indian Institute of Technology Kharagpur, the first of India’s IITs, has been established with a reported five-year commitment of 150 million Indian rupees, or approximately $1.8 million. The Vikram Sodhi Centre of Excellence for AI-Enabled Geological and Mining Systems will develop AI tools across the mining process.

The funding comes from Vikram Sodhi, founder of the Sodhi Foundation and Vice Chairman of Mineros S.A., a gold producer that is also the Centre’s inaugural industrial partner.

IIT Kharagpur is positioned for this kind of work in part because mining AI requires more than computer science. It requires geology, geophysics, mining engineering, metallurgy, data science and field knowledge to be integrated into one research platform. The model follows a broader trend among the IITs: industry-supported, interdisciplinary research units focused on applied technology. A well-known precedent is the Robert Bosch Centre for Data Science and AI at IIT Madras, launched in 2017 with multi-year industry funding.

The Centre’s research will focus on five connected areas: exploration, mine planning, mineral processing, predictive maintenance, and environmental, social and governance (ESG) analytics, with initial work concentrated on environmental monitoring and the governance frameworks around it. The goal is not to build isolated tools, but to connect data from across the mining process — from geological modeling to equipment performance to water-quality monitoring — into an integrated decision-support system.

A defining feature is the Industry-Linked Grand Challenges program, in which mining companies provide real operational data as the basis for annual competitive research. Teams develop deployable AI models for specific, operationally defined problems. The emphasis on real mine data — rather than relying only on synthetic or classroom datasets — is intended to keep the research relevant to active mining environments. It will also require clear rules on confidentiality, data rights, model validation, publication and environmental reporting.

The Centre’s first industrial partnership is with Mineros S.A., where Sodhi serves as Vice Chairman. Under a “Living Laboratory” model announced by Mineros in April 2026, selected operations in the Bajo Cauca region of Antioquia, Colombia, will be available for field validation of AI-based mining tools developed at the Centre.

The collaboration is expected to support exchanges among researchers, engineers and doctoral students in India and Colombia. Initial environmental analytics work will focus on water quality, particulate emissions, geotechnical stability, tailings performance, biodiversity indicators and carbon intensity per metric ton produced.

Governance will be important. The Centre operates within IIT Kharagpur, reports to the Dean of Research and Development, and is led academically by faculty. Its external advisory board provides strategic guidance but does not control research decisions. Industry partners may provide data and field problems, but research direction and academic standards remain with IIT Kharagpur faculty. The Centre’s ambitions will depend on whether it can attract researchers who combine geological domain knowledge with advanced computational skills — a difficult combination to find — and whether models developed in academic settings can perform reliably in active mining environments.

Over the next five years, the Centre should be judged by the quality of its published research, the number of models tested in field conditions, and measurable improvements in safety, maintenance, planning and environmental monitoring. The potential of AI in mining is likely to be measured not by how many models are built, but by whether those models improve decisions in real operating environments — and make the industry more efficient, safer and more accountable.