Insights
Beyond the Hype: What AGI Actually Means for the Future of Mining
18 minute read
18 minute read
•
•
Cuts through the market projections to examine what advancing AI capabilities (particularly agentic systems) mean operationally and strategically for mining
Cuts through the market projections to examine what advancing AI capabilities (particularly agentic systems) mean operationally and strategically for mining
There is no shortage of breathless predictions about artificial intelligence in mining. Market reports project AI spending in the sector growing from $2.7 billion in 2024 to $13.1 billion by 2029. Headlines promise autonomous operations, intelligent exploration, and AI-driven decision-making at every stage of the value chain. But beneath the forecasts and the vendor pitches, a more interesting, and more difficult, question is taking shape: what happens when AI moves beyond narrow task automation toward something approaching general intelligence?
This is a question mining executives should be engaging with now, not because artificial general intelligence (AGI) is imminent, but because the trajectory of AI development is already reshaping what is strategically possible, and the industry's readiness gap is widening.
The current state: useful but fragmented
The honest assessment of AI in mining today is that it works, in pockets, but has not yet transformed how most companies operate. McKinsey's December 2025 analysis of innovation in mining describes an industry where AI is "moving from pilot projects to everyday practice," with leading miners applying deep learning to optimize fixed plants and using agentic AI to reinvent complex processes such as source-to-pay and issue identification. China's mining industry now operates nearly 4,000 autonomous vehicles, more than half of them electric, running continuously with AI-optimized routing.
Yet McKinsey's broader State of AI survey (November 2025) reveals a sobering pattern across all industries: 88 percent of organizations now use AI in at least one function, but only about one-third have scaled it beyond pilots. Only roughly 6 percent of companies attribute more than 5 percent of EBIT to AI. The problem is not the technology itself. It is that most organizations remain stuck in what many observers now call "pilot purgatory," launching isolated proofs of concept that never converge into redesigned workflows or enterprise-level capability.
Mining faces this challenge acutely. The sector's dispersed operations, legacy systems, fragmented data environments, and deeply siloed organizational structures make scaling AI harder than in more digitally native industries. The operational complexity is real: ore variability, equipment heterogeneity, remote sites, and regulatory constraints create an environment where generalizable AI solutions are difficult to deploy.
From narrow AI to agentic systems: the intermediate horizon
The more immediate and consequential shift is not the arrival of AGI in the theoretical sense, a machine that can reason across all domains the way a human can, but rather the emergence of agentic AI systems. These are AI architectures capable of planning and executing multi-step workflows autonomously, not just answering questions or making predictions but taking sequences of actions across interconnected processes.
McKinsey's 2025 research found that 62 percent of organizations are experimenting with AI agents, and 23 percent report scaling them in at least one function. However, in no single function has the "scaled" share exceeded roughly 10 percent. To make agents operationally useful, companies need standardized process steps, modern interfaces around legacy systems, and governance frameworks that make autonomous action observable and interruptible.
For mining, agentic AI has distinctive implications. Consider the mine-to-market value chain: extraction, processing, logistics, and sales are typically managed in organizational silos with limited real-time data exchange. An agentic system could, in principle, coordinate across these boundaries—adjusting blast schedules based on downstream processing constraints, rerouting logistics in response to grade variability, and rebalancing production plans against shifting commodity prices—all without waiting for human intermediation at each handoff point.
The concept of "Just in Time Mining," where integrated AI and potentially quantum computing enable real-time optimization of extraction based on demand signals and operational conditions, represents the far end of this trajectory. It would mean treating the mine not as a sequence of independent operations but as a single adaptive system.
What AGI would actually change
If we look beyond today's agentic systems to a future where AI capabilities approach genuine generality, the ability to reason across unfamiliar domains, handle ambiguity, and learn from limited examples, the implications for mining are profound but also more nuanced than most projections suggest.
Exploration and resource characterization. The most transformative near-term application may be in exploration, where AI is already combining satellite imagery, geophysical surveys, geochemical data, and historical drilling records to generate higher-confidence targets. A system with more general reasoning capabilities could integrate geological models with economic constraints, geopolitical risk, ESG considerations, and infrastructure availability to produce genuinely strategic assessments of where to explore, not just where minerals are likely to be, but where developing them makes sense.
Operational decision-making under uncertainty. Mining is fundamentally a business of managing uncertainty—in ore grades, commodity prices, equipment reliability, and regulatory environments. Current AI excels at pattern recognition within defined parameters. A more generally capable AI could help leaders reason about scenarios that fall outside historical distributions: novel geological conditions, unprecedented market structures, or regulatory shifts that have no close precedent.
Capital allocation and portfolio strategy. Perhaps the most underexplored area. Mining companies make capital decisions with 20- to 30-year horizons based on necessarily imperfect projections. A system that could synthesize technical, economic, political, and environmental data to stress-test investment theses across a range of futures would represent a genuine step change in how the industry allocates capital.
The readiness gap: organizational, not technical
The central challenge is not whether AI will become more capable, it will. The challenge is whether mining organizations can absorb that capability and translate it into value. This is fundamentally an organizational and cultural problem, not a technical one.
McKinsey's research on operational excellence in mining (February 2025) found that the few companies achieving exceptional, sustained results share a common feature: they invested in building people's capabilities so that operators understand not just what to do and how to do it, but how it could be done better. Leaders in these organizations serve their teams, providing context so that decisions can be made close to the actual work. This human-centered approach to operational improvement turns out to be the same foundation required for effective AI adoption.
The pattern is consistent: technology deployed into unreformed organizational structures produces disappointing results. BCG's research on capability building reinforces this point, noting that companies in process-heavy industries like mining may find it reasonably easy to develop lean capabilities for existing processes but struggle to implement new digital capabilities that require changes to skills, technology, and daily behaviors. Without a systematic approach, behavioral changes are superficial and temporary.
For mining companies considering how to prepare for increasingly capable AI, several priorities emerge:
Invest in data architecture before algorithms. The most sophisticated AI system is useless if it cannot access clean, integrated data across operations. Many mining companies still operate with fragmented data environments where different systems do not communicate. Fixing this is unsexy but essential.
Build cross-functional integration. The value of advanced AI comes from its ability to optimize across boundaries. Organizations that remain deeply siloed, where the mine, the plant, and the commercial function operate as separate fiefdoms, will not capture that value regardless of how capable the AI becomes.
Develop governance frameworks for autonomous systems. As AI moves from recommendation to action, the governance question becomes critical. Who is accountable when an autonomous system makes a decision? How do you audit AI-driven choices? How do you maintain human oversight without creating bottlenecks that negate the efficiency gains? These are questions that need answers before deployment, not after.
Treat capability building as strategic investment, not cost. The single strongest factor McKinsey identified correlating with AI success is workflow redesign. High-performing organizations do not add AI to existing processes, they build new, AI-native processes. This requires sustained investment in people, training, and organizational change.
The strategic question
The future of AGI in mining is not a technology question. It is a strategy question about which organizations will build the foundations, in data, in people, in governance, and in organizational design, to absorb increasingly powerful tools and translate them into sustained competitive advantage.
The companies that will lead are unlikely to be those that deploy the most algorithms or spend the most on AI vendors. They will be the ones that treat AI as a catalyst for rethinking how they operate, make decisions, and allocate capital. The technology is advancing. The question is whether the industry will advance with it.
Sources:
McKinsey, "Unearthing a New Era of Innovation in Mining," December 2025
McKinsey, "The State of AI in 2025: Agents, Innovation, and Transformation," November 2025
McKinsey, "Mining for Operational Excellence," February 2025
ResearchAndMarkets, "Strategic Intelligence: Artificial Intelligence in Mining," November 2025
MarketsandMarkets, "AI in Mining Market — Global Forecast to 2032," 2025
ScienceDirect, "Human-Machine Collaboration in Mining: A Critical Review," August 2025
BCG, "Building Capabilities for Transformation That Lasts," 2016
Global Mining Review, "Five Ways AI Will Transform Mining in 2026," November 202
There is no shortage of breathless predictions about artificial intelligence in mining. Market reports project AI spending in the sector growing from $2.7 billion in 2024 to $13.1 billion by 2029. Headlines promise autonomous operations, intelligent exploration, and AI-driven decision-making at every stage of the value chain. But beneath the forecasts and the vendor pitches, a more interesting, and more difficult, question is taking shape: what happens when AI moves beyond narrow task automation toward something approaching general intelligence?
This is a question mining executives should be engaging with now, not because artificial general intelligence (AGI) is imminent, but because the trajectory of AI development is already reshaping what is strategically possible, and the industry's readiness gap is widening.
The current state: useful but fragmented
The honest assessment of AI in mining today is that it works, in pockets, but has not yet transformed how most companies operate. McKinsey's December 2025 analysis of innovation in mining describes an industry where AI is "moving from pilot projects to everyday practice," with leading miners applying deep learning to optimize fixed plants and using agentic AI to reinvent complex processes such as source-to-pay and issue identification. China's mining industry now operates nearly 4,000 autonomous vehicles, more than half of them electric, running continuously with AI-optimized routing.
Yet McKinsey's broader State of AI survey (November 2025) reveals a sobering pattern across all industries: 88 percent of organizations now use AI in at least one function, but only about one-third have scaled it beyond pilots. Only roughly 6 percent of companies attribute more than 5 percent of EBIT to AI. The problem is not the technology itself. It is that most organizations remain stuck in what many observers now call "pilot purgatory," launching isolated proofs of concept that never converge into redesigned workflows or enterprise-level capability.
Mining faces this challenge acutely. The sector's dispersed operations, legacy systems, fragmented data environments, and deeply siloed organizational structures make scaling AI harder than in more digitally native industries. The operational complexity is real: ore variability, equipment heterogeneity, remote sites, and regulatory constraints create an environment where generalizable AI solutions are difficult to deploy.
From narrow AI to agentic systems: the intermediate horizon
The more immediate and consequential shift is not the arrival of AGI in the theoretical sense, a machine that can reason across all domains the way a human can, but rather the emergence of agentic AI systems. These are AI architectures capable of planning and executing multi-step workflows autonomously, not just answering questions or making predictions but taking sequences of actions across interconnected processes.
McKinsey's 2025 research found that 62 percent of organizations are experimenting with AI agents, and 23 percent report scaling them in at least one function. However, in no single function has the "scaled" share exceeded roughly 10 percent. To make agents operationally useful, companies need standardized process steps, modern interfaces around legacy systems, and governance frameworks that make autonomous action observable and interruptible.
For mining, agentic AI has distinctive implications. Consider the mine-to-market value chain: extraction, processing, logistics, and sales are typically managed in organizational silos with limited real-time data exchange. An agentic system could, in principle, coordinate across these boundaries—adjusting blast schedules based on downstream processing constraints, rerouting logistics in response to grade variability, and rebalancing production plans against shifting commodity prices—all without waiting for human intermediation at each handoff point.
The concept of "Just in Time Mining," where integrated AI and potentially quantum computing enable real-time optimization of extraction based on demand signals and operational conditions, represents the far end of this trajectory. It would mean treating the mine not as a sequence of independent operations but as a single adaptive system.
What AGI would actually change
If we look beyond today's agentic systems to a future where AI capabilities approach genuine generality, the ability to reason across unfamiliar domains, handle ambiguity, and learn from limited examples, the implications for mining are profound but also more nuanced than most projections suggest.
Exploration and resource characterization. The most transformative near-term application may be in exploration, where AI is already combining satellite imagery, geophysical surveys, geochemical data, and historical drilling records to generate higher-confidence targets. A system with more general reasoning capabilities could integrate geological models with economic constraints, geopolitical risk, ESG considerations, and infrastructure availability to produce genuinely strategic assessments of where to explore, not just where minerals are likely to be, but where developing them makes sense.
Operational decision-making under uncertainty. Mining is fundamentally a business of managing uncertainty—in ore grades, commodity prices, equipment reliability, and regulatory environments. Current AI excels at pattern recognition within defined parameters. A more generally capable AI could help leaders reason about scenarios that fall outside historical distributions: novel geological conditions, unprecedented market structures, or regulatory shifts that have no close precedent.
Capital allocation and portfolio strategy. Perhaps the most underexplored area. Mining companies make capital decisions with 20- to 30-year horizons based on necessarily imperfect projections. A system that could synthesize technical, economic, political, and environmental data to stress-test investment theses across a range of futures would represent a genuine step change in how the industry allocates capital.
The readiness gap: organizational, not technical
The central challenge is not whether AI will become more capable, it will. The challenge is whether mining organizations can absorb that capability and translate it into value. This is fundamentally an organizational and cultural problem, not a technical one.
McKinsey's research on operational excellence in mining (February 2025) found that the few companies achieving exceptional, sustained results share a common feature: they invested in building people's capabilities so that operators understand not just what to do and how to do it, but how it could be done better. Leaders in these organizations serve their teams, providing context so that decisions can be made close to the actual work. This human-centered approach to operational improvement turns out to be the same foundation required for effective AI adoption.
The pattern is consistent: technology deployed into unreformed organizational structures produces disappointing results. BCG's research on capability building reinforces this point, noting that companies in process-heavy industries like mining may find it reasonably easy to develop lean capabilities for existing processes but struggle to implement new digital capabilities that require changes to skills, technology, and daily behaviors. Without a systematic approach, behavioral changes are superficial and temporary.
For mining companies considering how to prepare for increasingly capable AI, several priorities emerge:
Invest in data architecture before algorithms. The most sophisticated AI system is useless if it cannot access clean, integrated data across operations. Many mining companies still operate with fragmented data environments where different systems do not communicate. Fixing this is unsexy but essential.
Build cross-functional integration. The value of advanced AI comes from its ability to optimize across boundaries. Organizations that remain deeply siloed, where the mine, the plant, and the commercial function operate as separate fiefdoms, will not capture that value regardless of how capable the AI becomes.
Develop governance frameworks for autonomous systems. As AI moves from recommendation to action, the governance question becomes critical. Who is accountable when an autonomous system makes a decision? How do you audit AI-driven choices? How do you maintain human oversight without creating bottlenecks that negate the efficiency gains? These are questions that need answers before deployment, not after.
Treat capability building as strategic investment, not cost. The single strongest factor McKinsey identified correlating with AI success is workflow redesign. High-performing organizations do not add AI to existing processes, they build new, AI-native processes. This requires sustained investment in people, training, and organizational change.
The strategic question
The future of AGI in mining is not a technology question. It is a strategy question about which organizations will build the foundations, in data, in people, in governance, and in organizational design, to absorb increasingly powerful tools and translate them into sustained competitive advantage.
The companies that will lead are unlikely to be those that deploy the most algorithms or spend the most on AI vendors. They will be the ones that treat AI as a catalyst for rethinking how they operate, make decisions, and allocate capital. The technology is advancing. The question is whether the industry will advance with it.
Sources:
McKinsey, "Unearthing a New Era of Innovation in Mining," December 2025
McKinsey, "The State of AI in 2025: Agents, Innovation, and Transformation," November 2025
McKinsey, "Mining for Operational Excellence," February 2025
ResearchAndMarkets, "Strategic Intelligence: Artificial Intelligence in Mining," November 2025
MarketsandMarkets, "AI in Mining Market — Global Forecast to 2032," 2025
ScienceDirect, "Human-Machine Collaboration in Mining: A Critical Review," August 2025
BCG, "Building Capabilities for Transformation That Lasts," 2016
Global Mining Review, "Five Ways AI Will Transform Mining in 2026," November 202