Why Mining Suppliers Are Betting on AI and Diversification Together

Mining equipment and service suppliers are often assumed to follow a simple rule in a downturn: cut back, wait, then ride the next upswing. The third Santiago Mining Summit, held on 24 September 2026, suggested a different approach. In Latin American mining technology, the push into AI and autonomy and the advice to diversify beyond mining turned out to be the same strategy.

Chile’s project pipeline gives the timing its urgency. Cochilco, the Chilean Copper Commission, valued the country’s 2025-2034 mining investment portfolio at US$104.549 billion. Several of the largest projects in it are still waiting on environmental approval.

That gap between planned investment and approved projects shapes where suppliers are placing their bets. If you follow mining equipment makers, automation vendors or the miners that buy from them, it affects how you judge growth claims.

Here is which technologies are already deployed at scale, which remain vendor ambition, and how suppliers are positioning to absorb the next delay in the cycle.

What is Honeywell’s “Digital Cognition,” and why does it matter for mining?

Picture a concentrator control room today. An operator watches grinding and flotation data on screen while software adjusts setpoints in real time to keep throughput steady. That software layer is advanced process control (APC), a system that uses process models to automatically tune plant variables so output stays stable and efficient.

Honeywell used the summit, organised with AmCham Chile and with Microsoft participating, to pitch what sits on top of that screen. Leaders from Codelco, BHP, Rio Tinto and Albemarle were among the participants. The company calls the concept “Digital Cognition.”

José Fernandes, Honeywell’s Latin America President, framed blending AI with advanced automation as the way to speed up production efficiency in the region. Jason Urso, Technology Director for Process Automation and Industrial Automation, described Digital Cognition as sensors, process models and AI combined to interpret operating conditions and advise operators on what to do next.

Both sit inside a wider idea Honeywell calls Mining 5.0. It is a human-centred model that combines AI, advanced automation and industrial expertise, with the aim of giving workers more analytical capability rather than removing them. The layers build in sequence:

1. Sensors and process models that measure and describe the plant
2. APC and operator training simulators
3. Remote operation centres and asset performance management (software that tracks equipment health to predict failures)
4. AI-enabled recommendations to operators
5. Autonomous operation

Company claims, not independently verified Honeywell states that more than 45% of global copper production uses its APC solutions, and that it has over 600 computer-vision solutions deployed in mining worldwide. Neither figure is broken out by country.

The pitch has also moved year by year. The 2024 summit centred on digital technologies, data and cloud. The 2025 edition focused on digitalisation and AI. In 2026, the theme was autonomy.

From automated to autonomous: what actually changes

Steps one to four are decision support: the system recommends and a person decides. Step five is where the system acts. Honeywell frames that step as “autonomy connected with people,” and José Simon, its VP and GM of Industrial Automation for Latin America, has argued that AI “agents” should work alongside staff rather than replace them.

The idea of AI agents working alongside staff reflects a wider shift toward agentic AI, where software does not just flag anomalies but plans and executes tasks within defined limits set by operators.

Summit commentary projected that Chile’s autonomous fleet could triple this decade. Treat that as a forecast from interested parties, not a measured trend.

The read for you is that AI in mining is being sold as an add-on to infrastructure that must already exist. When you assess any AI claim, the control and data systems underneath it are the real test of credibility.

Why is the technology push accelerating while projects stall?

The headline numbers suggest a sector flush with new capacity. Cochilco’s portfolio, published on 11 December 2025, rose 25.7%, or US$21.369 billion, on the prior edition.

Look closer and much of that growth depends on approvals not yet granted. The new Rosario concentrator, a fourth line at Collahuasi, and the new Los Colorados concentrator at Escondida are both in pre-feasibility. Neither holds environmental qualification, yet both owners describe them as central to future growth and operational continuity.

That leaves miners with a timing problem.

Summit speakers tied AI and autonomy directly to closing the gap between current production and rising demand for critical minerals such as copper and lithium. When new concentrators are years and permits away, extracting more from existing plants becomes the faster lever. Technology spending, in other words, is partly a response to constrained supply growth.

The evidence base has limits worth stating plainly. No consolidated Latin America-wide capex figure was found, and country-level investment or deployment data for Peru, Brazil and Mexico was not available. Chile carries almost the entire quantitative argument.

The caveats vendors rarely lead with

The productivity case comes with friction that slows rollout. Simon has said near-autonomous operations need cybersecurity, AI and the Internet of Things (IoT, networks of connected sensors and devices) working together, which also signals how much more connected, and exposed, these sites become.

Honeywell’s deployment figures do push back against the idea that this is all hype. They are, however, vendor-reported. For you, the practical signal is that adoption pace will track permitting delays and proof of return as much as technical capability.

How are suppliers hedging the next cycle?

If miners are buying technology to work around stalled projects, suppliers face the mirror image: their own order books depend on those same projects. Diego Torroella, Managing Director of TAKRAF Mexico, offered a blunt answer.

Diego Torroella, Managing Director, TAKRAF Mexico Firms should innovate actively rather than idly wait for the next mining cycle, applying core engineering skills during project delays.

His targets are specific: material handling, water recovery and energy efficiency. Those problems are not unique to mines, and Torroella pointed to ports and petrochemical facilities as neighbouring sectors where the same engineering applies.

Honeywell’s version of the hedge looks different on the surface. Its core platforms are positioned across mining, energy and process industries:

• APC: process optimisation in mining, energy and petrochemical plants
• Computer vision: safety and equipment monitoring across industrial sites
• Remote operation centres: centralised control for mines, ports and logistics hubs
• Asset performance management: equipment health tracking in any capital-heavy process industry

Put the two side by side and the logic converges. TAKRAF redeploys engineering skill into adjacent sectors; Honeywell builds platforms that already serve them. The 45% copper APC share and 600-plus vision deployments show how widely one platform can be stretched.

One caution applies. This cross-sector pattern is inferred from how suppliers position themselves, and no revenue-mix data was found to confirm how much non-mining income either company actually earns.

The lens for you is transferability. When you evaluate a mining supplier, ask how easily its core technology moves into other process industries, because that determines how well it can absorb a delay in mining spend.

What the summit signals, and what to watch before the next cycle

The summit’s message was that technology adoption and supplier diversification form one strategy, and both rest on control infrastructure that is already deployed rather than on AI alone. Three variables will show whether it holds:

1. Permitting progress on Rosario and Los Colorados, which sets how long miners lean on efficiency over expansion
2. Evidence of autonomy returns from miners themselves, beyond vendor claims
3. Signs that suppliers’ non-mining revenue is growing in reported results

If the first stalls while the second and third strengthen, the suppliers with the most transferable platforms are best placed to absorb the wait.

Investors tracking autonomy returns beyond vendor claims can use our full explainer on autonomous mining adoption barriers, which examines how integration cost and workforce readiness shape deployment.

Past performance does not guarantee future results. Projections cited are subject to market conditions and various risk factors, and forward-looking statements are speculative and subject to change.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions.