# Rio Tinto’s AI Push: What’s Real, What’s Press Release
Here’s what this really means: Rio Tinto is one of the few miners actually extracting value from data, not just talking about it. I’ve watched plenty of industrials slap “AI” on their investor decks since 2023 and deliver nothing but conference slides. Rio is different — not because it’s cutting edge, but because it’s been grinding at this since the autonomous haul truck days in the Pilbara, well over a decade ago. Patience and boring iteration beat hype cycles every time.
The company runs global Centres of Excellence covering analytics, automation, asset management, energy and climate, orebody knowledge, and both underground and surface mining. Translation: Rio concentrates its best technical people and deploys them across operations worldwide, primarily to manage major hazard risk and keep assets running properly. That’s not glamorous. It’s also exactly where the money is. Mining margins are won and lost on equipment uptime, blast quality, and not killing anyone.
Its Mine Automation System — MAS — pulls data from 98% of its sites into a common format, processed through algorithms and served up either through RTVis, its visualization layer, or standard dashboards. Ninety-eight percent coverage is the number worth noting here. Getting mid-sized operations in remote jurisdictions onto a unified data platform is a decade of unglamorous slog. Most competitors are still arguing about which vendor’s system to buy. That data foundation, not the algorithms, is the actual moat. Everyone wants the magic model; nobody wants to do the plumbing.
On top of that foundation, Rio applies AI to orebody modeling, equipment dispatch, and blast prediction and control. One concrete example: AI-optimized truck speeds and reduced queuing in its autonomous fleet, which produced material productivity gains. This isn’t the first time we’ve seen this circus — small operational percentage improvements pitched as revolutionary — but in mining, small is real. A 3% improvement on a fleet of 400-ton haul trucks moving iron ore at scale is serious money, year after year. Contrast that with software valuations where “AI” earns a 30x multiple for nothing. In heavy industry, you earn it ton by ton.
The newer headline is a memorandum of understanding signed in March 2026 between Rio Tinto and AI Singapore, a two-year collaboration facilitated by Enterprise Singapore to develop and scale AI solutions across Rio’s supply chains. The first initiative: AI-assisted freight invoice processing. Rio processes more than 10,000 freight invoices annually, and validating them currently means manual cross-checking across multiple data sources.
Now, let me be blunt. Freight invoice processing is not moonshot material. It’s back-office automation — valuable, sure, but this is where AI actually pays its way today: tedious, rules-heavy, high-volume work where errors cost real dollars and humans make mistakes out of boredom. Rio’s framing about “streamlining validation” and “reducing turnaround times” is credible precisely because the ambition is modest. My skepticism radar goes off when a miner promises AI will reinvent discovery or replace geologists. It doesn’t when the target is invoice reconciliation. That’s the right-sized bet.
The CFO of Rio Tinto Commercial, Will Millsteed, framed it as strengthening the ability to apply AI to complex supply-chain problems, noting that mining is seen as traditional but faces challenges demanding modern, AI-enabled solutions. Fair enough. Though I’d note the phrase “build new capabilities” is the kind of language that appears in every partnership announcement since the dawn of partnerships. Watch what ships, not what’s signed. MOUs are cheap. Deployed systems with measured throughput improvements are not.
AI Singapore, launched in 2017, aggregates the country’s research institutions, startups, and talent to do applied AI work. Singapore, for all its lack of mines, has spent years positioning itself as the digital and commodities trading hub of Asia — most of the region’s physical commodity flows are financed, traded, or coordinated there. Enterprise Singapore’s Lee Pak Sing made the play explicit: develop Singapore as the place where commodity companies leverage AI. That’s nation-building strategy wearing a partnership press release. Smart, actually. Singapore hosted the iron ore derivatives market’s growth for a reason — infrastructure, talent, and regulatory stability attract the real economy’s paper. Adding AI services to that stack is a logical extension.
Laurence Liew, AI Singapore’s Director of AI Innovation, talked up giving talent access to real-world industrial problems, calling mining a sector where digital transformation is “emerging.” Here’s what that really means: from the outside, mining looks digitally primitive. From the inside, the majors — Rio, BHP, Vale — have been running autonomous fleets, remote operations centers, and machine learning on processing plants for years. What’s emerging is the second wave: AI applied to the commercial and administrative layers, not just the pit. That’s where the Singapore collaboration sits.
The other thing buried in this announcement is a marketing play: Marina Bay Sands will host the inaugural “AI Mine” conference in September 2027, an annual event dedicated to real-world AI use in mining. Take note — when a dedicated conference circuit forms around a technology theme in a cyclical industry, you’re usually mid-cycle in the hype wave. I watched the same pattern with blockchain-in-mining conferences around 2018. Most of those died quietly. The difference here is that AI in mining has measurable, bankable use cases already deployed — autonomy, dispatch, predictive maintenance, grade control. The conference is riding something real, even if half the podium talks will be vapor and vendor pitch decks. Go for the case studies with audited numbers. Skip the vision keynotes.
For investors, here’s the practical takeaway. Rio’s AI program is not a reason to buy the stock — it’s a reason to trust the operational execution behind the guidance. Marginal productivity gains compound in a business with Rio’s volumes, and cost discipline is what protects margins when iron ore inevitably rolls over, as it always does. The risk side: data platform concentration, cybersecurity on connected autonomous systems, and the persistent temptation for management to dress up capex as “digital transformation” to justify spending. Watch whether these initiatives show up as lower unit costs in the accounts, not just in press releases.
This isn’t a story about AI revolutionizing mining. It’s a story about a disciplined operator doing the unglamorous work — data plumbing, incremental optimization, targeted partnerships — and letting the compounding do the talking. In my experience, that’s the only kind of technology story that survives contact with a downturn.
Source: https://im-mining.com/2026/11/13/rio-tintos-ai-journey-and-the-pursuit-of-operational-excellence/
