Schneider Electric wants mines to run as software-defined, AI-powered platforms — electrification, automation, and digital intelligence stacked into one tidy story. I’ve heard versions of this pitch since the late ’90s, when every industrial vendor promised the “digital plant.” Most of those promises died in pilot purgatory. That said, this one has a few elements worth taking seriously, and one angle that’s genuinely interesting for reasons the press release buries.
**The three-layer pitch**
The company frames its offering in three layers: connected devices that sense and control, edge computing and automation that keep systems running, and analytics on top that turn data into decisions. It’s a sensible architecture — sensors feeding a data backbone feeding models — and Schneider’s AVEVA acquisition gives it a legitimate seat at that table. Digital twins across the mine lifecycle round out the picture.
Here’s what this really means: the vendor that owns your data pipeline owns your upgrade path. That’s not necessarily bad, but understand you’re buying into an ecosystem, not a product. Switching costs later will be brutal. Negotiate accordingly.
**The Torngat deal — this is the part that matters**
Schneider signed a non-binding MoU with Torngat Metals around the Strange Lake rare earth project in Nunavik, with associated infrastructure in Labrador and a planned separation facility at Sept-Îles targeting 15,000 t/y of rare earths. On paper it’s a “360° partnership”: Schneider lends electrification and digital expertise to the project, and in exchange gets long-term access to a secure, compliant rare earth and magnet supply chain.
Let’s be clear about what’s happening here. A major industrial equipment maker is effectively taking equity-like exposure to a critical minerals project — through a non-binding memorandum, mind you — because it’s worried about where the magnets inside its own products will come from in ten years. When the buyers start vertically integrating upstream because China has too much leverage over the supply chain, that tells you something about the strategic reality that no policy paper will.
This isn’t the first time we’ve seen this circus — remember the lithium rushes of 2010-2012, or cobalt panic-buying — and plenty of rare earth hopefuls have flamed out on permitting, metallurgy, and cost. Rare earth separation is nasty chemistry, not software. Strange Lake is remote, requires a seasonal road and port, and depends on a separation plant that doesn’t exist yet. A non-binding MoU is not a bankable offtake, no matter what the press release implies. Schneider’s own executive admitted the intent resembles an early-stage offtake — which, translated from corporate speak, means: we want optionality without commitment.
Still, if you’re watching critical minerals, this is the kind of demand-side anchor that separates serious projects from brochure projects. It doesn’t make Strange Lake de-risked. It makes it watchable.
**The proof points**
Schneider cites three deployments that give the pitch some spine.
At SQM, AI-driven predictive and optimisation models delivered a 10-15% cut in process variability, up to 1% more nitrate yield, and 1-3% better raw material efficiency. Now, 1% sounds like rounding error until you remember that in commodity processing, 1% margin on billions in throughput is real money. Process variability reduction is the quiet, unglamorous win that actually pays — far more credible than “AI transforms everything” marketing. Models hit roughly 90% accuracy with faster retraining cycles, and the operation is expanding the approach to more plants with maintenance and lab data layered in.
At Roy Hill in Western Australia — 55 Mt/y of iron ore — Schneider built remote operations and demand chain planning spanning mine, plant, rail, and port, including the full power infrastructure for Samsung C&T: 14 switch rooms, transformers, drives, backup power, and power management. This one I’ll credit without much hedging; Australian iron ore operations are among the few places where remote pit-to-port integration has genuinely been proven at scale.
With Vale, Schneider and AVEVA centralized monitoring across Mariana and Itabira, and are applying hybrid process control plus AI to milling circuits — with furnace NOx emissions reductions shown in other industrial applications.
**The data architecture story**
Schneider’s actual differentiator, if it has one, is the CONNECT platform sitting on top of the AVEVA PI System — a unified data backbone replacing the disconnected sensor, control, and lab systems that plague every mine. At SQM, operational data flows into Databricks via Delta Sharing, enabling near-real-time machine learning workflows.
The “closed-loop” piece deserves attention: model outputs get written back into operational systems, displayed in tools like PI Vision, and in some cases drive automatic setpoint adjustments. That’s the difference between AI as a dashboard novelty and AI as an actual control layer. It’s also where the risk lives — any system that closes the loop from model to actuator without a human in the path needs fail-safes designed by people who’ve watched control systems misbehave at 3 a.m. Most don’t have them. Ask hard questions about override authority before anyone hands setpoints to an algorithm.
The hybrid modeling approach — physics-based simulation combined with data-driven machine learning — is the right instinct. Pure black-box models fall apart the moment an ore body or feedstock shifts, and mining feedstocks always shift.
**The takeaway**
The mining industry is late to industrial AI, and the gap between pilots and enterprise deployment is where most of these programs go to die. SQM’s expansion beyond the initial nitrate optimisation is the exception worth tracking, not the rule. If you’re an investor, the signal here isn’t the AI marketing — it’s that a global equipment giant is willing to trade technology for upstream mineral exposure. Watch whether this MoU hardens into binding offtake or equity, and whether Strange Lake clears financing and permitting. Until then, it’s a smart hedge by a company that read the geopolitical tea leaves — not yet a done deal.
Source: https://im-mining.com/2026/11/10/schneider-electric-explores-technology-convergence-in-mining/
