Autonomous drilling has quietly moved from pilot phase to business-as-usual at several major Pilbara iron ore operations. While autonomous haulage grabs most of the headlines, it’s the drilling programs that are delivering some of the most consistent productivity gains—and the data from the past two years tells an interesting story.

The Current State of Play

Rio Tinto’s autonomous drill fleet in the Pilbara now covers the majority of its blast hole drilling across Brockman and Marra Mamba deposits. BHP’s South Flank and Jimblebar operations have followed a similar trajectory, with Caterpillar and Epiroc rigs running autonomously across multiple pits.

What’s changed since early trials isn’t the technology itself—it’s the operational maturity around it. The rigs can drill. They’ve been able to drill for years. The real progress has been in pattern optimisation, multi-rig coordination, and integration with downstream blasting and loading sequences.

What the Performance Data Shows

After reviewing operational reports and speaking with drill-and-blast teams across three major operations, a few patterns stand out:

Utilisation rates are genuinely higher. Autonomous rigs consistently clock 18-22 hours of productive drilling per day, compared to 14-16 hours for operator-controlled rigs. The difference isn’t that autonomous rigs are faster—they’re not, really. It’s that they don’t take meal breaks, they don’t change shifts, and they don’t stop for vehicle passes unless the collision avoidance system requires it.

Hole accuracy has improved. Drill pattern compliance has increased to within 100mm on most operations, which translates directly to better fragmentation outcomes. More consistent blast results mean fewer oversize events and smoother loading cycles downstream.

Maintenance scheduling has become more predictable. When rigs operate within tighter parameters—consistent pulldown pressure, rotary speed, and bit loading—component wear is more uniform. Teams are reporting 10-15% improvements in bit life and reduced unplanned maintenance events.

The Challenges Nobody’s Advertising

It’s not all smooth sailing. Several issues keep showing up in operational reviews:

Ground condition variability. Autonomous systems still struggle with highly variable geology. When you move from competent BIF into weathered zones or clay-rich material, the drill parameters need to change quickly. Current systems can adapt, but there’s a noticeable lag compared to an experienced driller who can feel the change through the rig.

Dust management. Autonomous rigs run longer hours, which means more dust generation over extended periods. Several sites have had to upgrade their dust suppression systems—both on-rig and around the pit—to manage the increased exposure for nearby workers and equipment.

Integration with grade control. This is where the real opportunity lies, and it’s still underdeveloped. Drill cuttings from autonomous rigs produce consistent samples, which should feed directly into real-time ore grade optimisation models. Most operations are still handling grade control sampling manually, creating a bottleneck in what could otherwise be a continuous data stream.

Where AI Fits Into the Picture

The next wave of improvement won’t come from better rigs. It’ll come from smarter decision-making about where and how to drill. Machine learning models trained on geological data, past drill performance, and blast outcomes can optimise patterns in ways that static engineering rules can’t match.

Several operations are already working with firms that provide AI strategy support to build these decision layers. The challenge is integrating multiple data sources—geology models, drill performance logs, blast monitoring, and downstream processing data—into a system that can make real-time recommendations.

It’s a data problem more than a drilling problem. And it’s one that’s solvable with the right approach.

What Comes Next

The trajectory is clear: autonomous drilling will become the default for large-scale open pit operations in Australia within the next three to five years. The economics are too compelling, and the productivity gap is real.

The more interesting question is what happens in mid-tier operations. Companies running two or three drill rigs don’t have the same scale advantages as the majors. The capital cost of autonomous-ready rigs is coming down, but the support infrastructure—network coverage, remote monitoring, technical expertise—is still expensive.

For the mid-tier, the play might not be full autonomy. Semi-autonomous systems that handle the repetitive drilling while operators manage pattern changes and difficult ground could be the sweet spot. Epiroc’s BenchREMOTE and Caterpillar’s Command for Drilling both offer this middle path.

The Pilbara has proven the concept. The question now is how fast the rest of the industry catches up, and whether the technology adapts to the operational realities of smaller, more geologically complex deposits.

That’s where the real test begins.