METS 2.0

By Aidan Montague · First published 29 July 2026

METS 2.0 is the argument that Australia's edge in industrial AI is the same edge it had in mining equipment, technology and services (METS): applying technology to hard operations better than anyone else, then exporting that applied capability to the world. Australia does not need to build the frontier AI models, just as it did not build the haul trucks. The term comes from Aidan Montague, an Australian engineer and author of Design Before Scale, working with mining, engineering and energy firms on AI commercialisation.

The precedent

Australia has answered the question "where do we lead?" before. METS became a genuine export industry, and not because of the ore. Australians operated some of the hardest mines on earth, and the operating knowledge itself turned out to be the product. The machines often came from overseas. The know-how of running them, planning around them and improving them was built here, and the world bought it.

The 2.0

Industrial AI is the same play again. The frontier models will come from overseas, and that is fine. The edge is applying AI to mining and energy operations that Australian firms understand better than anyone, then exporting that applied capability the way METS was exported.

What it means for a firm

The exportable asset is the operating knowledge in the business: how its people run a plant, plan a mine, price a job or judge a risk. That knowledge only becomes an asset the firm can sell once it is captured in a system the firm owns. Poured into shared AI tools, it becomes everyone's baseline instead, which is Judgement Leakage. The question for any firm is the one the original post ended on: what is the operating knowledge in your business that the rest of the world would pay for?

First published

First published on LinkedIn on 29 July 2026. The post is reproduced below as it was posted, with the hashtags removed. View the original post on LinkedIn.

Graphic from the 29 July 2026 post: Australia's AI edge is METS 2.0.

As posted on LinkedIn, 29 July 2026

Australia doesn't need to build frontier AI models. We didn't build the mining trucks either. Everyone from the Prime Minister down is asking where Australia leads in the AI era. The answers keep coming back the same: dig more of what the data centres need, or spend billions building sovereign models of our own. Both miss the precedent. We've answered this question before. METS - mining equipment, technology and services - became a genuine global export industry, and not because of the ore. Australians operated some of the hardest mines on earth, and the operating knowledge itself turned out to be the product. Mine-planning software from Micromine, Maptek and Deswik runs on mine sites all over the world. Orica turned blasting into a digital science. And the haul trucks: the machines came from Japan and America, but running them driverless at fleet scale matured in the Pilbara. Now big mines everywhere are following. Industrial AI is the same play again. Call it METS 2.0. The frontier models will come from overseas, and that's fine. The edge is applying AI to mining and energy operations we understand better than anyone on earth, then exporting that applied capability the way we exported METS. It's already starting. MMG's Dugald River operation recently published a measured AI result: flotation control built with Hatch on a well-run plant, payback inside three months. Small example, right shape. So where does Australia lead? Same place as last time. In the operations. What's the operating knowledge in your business the rest of the world would pay for?