It’s groundbreaking technology. Pic: Getty Images
The brains behind Microsoft’s bid to build a quantum computer within three years is also helping BHP dig more copper out of the ground to wire everything from electric cars to data centres.
Copper demand has exploded – used in hotplates to EVs, solar panels and the most advanced AI ‘factories’. It is also found virtually everywhere in the Earth’s crust. But the easy ore is mostly gone, prompting BHP to take a radically different approach.
Microsoft has provided it with its most advanced artificial intelligence tools to pull copper from low grade rock, which is notoriously difficult. This type of rock normally contains about 1 per cent copper, requiring massive amounts of energy and chemical processing.
BHP turned to Microsoft Discovery, the division spearheading the tech titan’s quantum computing push. It now halved its timeline to build a working quantum computer, saying a machine will be ready by 2029.
This is possible thanks to its new Majorana 2 chip, unveiled at Microsoft’s Build developer conference in Seattle this week, and quantum chemistry, which is also helping BHP become more efficient and provide the core material needed to wire the AI boom.
BHP’s copper sits deep, spread thin and bonded tight to the minerals around it. Freeing it means leaching – a hydrometallurgical process that dissolves the metal so solvent extraction can draw it off – and a better reagent rewrites the economics of every deposit the industry has written off.
Working with Microsoft and computational chemists at Prescience Insilico, BHP ran tens of thousands of quantum chemistry calculations to narrow half a million molecules called reagents to a shortlist now facing the bench in Australian laboratories.
The job ran on Microsoft Discovery, an Azure platform made generally available on Wednesday that deploys teams of ‘‘agentic’ AI to mimic the scientific method.
“Copper is everywhere, from electric vehicles to solar panels, wind turbines and grid infrastructure,” said BHP vice-president of innovation Jessica Farrell.
“Much of the critical infrastructure and enablers of the energy transition require copper.”
The reagent hunt is a problem chemists have chipped at for a generation.
Ms Farrell said using AI was “increasing our speed, freeing up our technical experts to do more thoughtful analysis, and it’s increasing our chances of success”.
But the BHP project used AI, not a quantum chip. Still, modelling how molecules behave, atom by atom, is the exact problem a fault-tolerant quantum machine is built to crack.
Microsoft says the Majorana 2 processor’s qubits are 1,000 times more reliable than the chip it built a year ago.
It brings a scalable quantum computer, which has been a decade away for decades, now set to be available in just a few years.
Quantum computing promises to solve problems classical computers cannot touch. Microsoft Discovery vice president Jason Zander told The Australian in April that it would create AI models that are “100 times, 1000 times smarter than what you’ve already got”, allowing companies to tackle problems like the removal of ‘forever chemicals’ such as PFAs and accelerate drug development.
“Quantum computing is going to give us hyper-accurate AI for particular domains like chemistry and biology and other sort of areas like that. What that will do is give us the opportunity to discover new products and build them much faster than we’d ever have before,” Mr Zander said.
Microsoft has spent $US1bn on quantum computing over the past two decades and achieved a breakthrough last February with its Majorana 1 chip, which leverages a new state of matter.
The Majorana 2 advancement came from a swap of raw material. It ditched aluminium for lead, paired with a semiconductor mix of indium arsenide and indium arsenide antimonide.
That change widened what physicists call the ‘topological gap’ – the shield protecting a qubit from the environmental noise that scrambles its calculations – to more than double the previous processor’s.
The pay-off is in survival time. In the aluminium chip, qubits held their quantum state for between one and 12 milliseconds. In Majorana 2, the mean lifetime runs to 20 seconds. Some last more than a minute.
“This rapid progress, enabled by AI, has cut our timeline in half,“ said Chetan Nayak, the technical fellow who leads Microsoft’s quantum hardware team.
This article first appeared in The Australian as Microsoft’s quantum division uses AI to solve BHP’s major copper mining challenge