The AI arms race has a critical minerals problem, with Michael Walshe and Jeff Dickerson

Let's Talk Energy — Episode 38 The AI arms race has a critical minerals problem, with Michael Walshe and Jeff Dickerson Wednesday, 20 May 2026 SPEAKERS NB Noah Brenner — Host, Let's Talk Energy MW Michael Walshe — CEO, Metallium JD Jeff Dickerson — Principal, Advisory, Rystad Energy Here is the transcript reformatted with proper spacing: [00:00] NB This is Let's Talk Energy, your go-to podcast for smart energy insights. I'm Noah Brenner. Critical minerals have vaulted to the front of any discussion about energy and security, as countries grapple with China's dominance of a sector that is the building block for everything from renewable power generation and batteries to computer chips and radar systems. The US and many of its allies are trying to rapidly develop a viable alternative supply chain through a combination of industrial policy, trade diplomacy, and innovation. NB As we're taping this on May 11th, China's President Xi Jinping and US President Donald Trump are preparing for a summit later this week where critical minerals will play a central role in trade talks, though few expect major changes in policy. So how did China become so dominant in mining and processing critical minerals, particularly rare earth elements? And how has the country used that leverage? How can the West begin to close the gap, and why are new approaches needed for non-Chinese companies to compete? And finally, what does a viable critical minerals industry in the West look like, and how long could it take to develop? NB To help us understand one of the most important overlaps between energy and geopolitics, we're joined from Perth by Michael Walshe, CEO of Metallium, which is advancing a new technology for both processing critical minerals from ores and extracting them from waste electronics. Michael, thanks for joining us. MW Thanks for having me. NB And from Houston, we're joined once again by Jeff Dickerson, a principal in our advisory business who leads Rystad Energy's work on critical minerals in the Americas. Jeff, welcome back. JD Thanks Noah, good to be here. NB Well, gentlemen, let's talk energy. Michael, can you give us a quick introduction to yourself and to Metallium — what is the company, and why is this the one you want to lead? MW I've been leading Metallium for about two years as CEO. My background is process engineering and, in particular, mining technology and process equipment. What originally excited me was that we have the potential to break through a lot of the existing flow sheets for how critical metals are currently refined and processed. Why that matters is that the stranglehold China has on the rest of the world — and in particular on the US — is its ability to refine critical metals. It's not the ore in the ground that's important. It's the refining and turning those metals into something of value that goes into end products. For the last 50 years or so, the West has offshored all of that ability, knowledge, and technology transfer to China. Since then, the West hasn't really progressed any technology of its own and has pretty much allowed China to do that dirty, hard lifting. MW In that time, China now produces more engineers per year — I think it's around ten to one in terms of engineers they throw at mining, technology, and critical metals. They have massively leapfrogged the West. My personal opinion is that the only way the West can really catch up is through two things: getting metals out of industrial waste and recycling above-ground resources, and making advancements in processing. The West offshored a lot of processing because it's generally dirty — you need to build huge chemical plant complexes with vast footprints. The only way we can beat that is by using smarter and more innovative techniques. That's what Metallium is trying to do across several critical metals. [04:00] NB Jeff, let's quickly ground this discussion. The term critical minerals gets thrown around a lot and can mean different things depending on the conversation — even rare earths refers to a broad part of the periodic table. What is it that we're talking about today, and can you build on Michael's comments around China's dominance in this sector? JD Critical minerals are, by definition in the US, any mineral material or substance that is essential to economic or national security and has a supply chain that's vulnerable to disruption. But if you look at other countries, they may define them differently — supplier countries like Brazil, for instance, tend to use the term strategic minerals. In the US specifically, which has really been driving this push since January 2025, there are 60 minerals plus a dozen more materials — around two thirds of the naturally occurring elements on the periodic table — that are counted as critical minerals. And so if everything's critical, then nothing's critical. It's all just a gaping hole in the West's industrial capacity, as Michael was describing. That hole is being filled by low-cost, marginal production coming out of China. JD For nearly every critical mineral on the USGS list — which among other things tracks US import dependence on other countries — China is either one of or the leading supplier. In many cases, it's the only supplier. That's particularly true for the basket of elements called rare earth elements, which includes the lanthanides group at the bottom of the periodic table, plus yttrium and scandium. Kudos to China — they've really earned this moniker, the China miracle. That's through decades of rigorous central planning and meticulous execution of strategy. They built out an industrial economy, scaled and squeezed out efficiencies through ruthless domestic competition, and cornered the sort of dirty, polluting markets that the West was happy to export in return for access to cheap goods. But now in the West, we're coming to terms with this dependence on China for all of the inputs to our advanced economy. JD Yet when you list out the physical requirements to re-industrialise, as we are trying to do, we've lost them over a generation — raw materials, processing infrastructure, experienced specialised labour, coherent industrial policies and strategy. These would normally take another generation to rebuild. But with increasing US-China competition on multiple fronts, China's broad critical mineral dominance provides major economic leverage over other countries. This has slowly taken place over decades — it didn't happen overnight. But we've been kind of like the frogs in the boiling pot. Really since 2023, but particularly over the last year or so, the situation has become quite dire and evolved from economic competition to economic warfare and now, looking at Iran and Venezuela, proxy resource wars. De-escalation, which I think is in everyone's interest, requires mitigating this economic and industrial dependence on China, which is why so many nations are now seeking to diversify supplies of critical minerals. But reversing a 30-plus year trajectory backed by literally trillions of US dollars in systematic consolidation of industrial capacity is a monumental feat, particularly if it's to be accomplished in three to five years rather than a full generation. NB Why is it that China is keeping prices lower rather than ratcheting them up? JD This comes back to the long-term perspective China has and their ability to plan far into the future, compared to our quarterly earnings-driven focus in the West. From China's perspective, it's more important to achieve your strategic objectives by not only cornering the market but also eliminating competition. The competition, from their perspective, is any competing mine, smelter, or anything else that would limit their ability to use this consolidation of capacity as geopolitical leverage. [08:00] JD As China built up its industrial capacity — initially to serve domestic needs, through incentivising domestic competition — it figured out the most efficient ways to produce these minerals and materials. That led to low-cost exports that beat out the competition. But then, trying to avoid the resource curse of being simply a miner or a processor, China built manufacturing hubs and industrialised up the value chain. In doing so, at the scale they've achieved, they've consolidated perhaps 50 to 100% of the productive capacity across all of these minerals and basically caused all competing sources to become uneconomic. MW When I attend conferences in other countries — there was one in Canada most recently — you still hear people from the mining industry saying we need more government support. But that only plays into the trap China is able to set. What it ends up doing is withholding minerals or goods either to achieve a certain objective or in response to a perceived slight — and then flooding the market later to push prices down and disincentivise investment in competing infrastructure around the world. We saw this with the Solyndra investment in US solar panels 20 years ago. It's been occurring at different scales over the past two decades as China has grown more effective at targeting the sectors it deems necessary. MW There's a classic case from Australia involving gallium. Back in the 1990s, a prominent alumina company here was planning to add a gallium circuit to their bauxite-to-alumina process plant. Most of the world's gallium comes as a by-product of turning bauxite into alumina. This was a public company — they'd put out statements about what they were trying to achieve. Literally the week they turned the circuit on and were ready to go into production, China dumped a large volume of gallium onto the market and the price crashed by over 50%. These guys had to shut the project down within weeks. That's just one example of China's ability to control pricing so that it obliterates the competition. [12:00] NB So when you're thinking about competing in this space — is it simply a matter of restarting smelters and bringing on additional capacity, or do we need to leapfrog to a next-generation technology? What's going to win here? MW We'd be foolish trying to simply replicate what China does and thinking it'll be economic. They've spent so much capital building up the vertical integration of all their industries, and the margins between each step are razor thin. Some of their projects don't even need to make a margin — the way I've heard it put is that they don't play by free market capitalism. They have state capitalism, or what some call neo-mercantilism, where the government is fine not making a return on equity as long as they achieve strategic dominance. So yes, we'd be foolish just trying to replicate their existing flow sheets. MW What we're trying to do is pick very targeted feedstocks, like printed circuit boards in a recycling context, rather than trying to turn copper ore into copper products and gold ore into gold products. Recycling is actually a much more viable source of metal than a typical ore body in the Western sense, because permitting times for new mining projects in the West are around 50 times longer than in China. A recycling project gets metal units to market much faster. So we're focused on recycling from electronic waste, which is a very rich source of many critical metals including rare earths, antimony, and niobium. And in mineral processing, we're targeting problematic flow sheets that the West has 100% outsourced to China — including lithium downstream refining and rare earth element processing. MW There's almost no complex lithium refining left in Australia as one example. A few years ago, Albemarle, a large American company, tried to establish a project there. Several months ago, they shut it down because the economics didn't work. That flow sheet used traditional 1950s IP, which was actually developed in Europe and the US and transferred to China in the late 1980s and 1990s. If we tried to just replicate that flow sheet, it wouldn't make sense. What we're trying to do is delete a number of the steps. For lithium, we're aiming to cut about 40 to 50% of the steps in the traditional flow sheet to reach the same end target — lithium carbonate. We can compete in recycling because that's a viable source of high-value metals. And we can compete in mineral processing if we use innovation. We can't compete with China using traditional methods at their scale. NB Michael, this is a good moment to explain flash joule heating — Metallium's flagship technology. Why is this the innovative leap forward that can start to counter China's dominance? MW Flash joule heating was developed at Rice University in Houston, Texas, originally sponsored by DARPA — the research arm of the Department of Defense. The DoD was interested in finding non-mining-based alternative sources for critical metals. That's what Dr Tour at Rice University originally began studying back in 2017. The original techniques were based around rapidly heating materials like red mud — the residue left over when you turn bauxite into alumina, which contains trace amounts of rare earths and other critical metals. After that, lots of other feedstocks were added, including electronic waste and spent lithium-ion batteries. [16:00] MW The technology combines ultrafast electrical-based heating with some proprietary chemistry. Most mining-based metal processing involves some form of heat liberation — you heat a metal above a certain temperature to liberate it from unwanted material in an ore. The industry workhorse for the last 70 or 80 years has been the rotary kiln, with electric arc furnaces introduced around the 1960s — but nothing really new and innovative has happened since then. What we do is heat much more efficiently and rapidly, and we also add chemistry on top. We have the ability to add reagents like chlorine and fluorine — very reactive species — which lower the boiling point of metals. For example, to boil iron oxide into a vapour would normally require over 3,500 degrees. Under a chlorine atmosphere, it comes down to around 350 degrees. There's a huge temperature efficiency effect when you do vaporisation in a different chemical environment. We combine proprietary ultrafast electrical heating with this chemistry — something no one has done before — and we're now commercialising that technique from the lab at Rice University to industrial scale. NB And you're able to apply this to waste streams, not just virgin ores — is that right? MW Correct. One of our initial target feedstocks for commercialising the tech is printed circuit boards — the green motherboards inside most electronics. They're an attractive feedstock because they're orders of magnitude richer in metal content than a typical ore body. A typical copper concentration in a printed circuit board can be over 10% by weight. Most of the world's copper today comes from ore bodies in Chile with concentrations below 0.2% copper. There's a massive difference in metal concentration in these recyclable materials. We're also working with industrial residues — as one example, a material sourced from Indium Corporation in Syracuse, New York. They refine metals like gallium and germanium and have a residue from their processing that's still quite rich in those metals. We've shown we can recover gallium and germanium from that industrial residue. So we're focused on both electronic waste and industrial residues as feedstock types. JD One of the important things to consider here is what the US and other Western nations currently do with that waste. Europe is ahead of other regions but still only collecting and processing around 20% of the electronic waste it produces. In the US, it's slightly lower — most of this is simply exported to other countries, including China, India, and various nations in Africa. This urban mining concept represents a huge opportunity. Rather than building brand new mines in the US, which as Michael mentioned can take decades to get started and have much lower mineral grades, we have all of that material here in much higher concentrations and we're currently exporting it. [20:00] JD Right now, gallium is fundamental to Nvidia's latest chips and also to a lot of defence technology. There's a particular system — informally called the gallium gun — that fires a beam to fry the electronics in incoming drones. But right now, China is 100% the source of refined gallium into the United States. Having the ability to take gallium from waste or industrial residues already located in the country is a major strategic advantage, because China has put in place export bans for gallium and other metals like germanium, antimony, and rare earths over the past five years, doing so on multiple occasions. They've only recently paused some of those bans — not ended them. It seems to be tied to ongoing tariff negotiations with the Trump administration, but it's a real leverage point China holds over the US. I think it's only the Trump administration that has really woken up to how dependent the country is on China for defence technology and AI infrastructure. A lot of executive orders specifically related to critical metals have been issued since Trump came to power. Before that, people's heads were largely in the sand. NB We're sitting in the midst of a fossil fuel energy crisis with the closure of the Strait of Hormuz — something that people said could happen but was always considered very unlikely. You're pointing to the potential for China to unilaterally cut off critical mineral supplies to the world. Do you think the world appreciates the potentially precarious situation it is in? Or is this analogous to the Strait of Hormuz — where everyone saw the danger but perhaps didn't price it in? MW Personally, I do think the world is pretty much oblivious to this critical metals story and how fundamental it is to all modern life. One aspect of human nature is that the demand for energy is infinite — anytime a new technology like AI comes along, people just want to consume more and more of it. AI is one of the most energy-intensive technologies humans have ever developed. And I think there's a real lack of understanding between all these AI projections and data centre projections and the actual raw materials required to make that a reality. Most people probably still think the cloud is just some ether up in space, but every digital bit has to be stored somewhere physical. In the 1970s, metals and mining accounted for maybe 20 to 30% of S&P market cap — including energy companies. Now it's less than 5%. There's a real lack of understanding of how important that is. [24:00] JD This is one of the things the US government has been trying to understand — what is the full impact on the economy of losing access to any specific resource? At the same time, China has been collecting exactly this data. While they don't have outright mineral bans right now, anyone wanting to receive certain minerals from China must apply for an export licence and provide detailed information about where those minerals will end up. The ostensible purpose is for China to avoid exporting so-called dual-use minerals — those with both commercial and military applications. But what they appear to also be doing is collecting very detailed information about Western supply chains and end uses. JD I'll give one example of how crucial some of these components are and how close we are to not having them. One of the rare earths commonly mined in this basket is yttrium. Most people have never heard of it. Yttrium is used in several applications, but one that may be familiar is in manufacturing gas-fired turbines. Turbine blades require a thermal barrier coating — and that coating is yttrium-stabilised zirconia. There is no functional commercial replacement at scale. Without it, turbine blades would melt at peak performance — you either have to scale down capabilities or use a different form of energy generation. And those turbines are the primary power source being built at data centre sites, because simple cycle turbines can scale up power quickly and provide high reliability. GE Vernova raised the flag at the end of last year that this could become an issue. They mentioned some runway into 2026, but stockpiles are dwindling. In February and March, Reuters reported that executives from two suppliers said they would no longer be able to supply yttrium-stabilised zirconia, due to an inability to access yttrium. China continues to produce it but has so slowed the bureaucratic process of allowing industrial buyers to access this material that it has the potential for a very dramatic effect on the US's ability to produce the power needed to meet AI growth, reindustrialisation, and general power demand over the next several decades. [28:00] NB Michael, are those the types of companies you're having conversations with — the ones that can't get these critical minerals in the West? And what do those conversations look like? Is it licensing the technology, adding a plant? MW Most of the world's yttrium currently comes from Chinese ionic clay deposits in southern China. There are a number of Brazilian ionic clay projects hopefully approaching production — one of them is Meteoric Resources, with whom we have a collaboration. Their ore body contains yttrium, and we've shown that we can use flash joule heating to upgrade and super-concentrate their midstream product into a viable source of yttrium and other valuable magnet rare earths like dysprosium, terbium, neodymium, and praseodymium. We're definitely talking to companies like Meteoric about being their processing solution so they don't have to sell their product to Chinese refiners. As a rare earth developer, most can't afford the capex to go all the way to separated rare earth products — there are only two Western companies doing that, Lynas and MP Materials, precisely because the capex is so high. If you put in a traditional solvent extraction plant, the economics are very challenging. MW We've also had early engagement from some of the defence contractors — including testing some of their industrial residues and potentially being a source of some of these metals. Those conversations are at an earlier stage, but we are actively speaking to defence contractors who are looking for both heavy rare earths like yttrium and metals like gallium and germanium. NB If the West wants to compete in this critical minerals game, does it need the entire value chain to be developed in a way that supports itself? MW It'll need some of the capital that big tech players can bring to the table — the hyperscalers, the Mag Seven type companies. It'll need a real concerted effort to direct some of their massive pools of capital toward this problem, potentially establishing industrial parks like China has done effectively, putting similar companies in close proximity to one another. It will need not just government support but real capital from these players. Some of them are waking up to the fact that critical metals are fundamental. Tesla is probably one of the early Mag Seven companies that really woke up to that — they've established their own lithium processing plant in Texas to try to go further up the value chain, because they were struggling to get lithium units from China. If more companies adopt that Tesla-type approach, something could definitely be done to counteract that imbalance. MW The catalyst might be when the hyperscalers realise they can't get their turbines for their data centres or the gallium for their chips. I think that's when the fire will really be lit for them to do something about it. The capital the US government is putting in — or at least indicating — is a starting point, with tariffs, price floors, offtake guarantees, and the VAULT programme, a $12 billion stockpiling initiative. But the government isn't going to be able to buy everything. It's really going to take a sense of urgency among large, well-capitalised players. And that's also one of the things that has stopped China from pursuing a scorched earth policy — they want to make sure that industry, with all its capital, isn't incentivised to really magnify the impact of the government's initiatives. [32:00] NB Jeff, what does success look like? How much does the West need to produce to change the dynamic in this market? Is it meeting military demands, or producing enough to supply other critical industries? JD The innovations need to be revolutionary, not incremental. Innovation at a smelter is constantly occurring to squeeze out efficiencies. Revolutionary innovation is something that makes that smelter obsolete because you don't need that technology anymore. And it should be evolutionary in terms of economics, complexity, environmental competitiveness, and better cost, energy, labour inputs, and waste profiles. That's criteria number one. The second is that these technologies need to be flexible enough to serve applications across various minerals, feedstocks, and downstream materials in order to reduce exposure to the market dynamics of any single mineral supply chain — avoiding the problem of investing heavily in one thing and then having the market flooded. JD The third would be that it's competitive at both large and small scale, enabling distributed adoption and broad deployment, even down to the mine site. The reason this matters is that China's response to any of these technologies will inevitably be to copy them. If we develop one technology that can only be implemented at one large site and China somehow manages to build the same site, they'll find a way to subsidise it and make it more competitive. Whereas if a technology can be distributed in a decentralised and democratised way among whoever needs it, it eliminates the need for those processing choke points to be anywhere in particular. The fourth point is that these technologies all carry inherent risk. While I firmly believe we need to invest in early-stage technologies and advance them rapidly — and I think government actually does quite a good job of that with DARPA and other programmes — focusing large capital on technologies that have been sufficiently de-risked, moving from the lab to a proven pilot scale, enables us to focus on things achievable in the short term. Moving from first-of-a-kind to second-of-a-kind to nth-of-a-kind facilities allows rapid scaling and the ability to attract outside capital. [36:00] NB Michael, where does Metallium fit within that landscape, and what does success look like for you? MW We're trying to disrupt the traditional business model of having to establish a new technology, attract capital, and prove it out over a long commercialisation period. We're fast-tracking all of that for a targeted number of initial feedstocks — specifically those rich in gallium and germanium, and printed circuit boards. We're focused on those because they're fundamental and critical to defence and AI technology for the United States right now. MW In a much broader, longer-term view, success would look like attracting the best and brightest minds coming out of universities to join companies like us, and effectively making metals and mining exciting again. There's a huge problem in terms of the education system. If you asked most people in the US or Australia where a lot of materials come from, they would not have a clue — because mining is seen as a dirty, antiquated, unsexy business. Engineers who used to study mining are at an all-time low in Australia; they're being pulled toward digital technology or finance. Something has to be done to attract the best and the brightest away from those more glamorous industries, and that will need capital as well as support from government organisations and big industrial players. Near-term success for us would be standing up the first printed circuit board recycling facility in the United States and being the first to recover gallium and germanium from industrial residue. Bigger-picture success would be becoming the go-to technology in the West for these feedstock metal recycling applications. But for the overall system to succeed in the West, we really have to try to attract the best and the brightest back to the physical world. NB If you were going to recap this episode for someone in one idea, what would it be? Jeff, let's start with you. JD We are now finally seeing the repercussions of hollowing out the industrial capacity of different nations and offshoring those capabilities. It's incumbent on not only the government, which is trying to do something about this, but also industry to really figure out what the impacts would be if they lose access to various materials and minerals. I don't think any of them are really prepared for the potential impact that would have. On a positive note, there are various technologies being developed. As a Rice alumnus, I'm partial to Metallium, but there are plenty of others. One of the things I really like about Metallium's strategy is their explicit dual track — build and operate to prove out the facilities, with a technology campus under development in Texas, alongside an explicit strategy of licensing this technology out to other companies. That type of approach is what's needed to scale quickly: once you've developed a technology, open the doors for others to adopt it. That's something missing from a number of the other potentially viable technologies out there. [40:00] NB Michael, one parting thought. MW With critical metals, the term gets thrown around a lot — but the chokepoint is not the ore body, it's the processing. The West owns a significant amount of resources in the ground, but it really lacks strategic control if the conversion capacity sits offshore. It's not the upstream product that can be used in an end application — it's the refined material. And if that sits offshore, countries that hold that end product can really have leverage over the West. The second point is how energy and material intensive this AI boom truly is. It is the most energy-intensive technology ever developed by mankind, bar none, and similarly the most material-intensive. There's a real lack of appreciation for that. If the US is to compete with China, the next arms race is really an AI race — and to compete, you have to have the critical metals and the processing capacity onshore. NB And if I was going to give my one parting thought, I would say this is an area that's going to take innovation. The West can't subsidise its way out of this problem. It's going to take innovation across a wide, systemic landscape. Simply concentrating on the upstream, or the midstream, or the downstream, without making sure those industries are healthy and linked and viable, is simply not going to work. NB Michael, thank you very much for joining us. We really appreciate it. MW Thanks for having me. Appreciate it too. NB And Jeff, thanks for joining us from Houston. JD Of course. Anytime. NB Thanks for listening to Let's Talk Energy. This podcast is a Rystad Energy production, produced by Elliot Busby and Bade Og. Check out the show notes for further analysis on the topics we've discussed in today's episode and find us on social media — we're at Rystad Energy on all major platforms. While you're there, please leave us a review, subscribe, and hit that like button. You can also keep up to date with us on our website. If you'd like to send us a question or reflect on today's show, or maybe you've got an idea for a future episode, email us directly at podcast@rystadenergy.com. And don't forget to join us next week for more Let's Talk Energy.