Pod Insights a16z Podcast Mining the Data for Cobalt About this episode With increased demand for better batteries, cobalt has become one of the most important metals on the planet. Kobold Metals co-founder and CEO Kurt House, John Thompson of Cornell, and a16z general partner Connie Chan join Hanne Winarsky in discussing why this happened and how technology is rapidly changing the age-old mining industry.
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Speaker 00 While many of us increasingly live in digital worlds, and even a software eats the world, we still rely on a myriad of devices that depend on resources from the physical world. One of the most important resources on the planet is cobalt, which is used in batteries, among other things. In this episode from July 2019, Kurt House, CEO and co-founder of Cobalt Metals, John Thompson, professor of Earth and Geosciences at Cornell, and Connie Chan, A16Z general partner for Consumer, talk with editorial partner Hannah Wannarski about the way technology is transforming how we find cobalt and the mining industry as a whole, as well as the science behind why cobalt is so critical for batteries,
0:42 Speaker 00 the data and knowledge behind mining today versus the past, and more.
0:48 Hanne Tidnam? Hi and welcome to the A16Z Podcast. I'm Hannah, and this episode is all about the exploration for and mining of minerals, specifically cobalt. In this conversation, I'm joined by Kurt House, CEO and co-founder of Cobalt Metals, Professor John Thompson of Earth and Geosciences at Cornell, and A16Z general partner on the Consumer team, Connie Chan. We explain why it is that cobalt is suddenly one of the most important and in-demand metals on the planet, and how technology is transforming how we find it and the mining industry as a whole. Along the way, we touch on a little bit of battery tech history and science, and how entire chapters of human civilization are driven by the search for and mining of metals, from ancient civilizations' first finding copper
1:32 Hanne Tidnam? to the major ground shift in the 1950s with geophysics and knowledge of plate tectonics. And finally, what kinds of new data sources, technologies, and techniques we can use to find more cobalt today? Everything from geophysical and geochemical data to agricultural information to old boxes collected over centuries in the basements and attics of mining companies. All of this to satisfy the incredible new demand as we enter a new age of battery metals. Why are we even sitting around this table talking about cobalt today? What is it that's suddenly so interesting about cobalt? I think of it as a color, as a... Right?
2:06 Speaker 02 Well, that's actually right. And that was the very, very first use of cobalt. The very, very first use was in dyes. And to get a particular type of blue, that was the principal way to do it.
2:17 Hanne Tidnam? And when was it discovered?
2:19 Speaker 03 The actual metal was first isolated as a metal, I think. I'm pretty sure it's 1739.
2:23 Speaker 02 So if you go back 15 years or so, its principal uses were in, were in high, sort of, high-strength steels and things like that. So cobalt demand sort of grew gradually. But everybody listening to this podcast, and presumably, listening to this podcast on their device, has 10 grams of cobalt in that device, about.
2:44 Hanne Tidnam? Some might be listening from their cars.
2:46 Speaker 02 That's true. Good point. In which case, if they happen to be driving an electric vehicle, it could be closer to 10 or 20 kilograms of cobalt.
2:54 Hanne Tidnam? It's the battery that uses the cobalt.
2:56 Speaker 02 It makes, it makes the best batteries. Everyone knows they have a lithium, sort of a lithium ion battery in their, in their phone, or, but that's a, it's a chemical reaction between lithium and cobalt oxide. And, and so it's, the two parts of the chemical reaction are essential. And the greatest energy density, greatest, greatest rate capability, how fast you can charge and discharge the battery, greatest cycle life, that kind of thing, that has some or all cobalt in it.
3:19 Hanne Tidnam? Why is that? Can we get into the science of why that's the best battery?
3:23 Speaker 02 The reason cobalt makes a great battery is that the battery in your iPhone is what we call a lithium-intercalation cathode. That's a fancy name, but what it means is lithium is the mobile ion. So a battery is, a battery has an anode and a cathode, and an ion that moves from the anode to the cathode to react chemically and form a new molecule that's more stable. The cathode looks like, kind of like a layered sandwich. It has... It has cobalt oxide, then lithium, then cobalt oxide, then lithium, then cobalt oxide.
3:53 Hanne Tidnam? Oh, like a layered cake.
3:54 Speaker 02 A layered cake, in a fully discharged sense. And you can imagine, very intuitively, why that's good. Because the lithium has to get access to the cobalt. So when you fully charge it, when you push the lithium out of the cathode, the lithium can intercalate into those spaces very easily. And put simply, cobalt forms the most stable layered structures. So as you pull lithium out, it doesn't disorder or change. And other similar metals tend to change. And when they change, then you lose capacity. Your battery fades over time. Basically, because it forms that really robust crystal structure, it forms the longest-lasting batteries. And it also has the greatest energy per molecule for a lithium oxide battery.
4:35 Hanne Tidnam? And this is by orders of magnitude.
4:37 Speaker 02 So like, right next to lithium on the periodic table is nickel. And if you made a nickel oxide, lithium nickel oxide battery, it would work OK. But on your first cycle, it would have maybe 10% less energy density. Over 100 cycles, then it would have maybe 50% less energy density.
4:53 Hanne Tidnam? So it really adds up.
4:54 Connie Chan? In this day and age when you're shopping for your iPhone, battery life is one of the key things you think about.
4:59 Hanne Tidnam? And especially with your electric vehicle.
5:01 Connie Chan? And with your electric vehicle, right. It determines how long you can drive that car for.
5:05 Speaker 03 Batteries, the whole battery technology world is really interesting because it's such an important part of our life now. So if you go back 50, 100 years, the lead-acid battery, which is the first car battery, which is still the dominant battery for starting lights in ignition, the SLI battery. But once you've got an established battery that works, that people trust, it's quite hard to displace it. So you've got to convince people with any new battery technology that it's going to deliver the right amount of charge many, many times. If you bought your phone and tomorrow had to go back to get a new battery, you wouldn't be very happy.
5:38 Connie Chan? Right.
5:39 Speaker 03 This is where cobalt is a key piece of that puzzle because it offers... A level of reliability that it's going to be hard to substitute.
5:47 Speaker 02 To sum it up simply, it's the things you really care about are rate at which you can charge and discharge it, how much energy there is per unit mass and per unit volume, and how long the battery, how long it lasted a single charge, and then how much that charge fades over many cycles. And for all those, all those elements, cobalt is cobalt superior.
6:04 Hanne Tidnam? So we went from, from using it to make a pretty color to suddenly needing it all around us.
6:10 Speaker 02 You can almost define human history by the types of metals that we were pulling out of the ground during that time. And in fact, if you look at the 8,000 years from the beginning of the metal ages to about 1970, we produced a certain amount of metals we pulled out of the ground, call that X. In the last 50 years, we pulled out the same amount again. So throughout all human history, now we've pulled out 2X. We reproduced it in the last 50 years. In the next 30, we're going to pull out another 2X, just on current trends. Okay, that's the mass of materials coming out of the ground. But here's the thing, the types of metals that we're pulling out of the ground are changing. And they're changing for both society trends and society's needs, right? So in the next less than 100 years, we have to rebuild the entire energy infrastructure of the world. Some of that requires the types of metals that we've been pulling out of the ground for a long time,
6:55 Speaker 02 but some of it requires totally new metals. And we need lots of new materials. We need lots of lithium. We need lots of manganese. We need lots of cobalt. If we're going to convert the entire automobile fleet to an electric fleet, we need vastly more cobalt than we've pulled out to date. And in order to that, we need to find new sources of cobalt. And that's what we do.
7:14 Hanne Tidnam? So how do you do that? The metal that we need is changing, and now we need cobalt. How is the way we mine, the way we source it, changing too? And actually, let's go back and start earlier. What does the history of mining cobalt look like?
7:24 Speaker 02 The act of exploration and discovery is fundamentally an information problem, right? And so the mineral exploration business is an incredibly old industry, and it's essentially driven and been driven by, by the evolution of civilization.
7:40 Hanne Tidnam? Right, it's as old as humanity.
7:42 Speaker 03 It started pretty much with copper, and that's because people have actually found copper, the metal sticking out of the ground, so they used it. They could make it into different things and make it into ornaments and so on. So this was the very beginnings of metallurgy as a science, as a discipline. It's very cool because they somehow, very creatively, worked out how to extract metals from rocks that looked green, because they have copper in them, but didn't obviously show the copper until they were smelted effectively in a very hot fire. And a fire that was a lot hotter than a campfire. So somehow these people had figured out how to get the temperature of the fire up to a level where it could reduce the copper-bearing material and extract the metal from that. And that's amazing. So that's, you know, six, seven thousand years ago.
8:24 Hanne Tidnam? And that process of exploration back then just looked like kind of studying the land, understanding and reading the earth around us. I read an article recently about the Vikings and the Age of Iron and how the Vikings, they think, were able to identify where the iron, the bog iron, was through a kind of microbial sheen, apparently.
8:44 Speaker 03 It was all based on observation.
8:46 Hanne Tidnam? Yeah.
8:46 Speaker 03 And then sort of correlation of different factors that were purely observation and experiential. Effectively, that's the original prospector, the person who could go into the ground and recognize that this had potential. And that really was the way all exploration was done until about 1950. They went to places where they could still observe the metals or the minerals then that they knew had the metals in, even if they didn't understand exactly why or how or what the concentration was.
9:13 Hanne Tidnam? But oftentimes, the minerals aren't sitting above the earth's
9:16 Speaker 03 No, not now. They'd find it on the surface and they'd keep mining down. And originally, they would mine down until they hit the water table, and then they couldn't deal with it. The Romans then started using wheels to actually de-water, so they could then go deeper. And the breakthrough in the Industrial Revolution was the steam engine, which then allowed them to make pumps that then could take the water from much, much deeper levels. So that, at that point, then they could chase it further and farther. But they didn't know where it was going. They just followed it.
9:42 Hanne Tidnam? Blindly.
9:43 Speaker 03 Yeah, so the 1950s was when we started to develop technologies. Remote technologies, geophysical technologies that could predict where things were going and predict where things might be.
9:54 Hanne Tidnam? Really recent, actually. And what were those sort of different sources of information in the 1950s that changed, that we started becoming aware of these new processes, Earth's processes?
10:02 Speaker 03 So two, really. One was what was called the plateectonic revolution of the late 60s to 70s, when people realized that the planet is dynamic and that the planet is descending off the west coast here of North America, the ocean plate is going beneath us and that's giving us all earthquakes and volcanoes and so on. And simultaneously was the development of geophysics, so the ability to detect the signature of the Earth beneath the surface in terms of its physical properties. So how magnetic it is, how dense it is, how conductive it is, we started to be able to measure those things. Those measurements, the data that comes from that correlates in some cases with the presence of metals.
10:41 Hanne Tidnam? How much does it correlate? How predictive was it really?
10:44 Speaker 03 Poorly is the answer.
10:47 Speaker 03 It's, you know, there were many false positives.
10:50 Hanne Tidnam? Okay.
10:51 Speaker 03 So we generated maps of magnetic signatures. People would say, there's all these fancy-looking anomalous bumps in the data. And they'd drill them, they'd put holes in the ground, and it turned out that one in 100 would actually be interesting.
11:03 Hanne Tidnam? Incredible amount of investment and effort and tools. Interesting.
11:07 Speaker 03 And that hasn't changed. We've got more and more tools now, but statistically, when we look at the chance of discovery, our odds are still very low. You know, we're in thousands, one in 1,000, one in 5,000.
11:17 Hanne Tidnam? Well, let's go back to cobalt and talk about what that process has looked like for cobalt, because it hasn't, we haven't, there hasn't been a reason to invest a lot of discovery in cobalt, right? Up until this point.
11:29 Speaker 02 Yeah, that's exactly right.
11:30 Hanne Tidnam? Besides pretty pictures.
11:31 Speaker 02 Right. That's right. So one of the interesting things about cobalt in particular as a metal is that you have big copper mines that are principally there because of copper. And they also have a lot of copper and a little bit of cobalt.
11:42 Hanne Tidnam? And that just happens to be because cobalt and copper tend to be?
11:45 Speaker 02 Hang out together.
11:46 Hanne Tidnam? They hang out together.
11:47 Speaker 02 Not always, but in certain circumstances they do. And so, so it was effectively the marginal cost to produce cobalt out of these mines is very low.
11:53 Hanne Tidnam? A nice little extra perk.
11:54 Speaker 02 It's an extra perk, right? You would develop a mine anyway if the, if the There was no cobalt there.
11:58 Hanne Tidnam? Right.
11:58 Speaker 02 So you have this, you have this sort of kind of gift to the world. We're going to invest in copper production and we get a little more, a little cobalt. And same thing with nickel. You get, you get a lot of cobalt associated with nickel. So that byproduct production of cobalt alongside copper and nickel was, was more than sufficient to supply the world up until, up until now.
12:19 Hanne Tidnam? When you say more than sufficient, does that mean sometimes people didn't?
12:22 Speaker 02 Absolutely.
12:22 Hanne Tidnam? So there's just piles of cobalt sitting around?
12:25 Speaker 02 Even just in the last 15 years where copper mines were, were, were developed. The cobalt was well known, its presence was well known. And investment decisions were made not to extract the cobalt from the ore, to, to extract the copper and to throw the cobalt into the, into the tailings pile.
12:40 Hanne Tidnam? Why?
12:41 Speaker 02 Because at the time, though, if you, if you did make that investment and supplied the, that cobalt, it would have tanked the market.
12:48 Hanne Tidnam? Too much.
12:49 Connie Chan? Because the demand was not there from the smartphones and the laptops and the electric cars yet.
12:54 Hanne Tidnam? Nobody needs it, quite literally.
12:55 Connie Chan? Yeah, right.
12:56 Speaker 02 Exactly. And so now the whole situation has flipped, right? So now all of those mined tailings that are full of cobalt and that are known are being reprocessed or investments are being made to go, to go reprocessing.
13:09 Hanne Tidnam? So everyone that was sitting on a garbage pile of cobalt is suddenly feeling good about it.
13:13 Speaker 02 Oh, absolutely, yeah. In fact, the largest, the large, the two largest projects to come online in the next, in the next 18 to 24 months are exactly that, are waste pile reprocessing projects. But in order to, you know, in order to convert the entire global vehicle fleet to, to electric vehicles, we need vastly more than is available.
13:32 Connie Chan? Even just current predictions, if you include global demand, especially from Asia of electric vehicles, we're likely going to run out of cobalt from known supply in less than 10 years.
13:44 Hanne Tidnam? Wow. We have this sudden increased demand. What are the sourcing efforts starting to look like? How is it changing how we actually find and source cobalt?
13:52 Speaker 02 If you look into the scientific literature and you look at gold ore deposit formation, you'll find a very rich scientific literature on how gold deposits form. You'll find a very rich scientific literature on how copper deposits form. You will not find a rich scientific literature on how cobalt deposits form because people haven't looked for it and people haven't done that.
14:10 Hanne Tidnam? There was not the same incentive at all. It just wasn't important. So John, what does that science look like for how to find cobalt? How much do we know?
14:17 Speaker 03 Not very much is the simple answer. I mean, we're kind of starting pretty at a basic level. If we wanted cobalt, we'd have gone looking for copper. And we know how to do that quite well. And we'd go looking for nickel.
14:27 Hanne Tidnam? So to kind of piggyback on other knowledge.
14:29 Speaker 03 And we'd hope that we found some copper and, oh, this is terrific. We've got a bit of extra cobalt. But there's no science basis for that. So what actually makes cobalt tick? So how, understanding what kind of liquid will transport cobalt? So what, if we have a fluid that moves through the earth, you know, it's just water dominated, maybe salty water. It moves through the earth and it interacts with a rock. Will it actually extract cobalt? We know kind of how much copper it might extract, but we don't know with cobalt. So, and if, so if it did, and then it keeps moving, then that liquid kind of comes up on the surface or comes into a different environment, would it precipitate cobalt? We don't know the answer to that either. And yet we know it did that because we can find occurrences around the world where, which are rich in cobalt.
15:12 Speaker 03 And we can see the evidence that that came from the passage of a liquid through the rock and it left the cobalt behind in cobalt minerals.
15:20 Hanne Tidnam? So that's one clue.
15:21 Speaker 03 So that's one clue. And if we understand how that worked, then can we extrapolate to other areas and predict where it might work again or where it might even work even better and give us a greater concentration of cobalt? So that's the science kind of basis.
15:34 Hanne Tidnam? And what are some of the other clues that we're starting to gather and where you want to dive deeper into? Why is this, why is this happening with cobalt? Let's keep understanding it's not that rare.
15:43 Speaker 03 When you look at the distribution of cobalt, it occurs in environments that formed from liquid rocks. So very high temperature, plus a thousand degrees. But also it's precipitating on the ocean floor, deep, deep on the ocean floor, below 4,000 meters beneath the surface. We have the nodules, these little concentrations of metal that are precipitating out of seawater. And so at seawater temperature, that's two degrees in the deep ocean.
16:06 Hanne Tidnam? And how do we know those are there?
16:07 Speaker 03 They were found in late 1800s on an expedition that was just dredging stuff up at the bottom, for the heck of dredging stuff up at the bottom.
16:14 Connie Chan? Oh my gosh.
16:14 Speaker 03 Pulled up these little round balls called manganese nodules because they're a major constituent, actually, is manganese, but they contain significant cobalt.
16:22 Connie Chan? And how interesting that that's what exploration was back then.
16:25 Speaker 03 Yeah, let's throw a net.
16:26 Connie Chan? Like, let's just drag, yeah, let's just pick stuff up out of the ocean and see what's there.
16:31 Speaker 03 The result is we know cobalt can form at really high temperatures and we know it can form at really cold temperatures. That's a huge range of conditions, range of pressures, temperatures. Now we want to get a little bit smarter and cleverer, understand which of those range of conditions will give us more cobalt relative to copper or nickel or manganese or other things.
16:49 Connie Chan? And I think that's where the data play makes so much sense, right?
16:51 Speaker 03 Exactly, that takes us into now the data world.
16:54 Connie Chan? Because rather than like comparing various exploration efforts, they're looking at places that are fundamentally so different.
17:00 Speaker 03 Yeah.
17:01 Connie Chan? Now you can track every known source of cobalt and what aspects or what qualities around that land made it particular and then pattern match.
17:10 Speaker 03 There's pattern matching that process and then there's just looking at the patterns of data until and get them to tell us a story.
17:17 Connie Chan? That almost reminds me of like the initial way of looking at the landscape with very little information and trying to pattern match the green rocks or the, you know, the iron sheen or what have you, but this is like at a much higher resolution and much greater than any one human could ever do, right?
17:32 Speaker 03 What the data world can do for us now, AI, machine learning, it is the 21st century prospector because it's not biased and prospectors weren't biased, they were just observers. So the digital world can observe and integrate and interrogate the data in ways that we humans, geologists with all our biases will never do.
17:51 Connie Chan? And we can go back and look at all the other historical known traces of cobalt and all the reports that are written that are like in PDF form right now, actually, over the last several decades.
18:02 Speaker 03 Or not, or in paper form.
18:03 Connie Chan? Or not even in PDF, yes, or in paper form.
18:06 Hanne Tidnam? So where are you pulling these different information sources from? What are they? What are the mainstreams of different kinds of information?
18:12 Speaker 02 There's a huge amount of information out there. The challenge is that it's not well structured or even digitized in some sense. You have geophysical data, which is that we're talking about things like gravitational anomalies, magnetic anomalies, electromagnetic responses, things like that. It's a whole class of data. Geochemical data is compositional data in basically a point in space and a list of concentrations at that location. Then you have mineralogical data, which is like geochemical data, but it's more complex because it gets into not just a list of elements, but actually what molecules the elements were in. And then you have things like agricultural information, right? Which are sort of indirect or topological information.
18:54 Hanne Tidnam? Meaning like what is the soil like here?
18:56 Speaker 02 And this is used for inference, right? It's not necessarily direct, you know, direct observation, but that's really important. And then hyperspectral data, which is, you know, just the wide band of electromagnetic emissions and reflection from the surface of the earth.
19:09 Hanne Tidnam? Is weather a part of that as well?
19:10 Speaker 02 Groundwater is a great source of information. So you have these, you have very, very wide sets of data. That data has been collected over centuries, really.
19:20 Hanne Tidnam? Right, it sounds like basically every piece of knowledge we have about the earth and the way the earth works.
19:24 Speaker 02 That's exactly right. Every piece of, you know, to first order, every piece of knowledge about how the earth works and what about the earth is relevant. It's just a matter of how relevant, you know, the sort of relative weightings of importance. And so in certain jurisdictions, that data has been sort of aggregated in certain ways. And there's a lot of it is public. In certain places, it's, it's, it has not even been digitized.
19:43 Hanne Tidnam? Over what kind of time scale are you looking at? Is this all like fresh, new data? Is this data from 200 years ago when people were panning for gold?
19:51 Speaker 02 Both, new and old.
19:53 Connie Chan? The good thing is the rock doesn't move that quickly.
19:56 Hanne Tidnam? That's the one constantly happens. So 200 years ago is not really that. It's still kind of recent.
20:02 Speaker 02 Unlike a lot of, it's actually interesting, unlike a lot of data analytics and data science plays, right, we are, we are looking at effectively a static system. I mean, of course, the earth is dynamic, but on a time scale, we're looking at it, it's effectively static.
20:13 Connie Chan? It's slow, yeah.
20:13 Speaker 02 So that's a very different, it's a very different data science problem, but, but we're dealing with sparse data and then we're dealing with highly, highly disparate data. So we have, we have a program of trying to aggregate all these different data sources and then, and then do two different things on it. From one side, we have our basic science approach, which is sort of how, how these ore bodies formed, cobalt or any other material we might, we might be looking for, and then looking for those sorts of indicators. And then, and then you have the really exciting thing, which is the data, rather than us asking the data questions, the data tells us stuff, right? So in this, and this is where your machine learning or statistical association modeling becomes important, right? Because the data itself can make predictions based on the patterns that it, that it sees. And
20:54 Speaker 02 that's where you eliminate the, you know, human bias and the non-systematic approach of historic exploration.
21:00 Connie Chan? Yeah. And just like to tie that back to again, like the, the demand for cobalt is so new. So there's lots of reports out there where they'll say that there's known cobalt in these places, they just didn't go mine them. If you even get all those reports and digitize them and look on a map, oh, you know, all those reports are clustered in these areas. That already surfaces some interesting sites.
21:20 Speaker 03 And fun enough, in the, if you go back, you know, 30, 40 years, a lot of the people out exploring were, were very good at identifying minerals and better than we probably are now because we rely on a lot of extra tools to do it for us now. And they recorded that presence of that mineral. So now you're not looking through all these old texts for the word cobalt because they didn't write cobalt down. They wrote down the name of a scuteroite, which is a very, you know, is a particular cobalt arsenic.
21:45 Connie Chan? So associated, yeah.
21:46 Speaker 03 So yes, they would have found that interesting, but not from a commercial perspective, just they were, just thought it was cool. They found another mineral because they were, it's like being a bird watcher. They were mineral watchers. So they recorded it. So now you've got to go through the data and find those references to that kind of thing.
21:59 Connie Chan? I just want to drill one level down, just for like fun color. Where do those reports live now? Where are these types of data coming from? I mean, I can, the modern data must be easier to access, but the old data...
22:09 Speaker 03 Boxes, boxes and boxes and boxes and boxes of paper.
22:12 Connie Chan? Boxes in, in where? In libraries?
22:13 Speaker 03 Warehouses.
22:13 Speaker 02 In dusty basements.
22:15 Speaker 03 Basements, commercial organizations.
22:16 Connie Chan? How do you even get access to all these? I mean, where...
22:18 Speaker 03 It's hard. I mean, some of the mining companies are 100 year old companies and they have 100 year old data. And so they have boxes and boxes sitting there. Some would be well archived and cataloged and some is completely unknown. And inside those boxes could be anything. It could be actual good information, a mention of a mineral, or it could be some mention of a conversation between two people. And in the conversation, they made mention that, oh, when I was in the hills last, you know, I found this rock and it has this mineral in it. And that happens to be a cobalt mineral.
22:47 Speaker 02 Twenty years ago, if you went to a new city and you wanted to find a business in the new city, you had to get the Yellow Pages for that city and you had to get a paper map for that city and you had to look up the address in the book and then you had to, and then you had to look on the map for that location. And a lot of what mineral exploration does now is exactly that. It's very site-specific. You kind of collect all the data for a new project in a new area at a new time. It's relatively easy to image the surface of the earth and the infrastructure of the earth in the way Google Maps did and to catalog all the businesses. The problem, the problem we deal with is data sparsity, right? So in some locations, we have, we have tremendous amounts of surface data density and meaningful amounts of subsurface data density.
23:30 Speaker 02 And in other locations, it's very, very data poor. And so then we have to use really sophisticated statistics, really, to try to figure out and predict what is in those material, what is in those areas where there is no data.
23:42 Hanne Tidnam? How do you actually deal with that incredible variety and huge amounts of data in some areas and very little, very old data in other areas?
23:49 Speaker 02 It's about making predictions, right? So it's about using places where there is a high density of data and you can train and make predictions and then make those predictions in areas where there isn't high densities of data and then go out and validate it, you know, by collecting new data.
24:03 Hanne Tidnam? It's interesting because in some ways it feels incredibly modern and new, but in other ways, it also feels like a kind of old-fashioned way of exploring again.
24:11 Connie Chan? Actually, a lot of this reminds me of like the original idea of venture capital, which is when kings and queens would fund these exploration efforts to look for natural resources or look for new land or whatever they were looking for.
24:24 Speaker 03 The whole idea of exploring the earth to encourage people to look for the materials that we need such that society can improve and do new things, that's an old concept. That's what developed California ultimately, was the influx of people looking for gold. And governments also fairly early, 150 years ago, started mapping the rocks on the surface because they knew if they mapped rocks that somebody would recognize associations and realize that that might have potential and therefore they'd go exploring, therefore they'd find things and that would then open up and create economic activity and so on. It's a very old cycle we're repeating, but more efficient and more effective.
25:00 Hanne Tidnam? Because now one of those explorers is a computer.
25:02 Speaker 03 Yeah, it's also going to help us actually mine more efficiently and more effectively and more cleanly. And that's really important because to me, there's no point in us going electric and having electric cars using cobalt for batteries and so on to do that if we create a big mess in terms of providing those materials. So we've not only got to find it better, we've also got to then exploit it and develop in a way that's more efficient and cleaner and doesn't have the kind of problems that we've seen all over the world.
25:28 Hanne Tidnam? Well, I think it's fantastic to think about the way that the searches and exploration for these metals have driven sort of entire chapters of human civilization. And if we think about that, you know, the age of copper, the age of iron, if we think now we're entering the age of cobalt, what are some of the ripple effects that we're going to see, you know, as we begin to more smartly mine and access this new incredibly important mineral?
25:52 Speaker 03 Trying to solve like climate change and other major issues that requires very specific use of commodities that we're not so familiar with, like cobalt. And so that changes the way we need to think about them and the way we therefore also need to exploit them. And the exploitation part is to be more selective. We've bulk mined everything, so we make big, big holes in the ground in order to get iron ore or copper out of the ground. And if we want to be really clever, we've got to only find higher concentrations because higher concentrations are more efficient. But we also now need to try and be very selective about how we mine them.
26:24 Connie Chan? Have much higher certainty.
26:25 Speaker 03 Yeah, and it's great if we go back and reprocess the tailings, which as we're doing for cobalt, and actually, the Romans were the first people who started reprocessing waste rocks to see how they did a couple of cycles of this kind of stuff. So that's not a new idea either. But it'd be much better to be really efficient at the outset and extract as much metal as we can from the less and less volumes of rock instead of moving more and more rock to get, you know, to get the metal.
26:48 Connie Chan? And from a consumer perspective, this can result in much better batteries.
26:52 Speaker 03 Yeah.
26:52 Connie Chan? Because right now, even the amount of cobalt in the battery is kind of a financial decision, right?
26:58 Hanne Tidnam? Oh, really?
26:58 Speaker 03 And sourcing this.
26:59 Connie Chan? A financial sourcing decision, right. Like, if these companies could put more cobalt than they put in today in their batteries, it would still be a better battery. So if we want batteries or iPhones that last one, two weeks without a charge.
27:12 Hanne Tidnam? Yeah, we need more cobalt.
27:14 Connie Chan? Yeah.
27:14 Hanne Tidnam? You know, there must be a cobalt craze, right? Like, everyone, everyone and their mom suddenly wants to go mining cobalt.
27:20 Speaker 03 That's the other cobalt issue, is that a lot of it comes from the DRC.
27:24 Connie Chan? Over two thirds of the world's supply. And again, remember, cobalt's not a rare metal. And so the fact that two thirds of our supply comes from the DRC, the Congo right now, is largely a function of where those copper and nickel mines historically were.
27:38 Speaker 03 And it's mined at a scale of, you know, local people who don't aren't regulated, don't necessarily do it in an appropriate manner, and use child labor and may have links to, you know, all sorts of other potential products. It's basically done by local people and DRC and other parts of the world. And they're doing it because they are impoverished and they have, you know, they feel they can probably make a better living by scraping up the material than plowing their little piece of land.
28:06 Hanne Tidnam? Well, this highly valuable material.
28:07 Speaker 03 Highly valuable material.
28:08 Connie Chan? So it's these small teams that are taking the gamble and
28:11 Speaker 03 Yeah.
28:12 Connie Chan? using shovels and very, very basic tools to go look for this.
28:17 Speaker 03 Correct.
28:17 Hanne Tidnam? So that's actually how it's happening right now still, is just small groups of people with eyeballs and shovels in the dirt.
28:24 Connie Chan? Consumers care about where their products came from now. They care about the ingredients. They care about how they were made. They care about if this came from a local farmer. So they will eventually start also caring about where their batteries came from.
28:39 Speaker 03 In the DRC where this is done, they are mining material where, which was originally a copper cobalt deposit that then suffered thousands of years of weathering. So rain came down and dripped through the rocks and it actually separated the copper from the cobalt. So the cobalt stayed near the surface and deeper down it gets more copper rich. And the mineral that they mine now is it's got this great name called heterogenite, which as you might guess is something to do with being heterogeneous. It looks all over the place. It's really messy stuff and they can just literally dig that up and put it in bags. But unfortunately, that also concentrated quite a bit of thorium, which is radioactive. So now we have artisanal miners, local people and kids who are mining
29:22 Speaker 03 bags for this stuff, which is slightly radioactive. It's not super radioactive, but it's radioactive enough to cause concern. They don't know that. They're just interested in getting bags full and getting paid for the bag of dirt that they scrape out. Some people estimate as many as 100 million people on the planet involved in some kind of activity like this at this scale.
29:41 Speaker 02 Not just cobalt.
29:42 Speaker 03 Not just cobalt, gold, diamonds, other coltan, other minerals.
29:46 Hanne Tidnam? Okay, so let's talk about what this new kind of endeavor of exploration and mining and knowledge aggregation, what does that mean on the company building side? Who do you need? What kind of, what kind of people do you need to sort of represent all those different elements?
30:03 Speaker 02 It's a fantastic question. And basically, it's, it's, it's two very different classes of people and they're both essential. Our company is effectively half made up of economic geologists, geochemists, mineral explorationists, people who have spent their careers looking in sort of the conventional manner for mineral deposits of all kinds. But then the other half is data scientists, right? And so one of my co-founders has got his PhD in quantum computing. And the other one was the chief reservoir engineer for ConocoPhillips for, you know, for many years.
30:36 Connie Chan? Oil and gas has been incredibly sophisticated in how they use technology. Because there's a very clear financial reason to go find oil and gas.
30:45 Hanne Tidnam? Right, right.
30:46 Connie Chan? But that same sophistication has not run brought over to mineral and metal exploration.
30:51 Speaker 03 We relatively well understand the environment in which we find oil and gas, and we have very good sophisticated tools to then help us do that. So the mining business is playing catch up on discovery, and it's playing catch up on exploitation as well. And so just in the last 10 years, everything now in mining is being censored all over it. Data is being gathered in the mining process. Autonomous vehicles are coming into mining, so on.
31:14 Connie Chan? And that's why the team makeup, I think, is so interesting. Because you have data scientists, you have people who are truly experts in cobalt, and they already know and have this gut feeling of where to look. And then you have people from oil and gas, who can take that sophistication and kind of bring them up to date.
31:32 Hanne Tidnam? Okay, so we're entering a new era, not just about the importance of cobalt, but also about new ways of mining as a whole, transforming a whole industry and a whole model of how we find and explore in the earth. So what changes as a result of that entire model shifting so dramatically?
31:51 Speaker 02 That's a fantastic question. You know, I think when we think about the metallurgic epochs, right? The sort of copper age, the bronze age, the iron age, the steel age, giving rise to the Industrial Revolution, and then petroleum. We're basically at, we're still in the petroleum age. From a material standpoint, I would say we're entering the battery materials age. And so battery materials will be the backbone of energy infrastructure in the next 100 years. And that requires a staggering amount of new material and different materials than we needed in the past. Cobalt being a salient one, but not the only one. I think that the tools that we're developing specifically to look for cobalt actually have a lot of generality to them.
32:34 Speaker 02 And ultimately, I think we'll probably be looking for a lot of things other than cobalt to feed the need of the battery materials age broadly.
32:45 Hanne Tidnam? So a new kind of exploration for a new age of new materials.
32:48 Speaker 02 Exactly.
32:49 Connie Chan? When you create the Google Maps, you don't just know where that store is. You know where everything else is, too.
32:54 Speaker 03 And funnily enough, it's different, but it's the same. It's still source to material to technology to people's desire to change the world. And that's the way it's been when it was the sword 5,000 years ago to electric vehicle now. It's that same process. But what we're doing is going to speed it up. Make it more efficient, use the data more effectively. So we're bringing all our tools that we have now to do the same thing as was done 5,000 years ago.
33:19 Hanne Tidnam? That's wonderful. Thank you so much for joining us on the A16Z Podcast.
33:22 Speaker 03 You're welcome. Thank you. Thank you.
Previous Applying AI in B2B Next All About Synthetic Biology Page updated September 30, 2026 · transcript and summary are machine-generated