Josh Goldman - Moving the World to Electric Vehicles, & Using AI to Discover More Raw Materials for Batteries

(Joel Beasley at 00:00:00) Hello, my friends. Today, Joel is talking to Josh, the co-founder and CTO at Kobold Metals, and they discuss how Kobold is discovering the raw materials we need to fully transition to electric vehicles, the technology behind different approaches for mineral exploration, and how they are maintaining ethical business integrity while exploring and creating new mineral mines. All of this right here, right now on the Modern CTO podcast.

(Josh (CTO and Co-founder of Kobold Metals) at 00:00:32) Here we go.

(Joel Beasley at 00:00:33) This is the Modern CTO podcast. Yeah, so tell me a little bit about your background.

(Josh (CTO and Co-founder of Kobold Metals) at 00:00:47) So I'm a physicist by training. I've always been interested in the big questions, and I've been attracted always to problems that are really hard and really interesting. And continually feeding my own intellectual curiosity has been a big theme in my life. So I studied physics in undergraduate and graduate school, though I had a diversion into history of science for a little while. I actually wrote a master's thesis on the history of the first prediction of antimatter and the British physicist Paul Dirac, who was kind of the odd man out among the founders of quantum mechanics.

(Josh (CTO and Co-founder of Kobold Metals) at 00:01:23) And then I continued in physics. I did my doctoral work in quantum computing. I was working in a lab that was known for doing really, really precise experiments. And when I say precise, I mean that the graduate student at the next desk was measuring a fundamental property of nature—in this case, the intrinsic magnetism of the electron—to 13 digits of precision. And I was trying to take this system where you could get these exquisitely precise measurements from having this amazing control over a microscopic system and trying to adapt it to a microfabricated structure that we could print on a chip and replicate many times and use the motion of a single electron for storing quantum information.

(Josh (CTO and Co-founder of Kobold Metals) at 00:02:07) So that's what I did as a graduate student, but that was 10 years ago now that I finished my graduate work. And it was pretty evident to me at the time that quantum computing had a lot of really interesting, hard science problems ahead of it, but that we were definitely more than 10 years away from actually solving problems of practical import. And I wanted to work on really hard physical science problems, but ones that were more immediately relevant to major issues in society. And at least for the physical sciences—life sciences is another matter—but for physical science, that means energy and climate change. So that was the first big transition for me.

(Josh (CTO and Co-founder of Kobold Metals) at 00:02:47) I decided that the role that I really wanted to play was not as a pure scientist or the inventor of a new device, but the kind of impact I wanted to have was making decisions—really about decisions about investing capital and harnessing the amazing opportunities for moving capital markets towards investments in clean technology. And I wanted to make decisions based on really deep understanding of the underlying science. And I think that's just so critical. So much about the energy system is driven by the laws of physics and chemistry. You really have to understand what's going on in a lot of cases in order to make good decisions. In my case, the piece that was missing is I didn't understand the energy industry very well. And what was also important to me was that I knew that if I wanted to make a significant impact on the global energy system, I really needed to understand the incumbent industry really well. And so that, in my case, meant a turn through management consultancy. I worked with McKinsey out of Houston for four years.

(Josh (CTO and Co-founder of Kobold Metals) at 00:03:52) I worked with oil and gas production companies, electric utilities. I worked with the companies that design and manufacture subsea oilfield equipment, wind turbines, large power transformers, the firms who make investments in clean tech software and natural gas pipelines. I worked on corporate finance questions, strategy questions, and operations problems that these kinds of companies faced. And so that was a background in science and in energy. And then these really came together both in the venture before Kobold and the venture that is now, as of the last three years, Kobold Metals.

(Josh (CTO and Co-founder of Kobold Metals) at 00:04:28) I left the consulting world in 2015 and joined my now two co-founders, and we were investing private equity capital in upstream oil and gas in North America. And we were using a better understanding of the physics and chemistry to predict production from natural gas wells and then use that to value assets and underwrite private equity investments. And we were working with some of the largest private equity firms in the country on this. And the technology worked really well. I wrote the software that implemented our methods and the economics on top of that that valued these assets, and we used that for underwriting over a billion dollars in private equity deals.

(Josh (CTO and Co-founder of Kobold Metals) at 00:05:07) But as of the summer of 2017, there were really three big motivations that were different. One was the goal all along was to work on transitioning to sustainable energy. And natural gas is a lot better than coal, and the carbon impact of decreased use of coal, particularly in the power sector in this country in favor of gas, has been really quite significant. But ultimately, there's a lot of urgency to climate change, and I want to do better than switching to natural gas. And as we started working on this in the summer of 2017, once we stopped and thought about it, it was pretty clear which way the winds were blowing.

(Josh (CTO and Co-founder of Kobold Metals) at 00:05:45) And the opportunity for growth in electric vehicles was just so compelling. And the automakers were just starting to make big commitments to designing and producing electric vehicles, really driven by public policy. You had countries in Europe starting to commit to phasing out internal combustion engines. You have really aggressive policy in China favoring electric vehicles. And there were new announcements of model launches or commitments to produce this or that many more electric vehicles every week.

(Josh (CTO and Co-founder of Kobold Metals) at 00:06:21) And the same thing's happening again now, and they're getting more and more aggressive. And if we just looked at the way the industry was going, it was clear that this was a macro trend and this was a decadal trend. And the big question to us is, where are the raw materials going to come from to produce all of these electric vehicles? That's really the motivating market opportunity for Kobold. If you look at just electrifying the light duty vehicle fleet—passenger cars and light trucks—you look at just the batteries and the related parts of those electric vehicles, the amount of new nickel, cobalt, copper, and lithium that we need is $6 trillion worth of metal beyond all the current reserves of all four of those metals.

(Josh (CTO and Co-founder of Kobold Metals) at 00:07:07) So it's just an enormous opportunity, and we've got to go find these resources if we're going to have all these electric vehicles on the road. That was the opportunity that we set out to pursue. And ultimately, the question is, can we build a firm to do better? And I think we'll probably spend most of our time talking about technology for mineral exploration because that's the really exciting part and why there's so much opportunity to do that.

(Joel Beasley at 00:07:31) Is that your product? Is that what you guys build?

(Josh (CTO and Co-founder of Kobold Metals) at 00:07:33) Yeah. Well, so we do build technology, but ultimately we build technology and we use it ourselves. What we do as a company is we actually go acquire mineral assets. That means exploration licenses. And then we send our geologists to the field to go look at the rocks, sample the rocks. We commission new geophysical surveys, and then we drill the holes to make discoveries. And so then we own the mineral resources. We will eventually, most likely, sell the properties to mining companies to build mines. Our technological edge is the software that we write.

(Josh (CTO and Co-founder of Kobold Metals) at 00:08:12) We have an unusual team in that we have about a third exploration geoscientists, and the rest of the company are data scientists and software engineers who are building technology to help guide our search. But all the software we build, we're using to guide our own exploration decisions. What land do we buy? Where do we go explore? What kind of data do we collect?

(Josh (CTO and Co-founder of Kobold Metals) at 00:08:37) Given what we learned from collecting that data, what kind of data do we collect next? And where do we drill the discovery hole? Where do we drill the next hole after that? And it's all of those decisions that our technology is meant to guide. So we don't actually sell or license the software, but we put it into practice for our own use on those kinds of decisions.

(Joel Beasley at 00:08:57) And this is a fairly old trade as far as geologists figuring out where oil is or rocks are, correct?

(Josh (CTO and Co-founder of Kobold Metals) at 00:09:05) Absolutely. It's—humans have been looking for mineral resources for millennia, really. And if you think of mineral exploration in particular, it goes back to the Iron Age and the Bronze Age. We needed the metals for making tools. And humans have been walking the Earth for as long as there have been humans, and they've been noticing unusual things and noticing places that have valuable things—places to find water, places to find food, places to find the right kind of stone for making tools, and places where you can find rocks that you can turn into metals.

(Josh (CTO and Co-founder of Kobold Metals) at 00:09:43) And sometimes the rocks give you some pretty good clues. If we think about copper exploration, for example, copper minerals where the ore deposit forms, when they're exposed at the surface, the copper is oxidized, and one of the minerals you can form is called malachite. And so the places where you have copper ore deposits sticking out at the surface, many of those have been known about since ancient times. This is true across the industry where the really difficult to find deposits are the ones that are still remaining. The ones that are easy to find, the ones that are sticking out of the surface, those have been found.

(Josh (CTO and Co-founder of Kobold Metals) at 00:10:24) The next most difficult set of deposits are the ones that required modern technology to find. A good example of this—we spend a lot of effort and capital exploring in Central Africa for copper and cobalt. The deposits that stick out of the ground, as I said, have been known about for a very long time. The main way of finding ore deposits was to go sample the soil. And you could walk in every 500 meters on a regular grid, send a crew out, dig down below the topmost decayed plant matter, and dig down far enough, usually a foot or two, to where the soil you're sampling is the product of the rocks underneath it.

(Josh (CTO and Co-founder of Kobold Metals) at 00:11:02) The rainwater essentially, and other geologic processes, have eroded the rock, have weathered the rock, and the soil further down tells you about the rock underneath it. And then you send that soil sample to a laboratory, and you measure the amount of copper and cobalt and other metals in it. And you look for the places where you have unusually high copper in the soils, and then you drill a hole and see what's underneath. And that's the way that most of the ore deposits have been found in the Copper Belt. And now we have to do better than that.

(Josh (CTO and Co-founder of Kobold Metals) at 00:11:35) Those have largely been found too. So now we're talking about deposits that are deeper, and they're concealed by other layers of rock that don't have an ore deposit in them. And somehow we have to see through those, and we have to find the ones underneath. They're almost certainly there. There's nothing about the way that these ore deposits form that makes them form near the surface where we can find them.

(Josh (CTO and Co-founder of Kobold Metals) at 00:11:58) That's just where we can see them. We're looking under the lamppost because that's where the light is. But there's ore deposits all over the place, and it's about developing the right technology to be able to find the signals in the data.

(Joel Beasley at 00:12:11) How do you find them? Why can't you just take an X-ray of the ground?

(Josh (CTO and Co-founder of Kobold Metals) at 00:12:16) The reason is because you probably have to go deeper than you can see with some of those techniques. And so the ore deposits we're looking for are potentially hundreds of meters under the ground. And there are ways of detecting that from the surface, but when you're making measurements on the surface, we're trying to resolve things in 3D. And that's a very challenging problem.

(Josh (CTO and Co-founder of Kobold Metals) at 00:12:42) There's lots of techniques that can penetrate somewhat deeper. You can, for example, make measurements of the gravitational field at the surface. And that tells you something about the density of the rocks that are underneath. You can make measurements of the magnetic field at the surface, or you can also do this from an aircraft. And that tells you something about the magnetic properties of the rocks underneath.

(Josh (CTO and Co-founder of Kobold Metals) at 00:13:05) Iron is magnetic. Lots of iron minerals are magnetic, and some other minerals that contain nickel and cobalt are likewise. And you can do other really cool things. You can use electromagnetic methods where you actually have a big coil of wire. Either you set it on the ground or you carry it from a helicopter, and you pulse the coil of wire, and you cause currents to flow in the rocks in the subsurface.

(Josh (CTO and Co-founder of Kobold Metals) at 00:13:34) And so you can make a map of the conductivity of the subsurface. And sometimes the ore is conductive and the other rocks aren't. That's the best case scenario where you can look for the ore body directly. Sometimes there isn't a contrast between the place where there's a lot of cobalt and copper and nickel and the place where there isn't. But you can use it to make a map of the rocks under the surface.

(Josh (CTO and Co-founder of Kobold Metals) at 00:13:57) And then you can use that map to help figure out, well, where are the places where we might have high concentrations of these various metals? And those various methods—those are all called geophysical methods. Those things aren't unique to Kobold. Those are standard techniques in the industry. The problem is that we have to bring much more insight out of these data sources, because the real fundamental fact about mineral exploration—I said it's hard, I said the easy deposits have been found—what I mean by that is that success rates in the industry are well under one in 100. If you pursue 100 opportunities, you drill a hole in 100 different properties. On average, the number of discoveries that become mines is zero.

(Joel Beasley at 00:14:52) Wow.

(Josh (CTO and Co-founder of Kobold Metals) at 00:14:53) So it's really remarkable, right?

(Joel Beasley at 00:14:56) That's persistence. You guys must have a lot of drive and determination.

(Josh (CTO and Co-founder of Kobold Metals) at 00:14:59) Well, we certainly have that. But the goal is to do much, much better than that. Most holes don't find mineralization. And in the industry, what that means is just an enormous amount of capital and time gets wasted on properties that don't turn into mines.

(Josh (CTO and Co-founder of Kobold Metals) at 00:15:19) And what we need to do is—we're not expecting everything we do is going to be a success—but we need to have far fewer failures. And when we fail, we need to fail cheaper, we need to fail faster. And ultimately, if we have five or 10 discoveries out of each 100 properties that we pursue, that would be a smashing success. And that will be by far the leader in the industry.

(Joel Beasley at 00:15:48) I have a question.

(Josh (CTO and Co-founder of Kobold Metals) at 00:15:50) Mm-hmm.

(Joel Beasley at 00:15:51) So this is going to be probably a silly question, but I want to clarify something. You had said when finding these metals, one of the tests is measuring the gravity. Is gravity not the same everywhere on Earth?

(Joel Beasley at 00:16:06) That's a good question. Explain.

(Josh (CTO and Co-founder of Kobold Metals) at 00:16:08) Yeah. So gravity, as you know, is massive objects are attracted to each other. Right? You stand on the Earth because you are attracted by the gravitational pull of the Earth. And particularly, that means you're attracted by every other atom on the Earth is pulling you towards it.

(Josh (CTO and Co-founder of Kobold Metals) at 00:16:26) And so what that adds up to is you're pulled towards the center of mass of the Earth. But there are local variations in the density of the materials around you. And you know that a block of lead is much denser than the same volume of water. And that means if you are near rocks that are much more dense than other rocks, the gravitational field is going to be stronger over the area that has much higher density of rocks. And so we're talking about really minute variations in the gravitational field.

(Josh (CTO and Co-founder of Kobold Metals) at 00:17:01) It's almost exactly the same, but it's the small variations that are sensitive to the density of the rocks underneath you.

(Joel Beasley at 00:17:09) Is there any place on Earth where this is perceptible by humans?

(Josh (CTO and Co-founder of Kobold Metals) at 00:17:13) I don't think so.

(Joel Beasley at 00:17:15) So it's like micro changes. Very micro changes.

(Josh (CTO and Co-founder of Kobold Metals) at 00:17:18) That's exactly right.

(Joel Beasley at 00:17:20) I want to make sure next time I'm hiking, maybe stay away from those mountains. Be careful when I'm out in Colorado. I don't want to get sucked down.

(Josh (CTO and Co-founder of Kobold Metals) at 00:17:27) You will be, if your scale were sensitive enough, you would be slightly heavier.

(Joel Beasley at 00:17:36) This is good.

(Josh (CTO and Co-founder of Kobold Metals) at 00:17:37) Near denser rocks, when you're standing on top of denser rocks.

(Joel Beasley at 00:17:41) Yes. I am learning a lot. Thank you for that.

(Josh (CTO and Co-founder of Kobold Metals) at 00:17:45) Mm-hmm.

(Joel Beasley at 00:17:45) So I'm curious. You wanted to move into this whole industry because you wanted a direct impact. Right? We're making all these EV batteries. They're just on the rise.

(Joel Beasley at 00:17:55) Every couple months I hear Elon Musk is doing something new with the gigafactory. Right? So we're building all these batteries. They need these metals. You're helping create more efficient ways to find these metals.

(Joel Beasley at 00:18:08) And you're primarily doing it through working with geologists and data scientists and improving the methods that you use to reduce the amount of failures to find these models. In the process, you are buying rights to these lands so that you'll own them and do the proof of concept and everything like the proof drilling. But my question to you is, which comes first, the chicken or the egg? If you think that, like in South Africa, I think you mentioned you were looking for cobalt. If you're going to look for cobalt in South Africa, will you just start calling the government up and be like, hey. Or do you call private landowners?

(Joel Beasley at 00:18:49) Or do you have a geologist who, like, oh, in college, I think that felt rich in cobalt when I was there for some. So you get some inside information like, how do you even begin to form your process of where you want to go? And then to make the question harder, then how do you begin to negotiate those contracts?

(Josh (CTO and Co-founder of Kobold Metals) at 00:19:07) There's so many good questions in there about how do you find the areas that are interesting, and then how do you actually get access to properties. And ultimately, we want to own the asset, and we want to invest many millions of dollars in exploration on a property. We're only going to do that if we have secure interest in the property. You've asked the sort of central questions of the exploration industry there. I'll start with the deal side of it, and then we'll come back to the how we look because that's the much richer topic.

(Josh (CTO and Co-founder of Kobold Metals) at 00:19:40) Once you've found an area that you're interested in, you have to find out if somebody else already owns the exploration rights. And if no one else already owns the exploration rights, then you stake a claim. And one of the things that's really unusual about the United States, we inherit this from the UK, is private mineral rights. If you're a landowner in the US, usually the land ownership comes with the subsurface as well, the mineral rights. And your particular property at your house, it might have been severed.

(Josh (CTO and Co-founder of Kobold Metals) at 00:20:11) Someone else might own the mineral rights now. But at least those things are probably still in private ownership. That's really unusual in the US, and it's really one of the reasons that the oil and gas industry developed in the United States is because their private landowners are interested in getting royalties from people making discoveries and producing oil or gas or minerals on their land. Usually, not if it's their backyard, but if it's their ranch or their farmland or something like that, then they really are. But on federal lands, of course, the federal government owns the mineral rights.

(Josh (CTO and Co-founder of Kobold Metals) at 00:20:44) And elsewhere in the world, almost everywhere, the mineral rights are owned by the state or the crown. And so in the US, when you stake a claim, you quite literally stake a claim on federal land. You go to the four corners of your claim and you pound a stake in the ground. You leave a tag on the stake, and then you go do some paperwork with the BLM afterwards to say, I have staked a claim in this location. That's actually how it works in this country.

(Josh (CTO and Co-founder of Kobold Metals) at 00:21:13) In other parts of the world, the process is different. In most Canadian provinces, you can stake a claim on a website. And you have to qualify your company to be a mineral rights holder, but then you go click the boxes for the cells you want and you pay the fee, and then you own the exploration license. And you don't own it forever. You own it usually for a period of a year or two years, and you get to keep it so long as you continue to meet the obligations of a license, which means you spend money on exploration and you provide the receipts, and you submit a technical report to the government.

(Josh (CTO and Co-founder of Kobold Metals) at 00:21:48) And sometimes you have to submit the data that you collected too. And these things all, after a period of a few years of confidentiality, they become public record. That's one of the great data sources that we're using is the entire exploration history of all the companies who've worked on this ground. Usually, discoveries are made on ground that someone else has looked at before. And in fact, many companies have looked at it before.

(Josh (CTO and Co-founder of Kobold Metals) at 00:22:11) But they haven't asked the right questions. They haven't collected the right data. They weren't looking at it in the right way, or they weren't looking deep enough, for example. And we aggregate all of this data from historic exploration activity. We scan and OCR and index and do natural language processing and other techniques on the text of these reports, and we can go find insights in these things that geologists have recorded, but they haven't understood what they meant.

(Josh (CTO and Co-founder of Kobold Metals) at 00:22:40) That's one of many important data sources. So we have mineral claim. In other places, we work a lot in Zambia, and the process is relatively similar. You submit an application, you draw the boundaries of your claim, and you have to write down what your exploration plan is. And, again, there's a process for how you maintain the license over a number of years.

(Josh (CTO and Co-founder of Kobold Metals) at 00:22:59) In a lot of parts of the world, there's really interesting ground, but somebody else already owns the exploration rights. And so you have to go negotiate with the landowner. And that's what I spend a lot of my time doing is approaching other companies who own a property of interest and trying to establish a partnership with Kobold. We have to really understand what it is that the other party wants in that opportunity. Sometimes, it's a small company who doesn't have the means to do exploration themselves, and they're really happy to have a technologically sophisticated and well-funded partner come in and do all the work.

(Josh (CTO and Co-founder of Kobold Metals) at 00:23:36) And then we just have an option to buy them out for a large amount of money if we're successful, or we have a stream of payments they may get or something like that. Other times, the company, maybe they've invested in it for a while, but they don't want to take it further. They have, they're out of good ideas, and they have other uses for capital. Or they want us to have a go at it, but they want the opportunity to contribute again down the road. It's like, okay.

(Josh (CTO and Co-founder of Kobold Metals) at 00:24:06) Well, you guys made a discovery. Now we have to, there's, we can get into it, but there's many other steps you have to do before you create a mine. And those things cost millions to hundreds of millions of dollars, but they're much lower risk. Other companies may say, okay. If you guys are successful, I want the opportunity to invest again.

(Josh (CTO and Co-founder of Kobold Metals) at 00:24:23) And so we'll structure a deal with them where we become joint venture partners. And so we do a lot of these deals because there's lots of ground that we want to work on that's already owned by someone else. And there's lots of data that's not in the public domain, and we want access to it. And other companies are really excited to work with us because we're bringing great geoscience expertise, and we're bringing really new ways of looking at data and ultimately things that will be more likely to make us successful in exploration on their property. So there's been a lot of interest in working with us.

(Josh (CTO and Co-founder of Kobold Metals) at 00:25:00) And, obviously, you need the resources to be able to invest in exploration, which we have. So that's the answer to one of your questions about how you get access. The bigger and the much harder question is, where do you look? There's really different answers to that question at different length scales. There's the global question of where in the world do you even start?

(Josh (CTO and Co-founder of Kobold Metals) at 00:25:28) And that has a bunch of ingredients. Geology is one, call it one axis. Where are the places in the Earth's crust that are most likely to have had the right conditions for forming an ore deposit? And maybe let's digress on this for a second of, what is an ore deposit anyway? What are we looking for?

(Josh (CTO and Co-founder of Kobold Metals) at 00:25:50) We're looking for cobalt. We're looking for nickel. We're looking for copper, for example. Well, almost any rock you pick up has a little bit of these metals in it. And if you look at the average over the Earth's crust, well, we restrict ourselves to the continents and then to the upper part of the crust.

(Josh (CTO and Co-founder of Kobold Metals) at 00:26:08) And cobalt, copper, nickel, they're all 20 to 40 parts per million. That means in every kilogram of rock, you'll have 20 to 50 milligrams each of cobalt, copper, and nickel. Well, you could extract that. The technology exists to get that metal out. It's just really, really expensive to do so.

(Josh (CTO and Co-founder of Kobold Metals) at 00:26:30) And there's a reason that I think you can buy a car for 20 or $30,000 and not 20 or $30 million. If you had to make it atom by atom, it would cost much more. And so the things that you mine are places where nature has taken the sort of background average in the rocks, managed to concentrate them, in the case of these metals, by something like 200 to 500 times concentration, up to something like 1%. That's what you're actually mining. An ore deposit is a place where nature has collected the metal from a really large volume of rocks, and it's dropped it out in somewhere that's much more concentrated, where it's more like 1% of the mass of the rock is the metal that you want.

(Josh (CTO and Co-founder of Kobold Metals) at 00:27:16) And then we do the rest of it with industry, taking it from 1% to 100%. And that's something that you can do in the case of copper for, you know, a dollar or two a pound. And that's a remarkable thing. So you start with a kind of global understanding of geology, of where in the world is it likely that these kinds of processes have happened. And that includes both places where you've found some ore deposits already, and so you know those processes at work, and places where there aren't mines, but we have really good reason to think that there really could be.

(Josh (CTO and Co-founder of Kobold Metals) at 00:27:50) The other axis is everything else that's not geology. You got to be in a part of the world where we can operate in high standards of business integrity, where we can maintain our interests and properties without having to do anything unethical. And that's true of most of the world, but unfortunately not all of it. And there's really great rocks in places where it's really hard to do business. And, you know, if ultimately, there's no value if we find something but you can't mine it, or you don't want to mine it because it's a really sensitive environmental area, or the community where it occurs is not supportive of mining.

(Josh (CTO and Co-founder of Kobold Metals) at 00:28:32) Ultimately, there are really great projects that get stuck because you can't move the ore deposit. It's under the ground and it's a, you know, it's a large volume of rock. And so if you want to mine it, you have to mine it where it is. And there are some places that are much more receptive and much less environmentally sensitive. And so you start with those places.

(Josh (CTO and Co-founder of Kobold Metals) at 00:28:52) Then we aggregate enormous amounts of data, data we get from our partners, data that we scrape from geologic surveys or that they provide to us, data that we go collect ourselves, many, many thousands of different datasets, and many millions of pages of these reports. And we generate models that tell us regions that are the most interesting. And then we go into those regions, and then we do really detailed studies where we're mapping out the specific features that might contribute to the formation of an ore deposit, making predictions about whether the right rocks are there, whether the right kinds of faults are present. And then we go out and test our predictions in the field. We actually send our geologists to go out and sample the rocks where we predicted the right kind of rocks would be.

(Josh (CTO and Co-founder of Kobold Metals) at 00:29:38) We commission airborne surveys to go make those measurements, and then we update our models according to what we learn in the field.

(Joel Beasley at 00:29:44) That's pretty smart. About five or six years ago, I live in Florida, so I was really frustrated. It was the summer and it was very hot yet again. It's like there's hot and hotter in Florida, right?

(Joel Beasley at 00:29:57) And humid. But I had gotten to do some traveling and go to other cities and I found it was very difficult to understand what type of weather I wanted. Like heat, humidity, the mixture of the two, it's just so different everywhere. And so I built this application where I pulled all the data from the Weather Channel. They had something called like Weather Underground or something.

(Joel Beasley at 00:30:19) Put it all in through an API and then I wrote this small app for myself on my local computer called similar to. And I had to look at the different averages across the seasons in Florida and then different averages across different places I have been to before, spent time in different seasons. And I found, like I was trying to find the ideal weather for me. Like, I wanted it to be similar to where I'm currently at.

(Joel Beasley at 00:30:47) But just, like, on average, slightly less humid and slightly cooler. So I was having to put in these properties of, you know, the different weather attributes that I wanted and find things that were similar to it. It sounds like you guys are doing something similar where you find places where there are ore deposits. You create all these attributes, but you're collecting, you don't have a Weather Underground. You have to go around and pull all this data together from all these other sources.

(Josh (CTO and Co-founder of Kobold Metals) at 00:31:08) You summed it up so well. That's, the latter part is so critical. There wasn't a data system around that allowed us to store and to programmatically access all the different kinds of data. That's one of the things that's been, I think, a big limitation in the industry is there's many different types of data. And then there are many sources of each of those types of data.

(Josh (CTO and Co-founder of Kobold Metals) at 00:31:32) So that gravity and the magnetic data that I talked about, that data is difficult to integrate effectively with geologic maps and satellite imagery and geochemical measurements. There isn't a system that just allows you to work with all that data at the same time. And so we built it. Your comment about, you know, similar to—you're spot on, which is we are gathering as much data as we can get on known ore deposits so that we can train models on those, and we can go predict where there are similar kinds of features. And sometimes that data exists in the public domain. Other times, we're going out and collecting it.

(Josh (CTO and Co-founder of Kobold Metals) at 00:32:14) We'll partner with companies who own deposits that exist today that haven't been mined yet, and we will go collect data on those deposits so that we can build those models. And that means taking measurements of holes that they've drilled previously. That means taking measurements from aircraft or on the surface above those deposits as well. At the same time, no two ore deposits are alike. You know, you go look at a mountain or even a small rock outcrop, and the variability of the rocks is really striking.

(Josh (CTO and Co-founder of Kobold Metals) at 00:32:47) And ore deposits are much the same. They're, you know, different geometries and different compositions. And that's one of the things that makes the problem so hard is that the next ore deposit you find isn't going to look exactly like something else. Understanding what are the underlying processes that led to create these ore deposits in the first place? And how can we predict where those occurred?

(Josh (CTO and Co-founder of Kobold Metals) at 00:33:09) How can we identify enough features that are the same that we can have confidence that a model that we train here is going to be useful there? We also do a lot of work with synthetic data, where you can imagine an ore deposit with particular properties or geometries. And you can say, well, these rocks typically have this distribution of density and this distribution of magnetic properties. Let's create a million hypothetical ore bodies, and then let's model what the data would look like if that's what was going on under the surface. And now let's train a model on that.

(Josh (CTO and Co-founder of Kobold Metals) at 00:33:47) And let's use that to go try to find ore bodies that no one has seen before. And I think this is a really powerful method. It's much more general than what I just described. And it's only possible now. It's only possible because it's really easy to run large ensembles of computations.

(Josh (CTO and Co-founder of Kobold Metals) at 00:34:06) Getting access to distributed computing infrastructures is really easy today in a way that wasn't even just a few years ago. Of course, it was possible. But just being able to have this as a kind of day-to-day thing that data scientists can do without a whole lot of thought is only possible just in the last couple of years.

(Joel Beasley at 00:34:25) Has this been hard for you to attract these top talent? I mean, you have to have brilliant people working for you to do this. What you're describing is whole other level. Right? Because you have to have self-motivated, brilliant people at scale.

(Joel Beasley at 00:34:42) That's the most difficult thing in the world.

(Josh (CTO and Co-founder of Kobold Metals) at 00:34:44) Yeah. It is. I'm glad you brought this up because the team is totally critical. The only thing that I disagree with what you said is the at scale part. We actually have a relatively small team.

(Josh (CTO and Co-founder of Kobold Metals) at 00:34:56) We will have a modestly larger team. But this is not something we do with thousands of people. This is something we do with dozens of people. And, of course, we have larger teams actually in the field on our assets. Finding the great people is not easy.

(Josh (CTO and Co-founder of Kobold Metals) at 00:35:13) Attracting them has been actually really easy. And, you know, if we think about the data scientists and software engineers we hire, the data scientists on our team, it's not sufficient to be a great data scientist. They have to be great scientists. They have to work with physical data. And we've talked to a lot of great data scientists who've applied for opportunities at Kobold.

(Josh (CTO and Co-founder of Kobold Metals) at 00:35:37) The kind of data that you work with that's generated by user-triggered user actions in applications or server events, the kinds of things that, you know, fill up your logs, that data is very different than physical measurements and the kind of uncertainties and the kinds of problems that you get from making measurements of gravitational field or the magnetic field and the way that you have to handle it. And then it's even more challenging in geoscience because there's a lot of qualitative data. There's long text descriptions of rocks with variable kinds of names and meanings. Our data scientists have to have worked a lot with physical data. They also have to just be great scientists.

(Josh (CTO and Co-founder of Kobold Metals) at 00:36:23) They have to learn really hard concepts, and they have to think really deeply from physical principles. And that really characterizes all the folks we've hired. They have been really excited to join Kobold because they're using data science not to increase user engagement in an application, not to serve up a product recommender system. They're using it to solve really hard scientific problems. And there's a lot of folks like me who are just really intellectually curious and want to work on cool things.

(Josh (CTO and Co-founder of Kobold Metals) at 00:36:56) And then the impact of it, if we're successful, is that we're providing the raw materials for energy transition. And so it's just—yeah, that's been really compelling. The flip side of it is we need awesome exploration geoscientists too. We needed to attract people to come join a startup company and a company not run by people coming from the mineral exploration industry. And there's just this really strong sense of need in the industry that we have to be more successful as explorers.

(Josh (CTO and Co-founder of Kobold Metals) at 00:37:27) We have to think differently and use data differently in order to unlock the next wave of discoveries. And so we have an extraordinary team, and we have people who've made—we've made nearly 20 discoveries between them, including some of the largest discoveries of the last generation. And they work kind of arm in arm with data scientists. They're on one team. They're not—there isn't a geoscience group and a data science group.

(Josh (CTO and Co-founder of Kobold Metals) at 00:37:55) There are exploration teams looking for copper and cobalt in Zambia. And that team has data scientists and geoscientists on it working together. And then there's an awesome but small software engineering team that's building really the tech architecture and the backbone that enables all of this work. So, yeah, it's a small team, but really just an amazing group of people to work with. And it's—ultimately, the product here is insight.

(Josh (CTO and Co-founder of Kobold Metals) at 00:38:26) We're trying to come up with better ideas about what decisions we should make, what ground to acquire and what data to collect, better ideas about how to figure out what's under the surface of the Earth. And we need a lot of big brains on that problem with really, really creative ideas and ways of looking at the data. And that's what we've built.

(Joel Beasley at 00:38:48) And people are taking notice too.

(Josh (CTO and Co-founder of Kobold Metals) at 00:38:49) And people are—

(Joel Beasley at 00:38:49) I saw in my prep notes that Bill Gates and Jeff Bezos both invested in the company. Is that for real? Or is that just—what is that?

(Josh (CTO and Co-founder of Kobold Metals) at 00:38:57) Yeah. That is for real. And the vehicle is called Breakthrough Energy Ventures, and it was Bill Gates' idea to create this fund. And he recruited a bunch of very well-known business leaders from around the world to co-invest with them because there's a bunch of really hard science and technology problems that can make a big impact on climate if we're successful, but they require a willingness to make investments in challenging technologies and to take a really long time horizon. And so that was the impetus behind creating a fund, and Kobold was one of their first investments.

(Joel Beasley at 00:39:40) Did you get to hang out with them?

(Josh (CTO and Co-founder of Kobold Metals) at 00:39:42) I have at the investor conference. Well, it didn't happen in 2020. But no, the investors in the fund are very keen to understand what it is that these technology companies are actually doing. And so I have in fact presented to them.

(Joel Beasley at 00:40:00) Nice. That's actually pretty cool.

(Josh (CTO and Co-founder of Kobold Metals) at 00:40:02) Yeah.

(Joel Beasley at 00:40:02) There's so many great people out there on the East Coast. So I'm in Florida. So all of my opportunities for in-person stuff that's easy is just the people here, but there's so many more people in Silicon Valley. I was telling my wife, I was like, maybe we need to go get a second place out there just so when I'm doing the show, I can start meeting some of these really great people that are out there.

(Josh (CTO and Co-founder of Kobold Metals) at 00:40:23) The world is a little different now. Of course, we're—

(Joel Beasley at 00:40:26) I know.

(Josh (CTO and Co-founder of Kobold Metals) at 00:40:26) We have our other foot in Silicon Valley, and Andreessen Horowitz was our co-lead for our initial investment round.

(Joel Beasley at 00:40:34) Oh, I know a couple of guys over there. They're really great people, man.

(Josh (CTO and Co-founder of Kobold Metals) at 00:40:37) They're amazing. Yeah. Yeah. We have a head office in Berkeley, California. But we are hiring all over.

(Josh (CTO and Co-founder of Kobold Metals) at 00:40:46) And we are a remote-first company. We have people in Sudbury, Ontario, and Vancouver, and in Madison, and Houston, and New York, and Baltimore, and Birmingham, and Salt Lake City, as well as in the Bay Area. And so we're looking for great talent in data science, software engineering, and geoscience, and we find them all over the world. We have a team in Zambia. We're building our portfolio and our program in Australia.

(Josh (CTO and Co-founder of Kobold Metals) at 00:41:18) We're looking all over the world and we're building a team with great people in those disciplines wherever they happen to be.

(Joel Beasley at 00:41:24) And how can people learn more about applying for jobs at Kobold?

(Josh (CTO and Co-founder of Kobold Metals) at 00:41:28) We certainly advertise positions in those disciplines. They can reach out to me or on our jobs page. We're always looking for awesome people in all three of those disciplines.

(Joel Beasley at 00:41:40) Can you spell the website?

(Josh (CTO and Co-founder of Kobold Metals) at 00:41:41) Koboldmetals.com. It's unusual too as a company. We're basically all technical. Everybody's job in the company is exploration. We don't have a sales team.

(Josh (CTO and Co-founder of Kobold Metals) at 00:41:55) We don't have a marketing team. We don't even really have a business development team. Obviously, we have to do a handful of transactions and there's a few of us who work on that. But really, it's about developing technology and working on our properties. And when we make discoveries, we'll have no problem selling either the asset that we discovered—an ore deposit ready to be mined.

(Josh (CTO and Co-founder of Kobold Metals) at 00:42:22) Or if we build the mines ourselves, that's not our base plan, but we may ever construct a mine, we'll be selling cobalt, copper, and nickel, and lithium. Those are global commodities. You have no problem finding a market for those.

(Joel Beasley at 00:42:36) So I was at the gym today with my trainer and she was like, oh, who are you talking to? I was like, oh, this cool guy with Kobold Metals. And it's kind of explaining the EV thing and the batteries and there's—and I didn't have a good explanation about how—I remember I saw this show called Adam Ruins Everything. Have you ever seen this Adam Ruins Everything?

(Josh (CTO and Co-founder of Kobold Metals) at 00:42:58) I don't know the show, actually. I've heard of it, but I'm not familiar with it.

(Joel Beasley at 00:43:02) So it's a show where this guy ruins standard concepts that we think we understand. And one of them that came up, it went for several seasons. It's kind of funny. It's a bit punchy, but he talked about electric vehicles. And my rough memory of it from a couple years ago watching it was something along the lines of the amount of inefficiency in actually mining and making the vehicles and the carbon and all of that.

(Joel Beasley at 00:43:32) Like, they're actually not that great for the environment. Can you help me develop a better understanding of that?

(Josh (CTO and Co-founder of Kobold Metals) at 00:43:40) Yeah. I think what is true is that producing raw materials takes energy. And depending upon the source of your energy, there will be carbon emissions associated with that unless you're producing carbon-free energy. The battery in an electric vehicle weighs hundreds of kilograms, maybe even the better part of a ton depending upon the size of the battery. So it is true that you have to account for the impact of the raw materials for the electric vehicle as well.

(Josh (CTO and Co-founder of Kobold Metals) at 00:44:14) What's amazing about electric vehicles though, is that once you produce the raw materials to supply that vehicle, you can recycle them. And so all of our assumptions, you know, that $6 trillion of new materials that I gave you at the beginning of our conversation, that assumes 100% recycling of all the raw materials in an EV. So the point is once you've produced enough of these critical materials to transition internal combustion engines to electric vehicles, you have a big enough stock and you can recycle them. And recycling rates are really high, greater than 90% for most of the raw materials. And so, basically, you have a circular economy or at least it's possible, because you can keep those materials in circulation.

(Josh (CTO and Co-founder of Kobold Metals) at 00:45:04) It's not like gasoline, where once you burn your tank of gas, you've got to refill it with more gasoline, which means you have to produce more oil from the ground, and that's very much not renewable. And so the materials in the battery, the materials in the electric vehicle aren't consumed by driving the vehicle the way that gasoline is. And so if you can fill up your EV with electricity from renewable sources, then you actually can have a low or no carbon transportation system. Yes, there's a cost associated with producing them in the first place, but you can keep them in circulation with much more modest impact associated with recycling afterward.

(Joel Beasley at 00:45:51) Yeah. My intent and thought of it was positive. I had it locked into my mind. I was like, well, even if it's not as efficient, because they're doing it and because they're putting money into it and there's a business model behind it, it will, over time, become more efficient. And so there's definitely—it's definitely the positive trend that we want.

(Joel Beasley at 00:46:11) Right? Because I like to question everything. I even question the things I like to believe the most because those are often things that need to be questioned the most. But, okay. And then last question was, what powers power stations?

(Joel Beasley at 00:46:27) So, like, let's say I have a Tesla and I'm powering it from my electricity. Who's generating that power? How is that being generated?

(Josh (CTO and Co-founder of Kobold Metals) at 00:46:33) It varies enormously based on where you are. So my dad lives in Oregon, and 80% of his power comes from hydroelectricity. So it's very low carbon. In other parts of the world, a lot of the electricity might still come from coal. And then your electric vehicle's not doing anywhere near as much good if the electricity that you're using to fill up your battery is coming from very high carbon sources.

(Josh (CTO and Co-founder of Kobold Metals) at 00:47:04) So transitioning to electric vehicles isn't enough. We also have to transition the electricity system to low carbon power sources as well. Much more wind, much more solar. And that's critically important to the whole thing, but it's possible. Producing liquid fuels from low carbon sources is a really, really hard problem.

(Josh (CTO and Co-founder of Kobold Metals) at 00:47:27) Replacing gasoline with biofuels or what have you, not possible to do at the scale of the consumption that we have for light duty vehicles. But it is possible to transition the electricity system to very high or entirely low carbon sources. Then we get all the benefits of moving our light duty vehicle fleet over to EVs. We have to do both of those things in parallel. Just getting EVs is not enough.

(Josh (CTO and Co-founder of Kobold Metals) at 00:47:56) We've got to change the electricity system too.

(Joel Beasley at 00:47:58) I love it. And I promise last question. While I have you, you're one of the smartest people I get to talk to, so I want to bounce all my ideas so that I am being a responsible citizen and repeating better quality stuff, right? Because I say stuff all day. So the last one is about nuclear. I've had this, I'm an armchair researcher type person. I'll go on these binges where I'll just try to understand something, whether it's quantum computing or nuclear power.

(Joel Beasley at 00:48:22) When I was exploring nuclear power last year, it seems to me like a really good idea, but people are just scared of it because it's publicly not that popular because of some incidents that happened when we didn't have great technology. Is that wrong or right?

(Josh (CTO and Co-founder of Kobold Metals) at 00:48:38) There's no simple answer to that question. Nuclear power has a lot of benefits because you can produce electricity with very low carbon impact and you can run it. It is dispatchable in the language of the electricity industry. You can run it when you want to run it. You don't have to depend on the sunlight or the wind.

(Josh (CTO and Co-founder of Kobold Metals) at 00:48:59) The reason we haven't built new nuclear plants in this country is not because people are afraid of it. It's because it's not cost competitive. The most recent nuclear builds were extremely expensive. And we, by not building nuclear plants, we've lost the ability to build nuclear plants, and so they will be even more expensive. So building safe nuclear power is, at least to first order, not especially cost effective.

(Josh (CTO and Co-founder of Kobold Metals) at 00:49:29) We don't have to use nuclear power in order to have a low carbon energy system. And so I think it would be too hasty to rule it out of hand, but economic considerations have been the overriding factor. There are new nuclear technologies that are potentially a lot safer. And they're designed around some of the kinds of problems that occur in conventional boiling or pressurized water reactors. But they are new technologies.

(Josh (CTO and Co-founder of Kobold Metals) at 00:50:05) They're not yet deployed at scale. And there may be a place for those as well. And the holy grail, of course, still is nuclear fusion. And there are companies working towards nuclear fusion. And that would be the at scale and very dispatchable and plentiful low carbon energy source.

(Josh (CTO and Co-founder of Kobold Metals) at 00:50:27) But there's a lot of technical problems still to solve. And so I don't think we should bank on that, but we should continue investing there too.

(Joel Beasley at 00:50:37) That'd be exciting.

(Josh (CTO and Co-founder of Kobold Metals) at 00:50:38) That'd be exciting.

(Joel Beasley at 00:50:38) I saw some articles that some universities and The UK were resuming some scientific studies on nuclear, what you just described as well. And so I was like, yay, the future. I just had heard once from some guy in the military that I was talking to. He said, yeah, we've got these nuclear subs and they can go like a hundred and thirty years without refueling, like an aircraft carrier.

(Joel Beasley at 00:51:00) I was like, what? That sounds incredibly awesome, right? Put one of those in my car, just drive around for a hundred years and never have to refuel.

(Josh (CTO and Co-founder of Kobold Metals) at 00:51:08) Yes. You would be carrying nuclear waste around with you, and you definitely would not want to crash your car and subject yourself to a meltdown.

(Joel Beasley at 00:51:21) My wife would be upset.

(Josh (CTO and Co-founder of Kobold Metals) at 00:51:23) Yeah. Yes. So would your neighbors.

(Joel Beasley at 00:51:35) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn or send me an email [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.