Oliver Gunasekara, Impossible Metals: CEO of Impossible Metals

Welcome to another co wisdom pod. We have Oliver Guna with us today. He's a CEO at Impossible Methals.

Uh he's a Forex founder to XCO, an investor and advisor mostly in deep tech and climate tech. He's a also an early employee at ARM uh which now has a 150 billion market cap. He's on to interesting stuff and I think um uh his company was at YC batch winter 2022 and yeah just an interesting guy alto together.

So Oliver, I'm glad to have you on. Tell us a bit more about yourself and what you've been up to as of late. Yeah, I'm really excited to to be here.

Um my most recent company that I started about four years ago is focused on deep sea mining with autonomous underwater robots. Um and so we've been making good progress on that. And we just recently had a lot of uh US government support.

There was a President Trump executive order on deepc mining. And uh this week we heard that our request to start leasing around American Samoa has been accepted. So yeah, big news.

Um before we get to that, what's your background? Uh what brought you to today? Yeah, I I spent about 30 years in semiconductors.

So, early employee at ARM um which went on to be a really big company. Uh but this this company is really focused on robotics. But, you know, at school I did electrical and electronic engineering, but I've really done more business and marketing in in the last couple of decades.

Interesting. And yeah, why deep tech? Why climate tech specifically?

Well, I think you know atoms are what really move the world and you know I'm personally much more interested with physical things uh than purely software. Now that being said robots have a lot of software in them but they physically interact with the world and and I think AI um in the form of robotics is the next wave. Uh so I think lots of opportunities.

Uh this particular company I think is is really about trying to uh accelerate the energy transition. We we desperately need lots of metal and the way that we get metal today isn't particularly good. We we do mines in Africa and India and Indonesia and there's lots of environmental and social issues.

So I think we can do things better. Right. and impossible methyl.

Where did you get that idea specifically? Uh I came up with the idea when I first discovered um these potatoiz rocks that lie on the seabed floor. So this is actually the product that we're after.

So our mining robot goes up to four miles deep to pick these up. And what other companies are doing is dredging. They're using these big dredging machines to effectively vacuum up the seabed with a riser system.

And this technology was uh tested in the 1970s and there really hasn't been much innovation. And so we took a blank sheet of paper and said could we do it better? Less environmental impact, less cost, no single points of failure.

And so that's how the idea came about. And what's so valuable about these rocks? Uh, and what's so impossible about them?

Well, what's what's valuable is that they're the planet's biggest source of critical battery metals. So, nickel and cobalt. Uh, there's also copper and manganese.

And there's huge quantities and we actually need massive amounts because we're trying to move away from fossil fuels to electrification. And electrification needs metal. It needs metal in the storage, in the transmission, in the generation and in the usage.

So the world needs massive quantities and right now we tend to go to places like Africa and Indonesia and dig up the earth and destroy the habitat to get at these metals. I think there's a better way of of going to the ocean. And why did we call it impossible?

Well, nobody believes what we were attempting to do could be done. Um it's technically very challenging. It needs a lot of new technology and so the name really works well with recruiting the best talent.

You know, the best engineers want to come and build something that others say cannot be done. And everything is impossible until someone finds a way, right? Yeah.

I have a hard time wrapping my mind around the economics of that. What's the value of that potatoiz rock? you know, if and if you need to build a a robot that may cost millions and uh have it go for miles deep, it can break.

Um and and returning these things like is there scale? Did you think there's like uh economic scalability into that? Yeah, we've discovered about 20 trillion dollars worth of minerals on the ocean, 4 trillion.

Um and the world last year consumed 400 million. So that's the total addressable market for nickel, cobalt, copper, and manganese. It's growing at over 7 and a half% every year.

So it's a huge market and uh we when we do our economics, we believe we'll be about 10x less expensive than the average mine on land. And that's because the grade of the resource is super high. We get four metals in one.

And we don't need to build land infrastructure on land. You need to build roads. you need to build maybe a railway line, uh, power connections, accommodation for workers.

The ocean, we just reuse ships and ports. So, 10x less expensive, 10x faster, 10x bigger resource, 10x less environmental impact. Basically, we think we are the SpaceX for mining, right?

And what's the cost of building that robot? What's the cost of the potato size uh rock that you'd carry? Yeah, the robot will cost about $2 million.

Um, but we will be able to advertise that over say 25 years. Um, each of these potato rocks could be worth about 50 cents once the metal is extracted and processed. And there are trillions of them on the seabed floor.

So at scale, each mine location could do about 4 billion a year in revenue and about 1 billion of profit. And how much potato size can one robot carry and bring up to the surface? So the robot we're designing and building now for production, we call it Eureka free.

Uh its payload is 4 metric tons. So 4,000 kilograms and it can deliver that every three and a half hours. So we reuse the same robot.

And so how much potatoes there? Uh I would have to do the maths because they they vary in size because uh let's do them by weight. Let's say they average about 100 grams.

So, you know, 0.1 kilogram versus 4,000. So, we could we could do the rough maths of of what that is. Quite a lot.

Yeah. Let's call it a round number, I don't know, 2,000 of these potatoes. And you said 50 cents, right?

So, a,000 bucks every payloads every three hours. Would that be the math? Yeah, three and a half hours, something like that.

And we we can operate 247. Um, and you know that obviously we have to get the rocks to land. We have to get the the minerals processed.

Okay. So, yeah, I I thought you would just bring them to a third party. So, you want to do the full vertical integration here.

We will use a third party. So we don't want to be building mineral processing facilities, but we want to operate with what's called a tolling relationship where we deliver the raw nodules, the raw ore, and they process them and deliver us the pure metal and we pay a tonage fee for that, right? Um, how do these rocks end up on the seabeds?

Um, yeah, it's an interesting question. Um, so first of all, the trace metals that we're after, these come up through volcanic action. uh to the surface and then through weathering over millions of years they get into solution and then the water runs into the ocean.

So now you have uh in parts per million the nickel, the cobalt, the copper and the maganese in the ocean. Now how these nodules form is that there is a kernel like a shark's tooth or a piece of clay and the metals start to solidify around that kernel and over millions of years they grow. So if you were to cut this uh nodule open, you would kind of see rings inside it as it's grown.

Typically each centimeter in diameter is about 1 million years. And so they literally grow and they form in in the trillions. So, if you were to look at the seabed floor, um it looks like a dry a gulf driving range where you just see huge numbers of these round potato sized rocks lying on the seabed floor, right?

Um but the seabed floor is not even. It's uh it's littered with other things, I I do believe, and maybe even corals, right? Like I saw your video.

Maybe I can show people just so that they can understand real quick. here is the sea seabed floor really like that like just pure sand and there's these cobbles. Uh in my mind it would be like way more uneven.

Uh it's to do with where these particular rocks form. So they they form on what we call the abyss abysmal uh plains. So these areas are very flat um and they're very deep.

They're typically uh around 3 to four miles deep. Um and so there's no sunlight. Um most of the life at this depth is bacterial life, microscopic life.

Uh but there is occasionally corals and sponges, but they're very rare. So you you see them um not very often, maybe every few hundred nodules. And our vehicle is designed to detect that with its AI.

And if we see it, we put a quarantine area around it and and don't disturb it. Is there like a Google Maps for the seafloor? Uh Google Maps has it, but it's not particularly great.

Um if you um if you want to see actual imagery, if you go to our YouTube page on um on YouTube, there's a video there that actually has real imagery data. Um, so yeah. Um, which one is it?

Keep keep going. I don't see it here. Keep scrolling all all the videos.

Uh, yeah, that one. Uh, the the one. Yeah.

Yes. The one that's that's animating now. That That's real image data from the seabed floor.

So that's exactly what it looks like. So why is it so clean? Why there's why is there like no corals or algaee's there?

Why is there no one there? Because it's there's so little energy, right? You you're talking in an environment that is so deep, so much pressure.

Yeah. You you you will see in this video a bit later on that there is a small coral that pops up, but they're very very rare. As I said, 80% of the life by by mass is bacteria.

It's not very common. And tell me about the technical challenge of building this. Like something that can go that deep and then a camera that can scan the floor, a robotic arm that can pick it up.

It seems like a a moonshot. It's a lot of technology, but fortunately our um chief technology officer and co-founder Jason Gillingham, you know, his previous company was building underwater robots. So he has a lot of expertise but fundamentally where possible we try to reuse offthe-shelf uh robotics components.

So the biggest market today for underwater robots is navies. You know the military that you know like all militaries they're moving to drones they're moving to autonomy. So that is a big market and we can buy components like sensors and propellers and batteries.

Um and so where possible we just buy off the shelf of components and we integrate them. Now we still have to integrate them electrically, mechanically we have to write software for them. So there's still work to do but there were three areas where nothing existed.

So we had to basically invent our own technology. So the first area is what we call the buoyancy engine. So this is the technology that allows the vehicle to move up and down the water column.

It's principle is pretty simple. Works a bit like a submarine. So we have a variable uh mass fixed volume.

So we have a sphere that we can pump water into and out. And the vehicle is designed to be neutally buoyant on the surface. uh and then we can with the spheres being empty and then we pump water into them and the vehicle's heavier and it dives.

So that's one tech. The second tech is the robotic arms. We needed something that could work really quickly.

So we took a delta style arm where the actual motors are off the limbs and put that onto the vehicle. Um the third area is the AI algorithms so that we can recognize the seabed, the modules and anything else. And so that's running on an Nvidia GPU.

All of that runs locally. We have stereo cameras. Uh and then the fourth area is our automated launch and recovery system.

We have this concept of a smart system where we can actually dock the vehicle below the wave affected area. And this then allows us to recover the vehicle onto the ship really quickly. So that's our our key inventions.

Rocket science. Um, so I see that you tested it in 25 m depth. Is it there way to test like in a lab, I don't know, like pressure, pressurized water or something so that you can really mimic four miles deep which springs a litany of uh issues and complexities.

Uh there is um but this equipment is super expensive. So it's called a a high. So why not build a startup uh about that?

Yeah. So, so we we take uh our components that we build uh to an external test chamber where they have a tank that can simulate the pressure at these depths. Um, and we do that.

And when we buy components, they often come with a test certificate saying they've been tested. Uh, but you can't really put the whole vehicle in the test chambers. Maybe the military have them, but we don't have access to that.

So, we have to actually take the vehicle to the deep ocean. And we did that twice last year. So, off Florida, we took Eureka 2 and we took it to about 2,000 meters.

So, well over a mile depth underwater, fully autonomous. And yet, like I would I would uh give the miners the toothpicks, the the mining picks, right? Like I would build a startup that would provide access to that big tank.

What what do they charge you to test that in the that artificially pressurized tank? It it's it's expensive. Um but you know, you pay uh per hour that you kind of have access uh to that facility, but but it's fairly specialized, you know, not there aren't that many.

Again, you know, military is the biggest use people selling components that go on underwater vehicles for for the military. And deep tech is intellectually challenging. I guess that's rewarding, but it's also most of the time less profitable or harder.

Is it is that fair to say as a niche? I would say it's harder, but I actually think it can be highly profitable. Um, the problem with pure software or or just AI on its own is that there are very little barriers to entry, which means competitors can duplicate what you do really rapidly.

And so that drives down your gross margins. um because you know the substitution is really easy and simple. The thing with hard techch is that it is hard, it takes time, but it's very very difficult to be duplicated.

Um and so once you succeed, you can actually have really high gross margins, right? Something that an AGI couldn't crack in the the next couple of years, especially the physical components, right? Yeah, exactly.

I mean, I think uh AI with robotics is going to be a huge market. Um, and that's what our vehicle does. Um, but I I I personally think it's really hard to make money these days with say a SAS business because uh AI is disrupting that and the barriers to entry are so low that it's very easy for competition which means you're not going to have a high price.

You're going to have a very low price. And why not build the SpaceX of deep tech just like bringing payloads in in and out and and being again like a a provider to all these uh deep uh deep sea companies. Uh have you somewhat thought of that or another way of monetizing what you're building right now?

Yeah, I mean we're interested in the components that we've built could be repackaged for for military or other applications. Um, and of course we we're interested in that, but few markets have a $400 billion TAM last year and and growing. That's that's a pretty unusual market and it's a market where there's no risk because unlike most startups, you have the fundamental concerns of uh do people want the product or service and secondly, can I make money at the price I need to sell?

Well, when you're selling a commodity like nickel or cobalt or copper, we have none of those issues. You know, the market is huge. It's growing.

It's all about controlling our cost. We don't need to worry about whether there anyone wants this. Um and so that is super attractive and I would describe them as like um our agricultural lands, right?

Like these these flat lands with these bunch of cobbles. uh how many of these are there uh in the world and and what's the the total kilometer surface? I'm asking that because I don't know maybe there's like 50 100 of these spots and every time you need to reach those you need to take a boat and there's logistics involved in that.

So how many of these playing fields are there in the world? So mining is is quite highly regulated. So it's not like you can just go to a random spot and and start.

So we we have three phases. We have prospecting which is basically where you kind of want to know where stuff is approximately. This really started in the 1970s.

Uh and then you have expiration. Now expiration is where you get an area exclusively for you to do the environmental work. And these areas are large.

They're the size of Portugal, 75,000 square kilometers. So huge. And you have that exclusively.

And with the expiration, what you're trying to do is concisely define the resource, what's down there, what's the cost, and also what is the environmental baseline and what would be the impact of doing the mining. And this started in in early 2000s. And then you have the final phase, which is what we're focused on, which is called exploitation or actual mining.

And this is where you take one of those areas, you take all the data that you've collected, including the environmental impact assessment, and then you submit it to the regulator and say, I want to mine this area for 30 years. This is the space. And because these areas are so large, even in 30 years, you're only going to probably pick up 10 15% of the material that's in that one location.

Interesting. How is YC for you? And uh yeah, I believe you raised a seed from there like are you still raising this year?

Yeah. Um YC was amazing. Um we went into YC having done a preede of about 2 million.

Uh but after demo day we raised another 10 plus um on the back as a as a seed round. And um I think the the network was very very powerful. The advice was great and it it made fundraising really easy.

So we we concluded YC winter of 22. So, it's it's been about three years since we raised our seed round. Um, and so we're we're starting to think about our series A, uh, which would be to finish the technology.

Um, and so, yeah, we're we're basically in the market now for for the series A. Did you mention you just unlocked the contract in America and Samoa? So, how does uh that play?

Yeah. So, we we formally requested to the the US Department of the Interior uh to start the leasing process around American SMOA and they agreed to um which is super exciting. So, just um just this Tuesday they put out a press release uh from the the US Department of the Interior and they actually mentioned that it was our request to start this process.

So, this is this is super exciting. Right. And uh where is it exactly?

So American Samoa is um a long way south. So if you uh I think you've got it there, but if you zoom out, it's if you Hawaii is right north. So it's basically a long way south from Hawaii in the Pacific Ocean.

So I think we can see Hawaii up there. And yeah, Hawaii's here. Samoa is is down there.

I was zooming on it, right? Like American Samoa. Yep.

Yep. Yeah. And where exactly are you going to uh see these things?

Like if you see on my screen here. Yeah. So there would be in in the within the 200 nautical miles to kind of the right corner.

Um you know so this is this is called the exclusive economic zone. So 200 miles around a country belongs to that country. Um and that's the area.

And this country borders what's called the Cook Islands, which is a country to to the east. And they have also been extensively um working on deep sea mining for about the last 20 years. Interesting.

Uh but yeah, how quick do you think this is going to fully unlock because you're still in development mode. It seems like a very long-term uh business, long sales cycles also and so forth. So uh what's your ETA on I don't know the first million bead mine in the next five years?

Yeah. Uh 2028. So just just under three years is when we think we can start production.

Um that would involve us finishing the technology and then building a fleet. But it also involves the partner that we would have uh applying for the the permission using the environmental data and getting granted. So we're we're trying to intercept those two together.

Very interesting. And how do you keep your sanity uh with such a long-term business and your patience? Yeah, I mean it you you have to focus on the mission like you know if we do this well um we'll solve some massive technical problems which is stimulating.

Um we'll allow the world to electrify which I think is good. We'll avoid some of the environmental and social damage that happens in mining on land. Um and we'll build a massive business.

So that's, you know, you you keep focused on the longer mission. Got it. Well, thanks Oliver for being on.

Where can people find out more about you? So I think primarily our website is the best place. Um, impossible medals.com.

Uh, we're also pretty active on on LinkedIn as well. So that was Oliver and Charles for another founder