DUG Technology Ltd (DUG) Earnings Call Transcript & Summary

Thank you for joining us, everybody, today. We'll just let the participants come in. We're delighted to have you on the line to hear both Matt Lamont and Louise Bower present on DUG Technology and the infrastructure in place and the growth strategy that's planned for the business. I think we have enough people now joining, Matt, but I'll throw across to you to take us through the presentation. We will have Q&A afterwards. But please type the questions as we go through, and then we'll moderate the Q&A at the end of the presentation. Thanks, Matt. Over to you.

Thanks, Ronn. It's nice to be with you all today, even if I can't see. So I think a lot of you will be familiar with the story today. So nonetheless, I'll recap who we are, where we're coming from to -- and -- but I'll go through it fairly quickly. And then we've made some -- a few new slides that hopefully help to illuminate what we are and what we do. So what we are is fundamentally mathematicians, physicists, computer programs that work --programmers that work around numerical algorithms and machine learning and artificial intelligence and running these big scientific programs on very large computers around the world. And we build these computers ourselves as we need and the computer room technology is ours as well, which we patented. We have a portfolio of, I don't know, 15, 16 patents, a couple of which are around the computer room technology and the rest, which are around mathematical and physics algorithms. We're global. We have 4 global offices, and we'll look at that and about 29 petaflops of compute. To give you a feeling for what 29 petaflops of compute means the national compute facility in Canberra is about 12 petaflops. The national compute facility in Perth is about 1 petaflop. So it's a very large machine, although it's several machines and we're about 300 staff and we've got a lot of clients, well over 200 customers, including your big blue chips. As we are traditionally in the oil industry, we're now breaking -- we're working beyond the oil industry as well. So Shell, Chevron, ExxonMobil, for example, Equinor, Total, are all our clients. Through to all your national, PETRONAS, Petrobras and so forth, we do work for all of these guys in 48 different countries. Our software itself is in 36 different countries. So that's the world according to DUG. The black dotted line is our global network that we have, fiber network which allows us to shift data and clients around the world. So data and staff to work on different problems around the world. The orange countries are where we've done projects. So you can see we've worked in most everywhere around the world. Bubba is our big boy at the moment. He sits in Houston. So he's 21 petaflops, big computer room with a lot of spare capacity. Bazza sits in the middle of London. So he's quite small, but all of our London guys, great business in London, great scientists in London, they work on Bubba. And then Bodhi sits in KL. And in KL, we've also got a lot of our rear-end business like our HR, for example, and a lot of support people for the business around the world, and then Bruce sits in Perth. But of course, Bruce's big sister, who's going to be built in Geraldton is going to supersede all of these machines and be a lot bigger and be completely green, which is a huge selling point in the world today. Recent announcements. So we've just settled a long-term patent litigation in PGS Australia. We did the trading update. But the big one is the Geraldton release, and that's really resonating around the world and it would be the first carbon-free high-performance computing data center. And it's really interesting, and we're going to talk about that and just how compelling it is to build this center in Geraldton. DUG Cool, so that's our computer room patents was just granted. We already had Australia, Malaysia, Singapore granted. We've now got U.S. granted and we're waiting for Europe to be granted for that. CSIRO was signing on, was a big deal. So that we're starting to demonstrate that there is a HPC business, and we'll talk about why I say it like that in a minute. And then there's this big alliance with Curtain that was also announced. So software-wise, software is a really important part of our business. It's something we often don't emphasize enough. We are really good at software as particularly scientific software, we have a very large -- a very great big data software scientific suite. And this traditionally is the seismic processing in the oil industry allows us to take a very large data set so-called Big Data and store it, slice it and dice it, so random access it. So you're ready to do machine learning and other things, you really want to random access your data so you can let the algorithms meander through the data sets. Well, that takes a lot of engineering that we've worked on for a long, long time. We can then process it over very large machines and then you can visualize it in 3D and 2D in all sorts of ways. And one of the big things that's happened to this software recent times is it's now being used by a radio astronomy. So you've not just got the coordinate systems to allow you look down into the earth from anywhere on earth, all the different coordinate systems around the globe. But you can now look up. And so you've got a galactic coordinate systems. So we've got people looking at stars in this package now. And what we discovered much to our delight, when we sat down with the radio astronomist, 95% of our algorithms, we've got about 250 algorithms in this package. And 95% of those algorithms are common between here and radio astronomy. Signal processing, data processing is data processing is data processing. So now we're working hard to broaden this software. So it's not just for the oil industry, it's but the radio astronomists, it's for the meteorologists and for other scientific endeavors as well. A big part of what differentiates us from other companies and other cloud providers is our knowledge around -- our deep knowledge around running numerical algorithms and writing algorithms. If you think about it, scientists, a molecular chemist is a molecular chemics, he isn't an IT guy able to write deep algorithms. But that is indeed what he needs. Most scientific endeavors today requires high-performance computing. And so how is he going to get his algorithm working really well on high-end, high-performance computing facilities? He needs help, right? Because it's a science unto itself and that's the skill set that we have. So we can help all the users come on board and get running on the high-performance computing, and that's a really unique proposition. And it's something that everybody wants. But when we talk to the military ,they say, we'll take that and we say, well, only if you're going to run on our machines. So we've got a couple of quotes for you. So, "All areas of contemporary research require access to advanced computing capabilities, big data, big science" and that's from Professor Andrew Rohl. And that's what we see. Everything you can think of from genome work, to molecular chemistry, to COVID research, to Bushfire research, anything military, anything military, anything spy, anything you could think of requires high performance computing. And that's what this next one says, radio astronomy and high-performance computing are absolutely wedded together. So the HPC outlook globally is huge. It's currently predicted to be about $151 billion. But as I said, it's a purview at the moment of governments mostly, and we're in this start of this wave of it becoming out of government into commercial reality. And so it is everywhere, and we've talked about this. But it's going to be a $1 trillion business well before the end of the decade. The unique thing about the high-performance computing world is that it's mostly around scientists, right? And they are vast majority of them. And it doesn't matter if they're in the oil industry or they're in other industries, are extremely concerned about the planet and carbon warming. And so much so that radio astronomists just published the paper, not on astronomy, but on their carbon footprint and what are they going to do about their carbon footprint. Coming out with green HPC like we are and our ability to do it, right? Our DUG Cool solution that we already have in place saves 51%. So we are already green is a massive sales. It's very important to the planet. It's very important for human beings but it's extremely important to this business. It's a really monster selling proposition. Our computers are also merged. We invented that -- the way we do it is what we invented, we've got patents on that and that saves us 51% of the power bill. So before we go and we marry this with the renewables and our hydrogen solution, we are already green. This is a green solution, we are already green. This is just -- this is what -- I'm not going to go through this, don't worry. But we're trying to think about how to get across to you what a high-performance computing system looks like, right? So in the middle here, you've got orange, you've got management servers, HPC storage, computer services and stuff. That's what you would find in another cloud vendor. Everything else that goes along with it, including the 2 security layers, including all the monitoring system on the left-hand side and all the support systems on the right-hand side and so forth. Everything else is unique to a high-performance computing system. So what happens is if you're in a company and you want to get high-performance computing from a traditional cloud vendor, then you've got to take all this knowledge and software with you and you've got to take your IT team with you and you've got to have that conversation with Google or Azure or AWS. This is what I need. How do I get it? I've got to ask for them to set it up. I've got to get -- the factory networks are not part of the way Google or Azure work. So you probably can get it, but you're going to have to ask for it, you're going to have to pay for it because it isn't a standard setup. That's part of scientific. That's about passing data between machines, not just between storage and machines. And then you're going to have to set it all up. You're going to have to set up your security layers, you're going to have to set up your monitoring, you're going to manage your users. You're going to manage the users, the way they use it, put on caps on how much they can use or what they can't use, all of that stuff is what we run and what we have. And then you get to the craft on the right-hand side, the green that we've already talked about, let's get you running, let's compile your codes better. Let's -- what's holding your code up, why isn't it running better? Let's get it running on higher end. Let's get it better threaded and so forth. That's what differentiates us from the other cloud vendors. So where does this business come from? Why don't we have people to compare to? Well, there is precedence. It is -- we are serving the new wave. As I said, the compute -- high-performance computing is traditionally provided by governments and scientific organizations that are government-owned, the National Labs in America, for example, or national compute facility on ANU in Australia. But it is changing, and there is precedent for change until we just mentioned a couple of precedents in the United Kingdom and in the United States here. So it is changing, and we are riding the wave, but that's why you don't see other companies looking just like us right at the moment. There are several companies that look like us that are actually just layered all that craft on top of an Azure platform, and that's fine, but that's where you'll find our competitors. But of course, they won't be able to compete with us because of our deep craft around the science of running and also our price point because we're providing it all ourselves. Now green is very important. We talked about that. And so is very large. We saw that we're predicting a $1 trillion business before the end of the decade. Well, that needs to start thinking and planning and building for that right now. And it's also green. Green is a massive. Because you're selling a lot into different institutions and governments now, you've now got to talk about the triple bottom line, not just commercial, it's environmental, it's social and it's commercial, the so-called triple bottom line. So in Geraldton, you've got world-class renewables. You've got amazing wind. If you just Google Geraldton, you come up with trees growing sideways. That's because of how strong the wind is up there. You've got fantastic sun. And the 2 wind and sun are quite out of phase. So you end up with only 6 hours a day where you not have -- where you -- on average, where you don't have wind and you don't have sun, and that's where the hydrogen comes into fill in those 6 hours. If you want to be completely carbon-free. We're not talking about carbon-neutral. We're talking about carbon-free. And it's only 3 milliseconds latency from Perth. So it's like it's in Perth. It's a unique -- it's a very unique spot, not just in Australia but on earth. That's why the world -- what you call those surfboards with sailboards. That's why the sailboards -- the global sailboard championships are in Geraldton every year, it's because of the wind and the strength and reliability of the wind there. But what we're forming there in the first data hall, it depends on the computing, right? And the compute we put in depends on the demand by the clients. If we put in pure CPUs, we end up at about 100 petaflops, 120 petaflops in that first data hall. If you put in GPUs, we can end up at 200 petaflops in that data hall. It depends on time, machines keep being more powerful all the time. So we've sort of said conservatively 130 in this little picture at the bottom. In other parts in here, we're saying 200 and that -- and the reason is that difference is it just depends on what the compute is we ultimately put in. But you can see the relative size of this, we're talking huge -- the ability to put in a huge large machine to cater for what's coming down the track. So why Geraldton? It's the best place on earth for renewables. It's got great fiber connections. It's got great fiber connections up to Singapore and the world, not just into Perth. It's got a -- it's a city of 40,000 people. It's a farming, agricultural, mining center, so it's got great workers and pragmatism and all the skill sets that we need to do what we want to do. It's got a TAFE, which is important to us. It's got normal construction costs. It's not remote. And there's a lot of -- the state government is very keen to develop that region and very keen to have hydrogen and renewable projects there because of the strength of renewables. The state of Australia, the state of Western Australia is going to go gas oil and over the next decade or 2, it's going to turn to a renewables energy hub. It's going to stay an energy hub -- is where we're going. So the capital model, what we're trying to do for you here is paint a picture of what's possible. This isn't the forecast. This is just showing you what's possible with what these -- with these facilities that we're building. So I'll start on the left. Our MO, and you would have heard us talking about this before, is to put the long lead time items in ahead of time. Anything that we can do just in time, we will do just in time. Anything that we could do almost just in time we'll do almost just in time but things that have to be put in long term, we put in long term, and that's the building basically. And then we put in the compute either opportunistically because somebody offers us a lot of compute for cents in the dollar or we put it in just in time. At the moment, we've got that 29 petaflops of compute. So what is this Data Hall 1. We're going to build Data Hall 1, the Board has signed off on building Data Hall 1. And as we said, Data Hall 1 could have 200 petaflops if we put in the current generation GPUs, it'll have 200 petaflops. If we put in tomorrow's generation GPUs, we'll have more than 200, we'll have 300, 400 petaflops. If we put in CPUs because that's what clients are demanding, it could be down to 120, 130 petaflops. So a medium mark, 200 petaflops. If we were to put in 200 petaflops of compute, which is what the building will take that we're building, then the revenue per annum would be $224 million at today's market rates, and the EBITDA margin would be 79%, the EBIT margin would be 62%. So -- and the capital expenditure would be $229 million. But I really, really want to emphasize with you, this is when that data hall is full, right? We put in compute just in time. So we gradually fill up that computer room as we need the capacity. It's sitting there, it's ready to go. We could get a huge order next week and we -- well, next year, and then we can put that computing in one hit, but it's unlikely. It's likely it will build gradually over time, but it will increase as we're building this business, it gets easier and we're known to more people, and it will accelerate. So that capital expenditure is spread over the entire life of that facility as we fill it up, not the life of the facility, as we're growing that facility, that capital expenditure. We don't spend that upfront. What we're spending upfront is $4.6 million. And then as we populate that machine room, we have to start building -- paying the rest of that capital because most of that capital is going into computers and the computers are putting just in time. Even the hydrogen solution we put in, in parts as we grow. We don't put it all in upfront. So that $229 million is associated with the revenue, the $229 million capital expenditure is not upfront. No revenue -- no capital or very little capital expenditure. So you can see what we're talking about is a fabulous business. And what we can't tell you is how long it's going to take us to fill up that first data hall, right? But when it's full, it's a really fantastic business. And remember, this is Data Hall 1 of 10. On that site, we can put -- we've got the plans in place for 10 data halls. That's for that site. And that's like we go beyond 10 as well. But at the moment, we're planning for 10 data halls. This is Data Hall 1. This is what Data Hall 1 could deliver to us. Anything you want to say about that one, Louise?

No, it's all good.

You got to say something, so I can drink coffee.

Just saying that we do actually have capacity to grow Houston as well. The facility that we've got over there is about 1/4 populated. And again, depending on the architecture that's put in, that could grow to about 250 petaflops.

Or beyond. It could go to 400, 500 if it went to full GPUs as well. But yes, and that's perfect MO. Put in the data rooms upfront and then populate them. And most of the money is in the populating of the room and that comes in when the revenue comes in. So there's no big capital expenditure upfront. The revenue model. So we just -- again, we're trying to just tease out and look at it. So our business is around software, high-performance computing and services. Our traditional business is services, it's low margin, right, which is why we're moving to software and -- emphasis on software and high-performance computing, which is what we're very good at, and they have much higher margins. You use software and high-performance computing. The reason we have software and high-performance computing skills and depth of skills that we do is because that's what we require to do our services. It's just that the services has been commoditized, seriously commoditized, especially over the last 5 years. We used to make 25%, 30% EBITDA margin on it. Now we don't. It's been really hammered in the last 5 years while the oil price has been down. And what you can see is, although this graph might look subtle, we are transitioning to software and high-performance computing. It is of low basis marks, it is of low points. We're still earning most of our money through services but it absolutely is occurring. And it's only over the last 18 months, 2 years that we've been building the non-services business. So it's very early days, but it's very exciting. Another thing we've been focusing on is building the sales force. It's incredibly important to get this global is sales force. At the moment, we're really focused on the non-oil and gas business in Australia, but that's just to generate a beachhead, prove up that it's a business which we've now done. We've signed on CSIRO, we signed on Harry Perkins. Now we've got to take it globally, and we need to get this sales force really strong and trained appropriately. We do have 16 salespeople globally because we're selling services globally. Now we've cross-trained those into selling HPC, and we are hiring to build up this team to take us to where we want to go. It needs to get much larger. It's going to take time. It's a continuous effort over a long time to build that team up. It is actually already a very good team. We've added regional sales leads. So we're getting structure into that global sales team. And we're changing our commission structure. It's all mapped out, and we're just rolling it out now. So make sure that sales guys are more incentivized to sell the high-margin business lines like software and HPC than they are to sell the low-margin lines like services. Equinor is a case study. They're a $60 billion company. You were seeing us talk about before, but this is a really important sign on Equinor, a known early adopter of quality and technology. So they were -- we knew they were a great target for us because we knew what we had is really core and really good. It's a really important one because we -- it was very disruptive because we displaced Schlumberger software running on Azure. So we can complete against Azure in our high-performance niche, no problems at all. And it's going really well and attests to the recurring nature of this revenue, we -- every month, we invoiced USD 238 million -- USD 238,000. Did I say million? That was a bit of a -- every month, we invoiced USD 238,000. It is growing, the number of users on the system is growing, is going super well. In terms of diversification, we signed Harry Perkins and that's a landmark signing in the genome space. It's not that material to us but it's allowing us a great base for us to then go to sell to other genome and bio-informatics by businesses. And SKA is the big one, and it was -- when we started looking, which is only about 12 months, 18 months ago, looking outside of the oil industry, we identified 2 big elephants we should go after, one was the SKA and the second was meteorology. We went after SKA, we said, "Look, come and run on our machines. We got a post doc. We took our code, we ran it. We looked at it run. We said, well, that's really awful, what it's doing." We worked on it for 2 weeks. We made the code go 125x faster. She then ran the entire 5-year backlog in data on our machine in 3 hours of the 1/4 of the machine in Perth. She'd been trying for 2 years to run that data set on the national facility and only got 15% through. So in 2 years, she got 15% through on the machine. We allowed it to get it all done in 3 days -- 3 hours on a 1/4 of our machine. That's that onboarding skill set, that's that deep knowledge about numerical algorithms that people can't get anywhere else. We've got a big effort in social media, and we've got a really great blog and I would encourage you all to follow the blog because it's really active and there's some great stories and it goes to the heart of who we are and what we're doing. Thank you very much. You've seen the financials before, they're the ones from 6 months -- from the previous 6 months. Thanks, Ronn.

[Operator Instructions] We have a question already. And it's -- the question is as follows: why put the new facility in Geraldton? Would it not be easy to fill capacity if you're located closer to a major non-oil and gas city? And how quickly can you transition from oil and gas? It's a triple barrel question.

So the first question, we covered off. Geraldton is the renewables hub. HPC is a huge user of electricity and it's a strong desire of everybody globally involved to have green energy and green HPC. You can go out there and you can find this very simply. If you look at Google, you'll see Google -- Azure, has just announced immersion cooling. So they got -- they've created their own immersion cooling like ours. It's different. I'm sure they're not infringing our patent. If they do, we'll go and talk to them. The computers in America use more power than the U.K. does in total. Power is a big deal. It matters to this business. It matters to the world. It matters to the governments of the world so being green is a huge sales proposition, right? That's number one. I can't emphasize that enough. You go out on social media, you'll see our announcement over green HBC is resonating throughout the world. We get any questions from Oracle cloud compute people, we're getting Google making comments on it. It is resonating. This is a very big deal. Secondly, it doesn't matter where you're computing it. There's no -- high-performance computing of its very nature is done by a queuing system, right? Whether your -- whether if the compute was sitting in the office next door to me here or whether it's in Geraldton, that it makes no difference. The user couldn't tell. The user doesn't know where it is and doesn't care. Our guys, we've got 75 -- or actually it's less than that. We've got 75 people in total. We have 60 scientists sitting in London. They all work on the facility in Houston. It doesn't matter, right? So what matters is how you power it, how green is it, how cheap is it, how much -- how are we able to add capacity quickly. That's what matters. Where it sits doesn't matter. We've got Imperial College in London running what our machines in Perth, right? It doesn't matter where it is. And we're doing things to make that more and more so about how you transfer data. We've got programs. But really, if you're sitting in a nearer capital city or sitting in Geraldton, doesn't matter. You still got to transfer your day, you still have the same protocol as it makes no different. What was the third part of the barrel?

And the third part was how quickly can DUG transition from oil and gas?

Well, there's the billion-dollar question. It's not a million-dollar question, that's a billion-dollar question. And we're working very hard to transition. And well, the first thing I would say is not actually a transition. We're not leaving services. We're not leaving the oil industry. We're just broadening way beyond it. And yes, the way beyond it will eclipse our oil industry business in time. But for the time being, there's still a lot of growth that can be done in the oil industry as well, and we're working very hard to bring those margins back up. And the oil price comes back up as it has and the margins will come back up as well. So that's the first thing. We're not transitioning, but we are broadening. And it's literally a billion-dollar question. And it's going pretty quick. There's a -- if you talk to Google, if you were sitting poolside in Singapore with a Google executive as one of our stock brokers was, the Google executive said, this is a fantastic business. The only thing annoying about it's got -- it's enterprise has a 2-year sales cycle. While we're sitting well within that 2-year sales cycle, we've signed on several really, really fantastic clients, and it's a case of land and expand, and we're really well into that cycle.

So staying on Geraldton for a little bit. And you mentioned just before, Matt, that Imperial College in London is using the facility in Perth, and that's not impacting their ability to utilize the facility. So there's a question that's come in around, are there any overland communications limitations Geraldton to Perth by having the facility in Geraldton?

No. There's 3 monster fiber packets. There's 3 monster fiber packets running from Perth to Geraldton. There's also a fiber running -- or there's a spur of a fiber running from Geraldton from -- there's one of the fibers that run from Perth to Singapore as a spur into Geraldton as well. So you've got great global connectivity and you've got huge connectivity. So for example, just to give you a flavor of it, we can have a 100 gigabits, very low latency, 3 milliseconds latency, 100 gigabits, which is a monster pipe for $10,000 a month from Perth to Geraldton. There's 3 providers of that, all of which are very keen to sign us on. Don't forget the SKA, the antenna for the SKA, which is a multibillion euro global project, the antenna is out of Geraldton, 360 kilometers northeast of Geraldton and all the data from that antenna goes straight into Geraldton before going to Perth.

Any security issues in Geraldton? Is another question.

No security issues in Geraldton. We are working on getting -- this will be a classified facility. We are working very hard on making it a classified facility. I think that was in one of our releases. There's a lot of work to do, to do that, and we're on that track. So it actually -- it meant we moved. We were going to put it closer to the airport in Geraldton, but there is -- it is more difficult to be really secure when sitting right next to an airport, funny enough. So we moved a little bit further away.

Reflecting on the green side of things with Geraldton, can you elaborate, please, on the hydrogen component of the facility?

So the hydrogen component of the facility is just a battery. The way hydrogen can be thought of in multiple different ways, right? It can be a mobile fuel source if you're using it in a car. In my opinion, it's the best mobile fuel source, and it will be what is certainly the best for Australia. So you'll see it as a mobile fuel source. You will see it as an export commodity, using renewables. There's a lot of projects that are going to create hydrogen and ship it to Singapore, for example. But in our case, it's just a battery. And the reason why -- it's a really nice battery. And the reason why listeners are probably thinking, well, why don't you just use lithium-ion, that's all proven. Well, I'll tell you why. So we -- just imagine this for a minute. You have a 1-megawatt demand for power from our data center. We'll have more than that, but just keep it in the units 1 megawatt. But I've got to fill in. I've got to fill in it 6 hours a night on average. So some nights, I might have 12 hours, some nights I might not have any gap in renewables power, right? But I've on average I've got to do 6 hours. And in the middle of winter, there can be some days where you go 48, 72 hours without power. So if I wanted to go for just 24 hours with no renewables power, then I need the lithium-ion battery that does 24 megawatts, right? Even though I'm only going to pull 1 megawatt an hour out of that battery, if I want to survive for 24 hours, I need 24 megawatts. If I go to a hydrogen, then all I have to do is store enough hydrogen so that my 1-megawatt fuel cell can last for 24 hours. So suddenly, all I'm doing is storing a lot of hydrogen. My fuel cell, which is the expensive component is still only 1-megawatt. And all I have to do is have enough plastic piping to store the hydrogen to take me through 2 weeks, however, long I want to survive, and I can just trickle hydrogen into that over the whole year and then my 1-megawatt fuel cell will keep the data center running. That's why hydrogen is so powerful. And hydrogen is absolutely carbon-free. It is the greenest of the green of the greens. Lithium-ion is not. So there is an alternative, and we are looking at it as well, which is flow batteries, which is where you store the electrolyte outside of the battery. Lithium ion batteries don't allow you to do this. So then you -- it's just like hydrogen, you build up all the electrolyte, which is all charged and you can suck that back into your battery system to power. So you can have a 1-megawatt battery if you like, which can last for 24 hours or however long depending on how much electrolyte you store. We are looking at those as well, but there's a lot going for it in hydrogen. We should all be driving hydrogen cars in the future. Believe me, as soon as there is one and the infrastructure to fill them up, I'll have one. They won't change your lifestyle. We'll still be able to go driving off around the world -- around Australia and go camping and pull into a service station and just fill it up like we do now. Sorry, I've gone off a little red herring there.

All right. So while we wait for some more questions to come in, maybe a question for me. What learnings have you taken from the Houston facility in thinking about and planning for Geraldton?