Green Giants – Episode 129

Kenneth O’Mahony (01:27)
Yeah, as an accountant, I suppose you have controls and you look at the big drivers for the cost and you’re always competing to try and find efficiency. And one of the big ones we saw in our location in Ireland and other locations we had throughout Europe, Middle East and Africa was the cost of data centers running our own EMC’s own data centers. And one of the big items there was electricity.

And as you imagine, electricity costs were rising, the rate per kilowatt hour was rising, and all of a sudden the numbers were rising by themselves with the same utilization. But bigger than that was the servers. They were becoming more guzzlers. They were eating into it. So we really focused, from an accountant’s perspective, I focused on the number. I focused on what could we do to reduce the cost of electricity. And ultimately, that led me down a road associated with engineering. We had an amazing engineering team.

and we started to look at the efficiencies of our own data center in Ireland and what could be done to make it more efficient. And at that point in time, we just focused really strong on the most advanced technology. What was we had? Nothing limiting us. So it really was an investigation supported by very good management that said, off you go, try and investigate what’s out there to do what we do today more efficiently, more effectively, get that number down. And that was the drive.

Wes Ashworth (02:43)
Yeah, it’s awesome. To me it’s always interesting how the kind of an operational perspective can reveal opportunities

that people closer to the technology may overlook and is always interesting to bring in that perspective. So take us back just a moment. Take us into the room when Professor Tony Robinson took apart a server and pointed to a tiny chip inside it. What did he show you that made the industry’s whole approach feel backwards?

Kenneth O’Mahony (03:06)
So I happened to think I was very proud of what we had done before until I met Tony Robinson and at that time I didn’t know him that well. There was three people in the room, was three co-founders, Professor Tony Robinson, Dr. Cahill Wilson and myself. And Cahill and myself knew each other a bit but Tony’s eyes started to raise when I explained to him about the heating, ventilation, air conditioning and how we used to do things. And he took off the lid while listening to me very attentively and he looked at Cahill and asked, said,

Will this guy get offended and calls went no just tell him the truth and he he showed me his thumb and he showed me He said that the chip is about the size of my thumbnail I started to take the server apart and he said all the heat is coming from this tiny component Which I understood and nodded in agreement so far we were getting on well And then he looked to the sky and said so you guys went away and did heating ventilation and air conditioning changed configuration

By the end of it. I said to my still remember the words very vividly I said if you can do what you’re saying you can do which is control it right there at the chip Give us the optimum performance right there. We’re going to change the world and To be honest that was that was the start of what Nexalus was I really believed what he had and when you’re talking to the now registered the top 2 % Scientist in the world is sitting opposite you kind of sit up and listen and he has proven it ever since and along with all the rest of the people we brought on board

Wes Ashworth (04:23)
What an incredible moment, right? It’s amazing to see those light bulb moments and that epiphany happen in real time. And then the power, I think a lot of people may have a moment like that and then don’t really do much about it, but then taking it and then doing something about it to try to change the world. I just love that.

When did that shift from an exciting scientific idea into a company you were willing to just bet years of your life on?

Kenneth O’Mahony (04:46)
we were sitting in the room talking and Cottle had explained to us how he had read an article on The Economist. And The Economist had said that data centers were the future, much to his absolute horror, they were still using air cooling.

When we explained amongst ourselves the difference between the tradition and what was possible, we got together, we applied to the Irish government for a commercialisation fund, papers were drawn up and we put the proposal in and we were lucky enough to be awarded it.

Wes Ashworth (05:16)
Yeah, absolutely incredible. And as I say, the rest is history, right? It’s amazing that kind

starting moment going from there. And I think, you know, w as we understand that origin story, as it starts with a very small chip, you know, but the forces bearing down on computing are anything but small. AI is just testing how much power and heat we can pack into a space. so when people hear about AI’s energy challenge, they usually picture electricity demand. Why should they be thinking just as seriously about heat?

Kenneth O’Mahony (05:44)
it’s very interesting because almost every watt that’s used that goes into a chip is also consumed, so we know understand the kilowatts, but it also generates the heat. so the AI energy problem is a heat problem too. If you can’t remove the heat, you can’t run the chip at full performance, you can’t pack density, you can’t spend, you know, it gets restricted in what’s happening. So cooling has become the asset to enable the AI generation to deploy further.

And if we really think about it, we’re all very familiar about Glen Dimplex So we understand we plug in our Glen Dimplex heater and it automatically converts it into heat. And we understand that we’re buying a heater and we’re getting an output. But realistically, when we buy our computer, our computer is deployed for a reason. But really strangely, it converts electricity into heat the same efficiency as your Glen Dimplex heater. And the industry, what has it done to date? It has seen it as a problem.

It is seen as something that needs to be resolved, whereas the world needs to understand that there is an opportunity here for this asset to be just as good as the Glenn Dimplex heater is at doing its job. The computer is the same thing. It’s a heater, and it’s an asset there that needs to be captured.

Wes Ashworth (06:54)
Yeah, I love that. I love when people see problems as opportunities and just seeing that from a different perspective. And as we look into It’s like power demand and heat management are just deeply connected and cooling isn’t just an expense in the background. I can show you how much computing a facility is able to deliver. so talk me through this. So for a listener out there who’s not a thermal engineer, walk us through what happens in a Nexalus system from the moment a GPU gets hot to the moment the heat leaves the server.

Kenneth O’Mahony (07:22)
And it’s a simple process you’re dead right and luckily me not being an engineer. It’s a good way to describe it So when a GPU heats up the liquid is pumped right onto the chip There are tiny jets that hit the spots directly the liquid picks up the heat from the chip Flows out of the seal to the seal server and goes to a heat exchanger From there the heat is either goes to somebody for use or it’s rejected into the environment now a normal

Cold plate will do that as well, but a normal thick cold plate flows water across the chip as you imagine Just like a power hose is flowed onto it versus the shower head It’s turned the nozzle and it touches at the pin points So our jet impingement pinpoint accuracy allows us to pull the heat off the chip in ultra efficiency and also a bit of power and not pressure which we can go into the power that pushes at the jet right to the point pushes the surface water out of the way which causes a friction and

and then it gives it the most ultimate efficiency. So we gather the heat ultra-efficiently and we make the chip perform better and the whole system to act better than it can if it’s just traditionally cooled.

Wes Ashworth (08:31)
Absolutely. Great clear clarification there. And I appreciate you spelling that out. And as you mentioned there too, so your team directs these tiny liquid jets at specific hotspots on a chip. why is that different from simply putting a conventional, you know, liquid cooled plate on it?

Kenneth O’Mahony (08:44)
Yeah, it really is the differentiating factor because it not only does one thing, it allows another thing as well. So it takes the heat away from the chip in a very efficient way. And ultimately, most other systems push high pressure through the system so it wants to flow the water at a fast pace. We get to the point of the chip and we just have velocity. And that velocity moves it quickly to the point without the pressure. When you reduce pressure like anything else in life, our hearts, everything, lower pressure.

better efficiency, more longevity. So ours allows us to reduce the pressure in the entire system, which everyone really likes, because you just don’t want that pressure throughout your entire system. But also it gives us the efficiency, the efficiency of grabbing the heat, taking it away, and also at a high temperature. You want what you recover, if you can recover it at all, to be at a usable high temperature.

Wes Ashworth (09:35)
That key there and just really maximizing the efficiency, obviously, and knowing how important that is and the impacts there. And that precision obviously matters when the very small area is carrying a disproportionate share of the thermal load. And you know, it reminds us too that liquid cooling describes a category, not one interchangeable technical solution. There’s a lot that goes into there. but you also talk about sealing the server so it no longer relies on air moving around the room to stay cool. what does that let a data center operator design differently?

Kenneth O’Mahony (10:03)
there are two patents, by the way, that we did when we left Trinity University, was the jet impingement and the sealed server. So the origin of the start of the company, the sealed server is really the one that is easy to understand and also has the biggest impact because it’s hard to say it, but it really is exciting. It literally changes everything because no one seals a server. The tradition has always been you leave the server as open as you possibly can.

you push air across it even if you do liquid cooling you touch down on spots and you leave the rest of it open so the efficiency is you take off as much as the metal as you can and you blow air across it and ultimately if you think about that in the structure and setup of a data center you might have a server where over on aisle 14 that is really really hot you got to get the air there so you push it across everything we do it exactly the other way we do we do a demand based so we seal the server

And you should see the eyes of some of the people that we talk to, and their origin of the company, when we start showing them the complete sealing of the server. When we seal it, we not only are able to optimize the jet impingement on the GPU or the CPU or the touch points like the NIC cards, we can also use the airflow internally in a sealed environment to grab the heat off everything else. In fact, we’re independently validated by the Research Institute of Sweden to recover nearly 90 plus percent, over 90 plus percent.

of the entire heat inside the server and that’s the asset. You recover everything. We see everything get hot and want to recover it. So the hotter it gets to a degree, the more efficiency we have in recovering it and then we can optimize the chip’s performance at the GPU and CPU level.

Wes Ashworth (11:36)
Yeah, absolutely incredible. And again, recovering ninety percent of the that heat is just a remarkable stat. as we think about this, so as AI chips get hotter and racks more crowded, where is that sort of air cooling finally reaching its limits and maybe where does it still make sense?

Kenneth O’Mahony (11:52)
Yeah, it’s interesting. suppose in principle at the low wattage it makes sense. It kind of makes sense because it does. But reality is we get a little bit, you know, hot under the collar with that even prospect. We kind of say you should recover everything. Why lose out on the opportunity? It’s like saying my Glen Dimplex heater should be only half on. You’re saying it’s still, you’ve paid for electricity already. You’re paid for it. Why not use it? So recover it. So it’s recovering everywhere. But I do understand the principle.

And it is that the wattages, know, at chip level now and server level, you’re up at, you know, 100, 200, 300, 400, 500 to a thousand watts in a server, not to mind at the rack level. So AI ultimately wants to rack and stack as much as it possibly can. And our technology allows them to do that. Single slash, GPUs, all capacity optimized and then the recovery. And then when you do that, when you recover it with a sealed server,

Ultimately it changes the design of the building infrastructure as well so you get rid of the hot aisles, the cold aisles, no air flow. So the combination of the AI pushing the knowledge base into, yes we must have opted up liquid and then the full ceiling allows you to do everything. Reduce the level of electricity in, optimize the performance and full recovery.

Wes Ashworth (13:05)
Absolutely incredible and easy to see. You I think the clear advantages of your technology and why, you know, probably just a just a better solution overall as you’ve just described for a lot of different reasons, even in those the small applications or what have you. So Nexalus also describes water usage effectiveness of zero. For someone outside

the industry, like what does that mean in normal operation and what water is still inside the system?

Kenneth O’Mahony (13:28)
Yeah, you can see the tradition or the movement in America especially to moving data centers to where there is water because ultimately they still draw water into the whole infrastructure and design. We don’t. What we do is when we seal the server we have a closed loop system, a bit like a combustion engine. One time fill, there’s a glycol inside there and industrialized water that has treatments inside there for antimicrobial bacteria and then in certain temperatures there’s different elements.

it’s 90 plus percent water based still, it’s got your green thumbs up. And then for the rest of the system, the system is water usage effectiveness of zero. It’s funny, so we have water usage effectiveness of, we do not attach to a district water system in any way, we don’t consume water, we don’t distribute it back out in any way. The second one is we can, we’re a cooling company that can cool with warm water.

So we’re a cooling company that can use warm water to get to higher water outage for the end user if we get full recovery. And then we’re a water cooling company that has water usage effectiveness of no water use at all. One time closed loop.

Wes Ashworth (14:29)
Yeah, absolutely incredible again. and I think that distinction matters a lot in places where households, industry, and data centers are all competing for resources. I think as if we’ve talked through, so like once you capture the heat, the question changes. You know, instead of asking how to throw it away, you can start asking what useful work it could do next. And so you’ve described a data center as an energy borrower. I love this term. Like, what do you mean by that? And help help explain that a little bit to us.

Kenneth O’Mahony (14:53)
you have to commend the industry. A lot of the big players in the industry, they’ve got green energy in. They invest a huge amount of money in green energy. So they consume green energy. And you have to give them thumbs up. But ultimately, at that point in time, is where the differentiator happens. We’re still consuming that energy. And then we’re engineering incredible ways to dissipate it, to get rid of it. Whereas ours, we believe that you could imagine a world where, just the same way as you might say, I get my energy from…

gas or oil or coal, you say, well I get mine, my hot water is actually given to me by the data center that’s in the basement of my apartment complex. So it’s connected to the district heating system, it’s on 24 hours a day, 365 days a year, it’s producing the seats, it’s like your heater is plugged in and now I’m recovering everything from it. So then the energy used by our liquid cooled server, one, they’re using ballpark 30 to 40 percent less electricity to produce the same output.

Then they’re producing the hot water and they’re giving that back. So there I feel they’re borrowing what they use, they’re using less than they need to and then they’re creating this commodity. That gets the end users off the grid. So now you have this perfect circular economy that the end users source of heat. I think again, on average in the States on the high 80s percent of entire electricity consumed at the domestic or utilization level is for heating hot water. And there you go, you now have it from your data center and your provider is your data center for your hot water.

Wes Ashworth (16:20)
Yeah, what amazing potential. And I really like thinking about the possibility of electricity doing one useful job as compute and then the resulting heat doing another. And I think changes the question altogether from how to get rid of the heat to who might actually value it and how can we kind of use this benefit of it. and obviously heat is only valuable if someone can use it. Why does the temperature coming out of your system just matter so much? Like what else is to that story?

Kenneth O’Mahony (16:43)
Yeah, you lot of certain you’ll hear a lot of stories in fact some of your listeners might think well I’ve heard this story before there is recovery in the system but well going back to my days and when I was with EMC we did that we did the most efficient that there was we end up with what’s categorized on the Fahrenheit of 77 to 95 Fahrenheit low-grade heat which means you need more pumps more energy consumption to bring it bring it to usable commodity size and if you find a user or located near a user

we give it 50, 60 degrees, which means it’s a usable commodity straight away. One of the concepts we have, and one of the things would be very interesting is, if you find a user that needs a 24-7, 365, for example, a chip manufacturer, they could be ironically using a server, data center, chips, heating their own fab manufacturer for the output of the end chip, which itself is made from the heat and generated the heat from their own chips. So this utilization location such as

apartment complexes, swimming pools, district heating as well as anywhere else that has high utilization. Then always knowing if you don’t have a use there’s more. So the extra bit there’s also huge technology associated with carbon capture and there one of the utilizations that need one of the commodities that they need is hot water. So getting the commodity that is water up to 50 to 60 degrees Celsius and then bringing it to the next level for the

Carbon capture if you don’t end up having the user directly connected. It’s always there’s

always some use for this hot commodity for sure

Wes Ashworth (18:11)
Absolutely. And tons of uses there and a lot of beneficiaries, when we think about who could benefit from that. I guess of all those, who is most likely to use that recovered heat first, whether it’s nearby homes, an industrial operation, agriculture, or somebody we’re not even thinking about.

Kenneth O’Mahony (18:25)
Yeah, do think in industry, because you have a constant flow, could be something to do with grow farming, but then you’ve got that ecosystem of being located nearby. So then I do believe there’s, you know, if you design cities of the future, if you design infrastructure, commodities of the future, we always associate, you must have your electricity connected, you must have your power connected, you must have your heating connected. So why isn’t it in there as a data center, as your end user?

And then as I said, I do believe there are other advanced technologies very far advanced on carbon capture. And one of our engineers said before is we have two responsibilities. One is not to produce anymore, but we have a secondary. There’s too much there already, so we’ve got to do carbon capture to offset what’s already happened. And that perfect combination will be really good with the end users.

Wes Ashworth (19:11)
Yeah, love it love it again. And again, I think clear wins for everybody around there. and obviously as we think about this, like there’s a difference between heat your system can capture and the heat somebody actually puts to work. what has Nexalus proven so far on the second part?

Kenneth O’Mahony (19:25)
So what we’ve proven is the capability of it, our company has been focused very strong to date on edge, rather than the specifics of internal data center and why that’s important, it’s bringing the capability. So what would have been perceived before is there’s a limiting factor that the only place in the world where you have a GPU and AI and CPUs configured in the stable space would have been a data center. There’s now a huge demand for also bringing that capability out to the edge.

And that’s where we’ve had the first wins in the company, deployment out on the edge, bringing AI GPUs, AI CPUs configured in much smaller. We’ve displaced buildings with small boxes, only 16 liters high. The building has gone and now the entire capability is derived from the box. And then the scalability of that is rack and stacking of that at the data center scale. So that’s the trajectory that we’re on and we’re on a growth trajectory for that and having very strong discussion. So very excited to see the edge deployment where

both the compute and the thermal challenge was the hardest. Getting that over the line and then bringing it to, the easier place, which is inside in this stagnant stationary building, the data center.

Wes Ashworth (20:33)
Yeah, very exciting trajectory there. And I think just that distinction overall helps separate the technical opportunity from its realized impact and gives operators a clearer way to decide what benefits they can count on today. If the nearest heat customer is miles away or doesn’t need heat in July, what becomes the hardest part of making this work financially?

Kenneth O’Mahony (20:51)
what we’ve talked about most of all is the end and we believe it’s right because we did this balance, did performance, profitability and planet and a lot of what we discussed there is the planet. The first one is performance and profitability as well. But the key there is if you utilise without recovery, which we always think there is a way, but you know, if maybe there isn’t, maybe the user is too far away, you’re also around 30 to 40 % reduced on energy consumption for the same output.

And that’s the stick. So regardless of if you can do this full circle, your utilization of energy consumed is so far reduced. Second benefit is the concept of the size of what you’ve built to perform it. Your data center size itself reduces so significantly, which are reduction in hot oils, cold oils, your plenums, you’re not moving air around. So the adoption of the system allows the infrastructure cost, the setup costs.

and the operative cost to be significantly reduced and then hopefully you still find the end user with a bit of creativity in the world I think there always will be one.

Wes Ashworth (21:53)
Yeah, no, that’s absolutely incredible. I’ll ask you this question. So does a Nexalus system still make economic sense for a data center operator who can’t reuse any of the heat?

Kenneth O’Mahony (22:01)
Absolutely significantly and you can see it coming from the start from the adoption of liquid cooling as a whole But ours gives you the second in the tradition. They’re still air blowing So they’re putting their on-chip cooling and they’re blowing the air. It’s still a reduction in energy consumption So it’s a good progress in engineering, but when you seal the server, that’s your significant reduction You’re no longer pushing air around as a commodity. You’re just pumping liquid in a confined space and that’s your

30, 40 plus percent. And the reason it varies between 30 and above is because these new configurations of the GPUs condense servers much more. The higher the density, the more the saving because of the benefits of the Nexalus cooler.

Wes Ashworth (22:45)
Yeah, a crucial question, I think, for operators who need the cooling investment to stand on its own merits and separates that

direct operating case from from any additional value a heat customer might create. And obviously those are there in most cases as well. But I think the idea is powerful here as we’re talking through it. But when an operator is betting critical infrastructure on it, the real test is whether the numbers and hardware hold up outside the presentation deck. And so want to just dive into this a little bit. So you’ve shared some significant energy saving figures across different applications.

when an operator hears one of those percentages, what baseline and what amount of computing work should they compare it against?

Kenneth O’Mahony (23:21)
it’s a very strong question because the… and it is a variable answer because the configuration of the server itself gives you the baseline to which it’s there and the output of it. As we look at the AI era, it’s probably the best one to focus on, is what tokens per second are you generating? In the world gone past, we focused on different litmus tests to check of what validity of our solutions were.

But now the world has aligned very strongly to the world of how many tokens per second per kilowatt. And that’s the one where Nexalus is really shining. It’s a simple equation. You’re using the same GPUs, you’re using the same form factor, and yet the tokens are… We will be launching a product in partnership with one of our partners, Renda, in November. And you’ll see that there’ll be a ballpark 20 % more tokens per second produced from the same output.

So now you’ve got a touch and feel of how efficient the system is and how effective it is and reliable.

Wes Ashworth (24:19)
Yeah, huge, huge numbers there, and impressive when you’re talking about 20% and all of that. And really it’s you know how an impressive result becomes something a customer can actually evaluate and budget for. So it’s cool to see those kind of metrics come together. Intel has described more than two years of collaboration and demanding thermal testing with your team. Like, tell us a bit about that and like what did that process force you to prove?

Kenneth O’Mahony (24:41)
Intel took it out onto the edge. Again, we met Intel. We’re a science and engineering company for design, so we really collaborated very well on what the success looked like for both companies. And they gave us a challenge in relation to taking a server, a data center configured server, putting it on the edge on telecommunications, and seeing how it worked.

To say it worked is an understatement because we actually put it in Texas. We were thinking one of the biggest challenges would be the heat, but strangely enough there was a cold spell minus temperatures in Texas during that period of time as well. So we got exactly what we wanted. We got the double challenge. It was left there for nine months, untouched, unserviced, and then opened and reviewed afterwards. And it did exactly what it was supposed to do, what we knew it would do.

and what Intel were hoping it would do. And Intel have been a phenomenal partner. They’ve introduced us to a good few of their client base and we have been, luckily, with the commercial benefits of the white paper that is published on our accounts. And we’re actually working on a new white paper for them at the moment because of the advancement in new chips and where they are. Those challenges that you present them like that, they’re the teamwork and the collaboration that really result in commercial output. They’re all very attractive.

Wes Ashworth (25:55)
Absolutely remarkable results there. And testing at that level makes that leap from promising science towards something an operator can trust and just looking at that scalability and things like that as well too. another area, so Formula One is another fascinating proving ground for your cooling technology. And I think it’s super interesting. So I wanted to bring that in there, but what does that world teach you about the kind of performance AI operators are now chasing?

Kenneth O’Mahony (26:17)
Yeah, look, there’s everything to do with engineering. F1 is an extreme heat in a tight space where weight and size and reliability all matter. Nothing is oversized. Every component is a challenge in how it fits in. And ultimately, it has to be high performance really quickly. And our partnership with Renda and Ian Parry, 8PACK, has enabled us to be very successful in the Formula One space.

performance is aligned directly to cooling. And every second that you can achieve higher grade performance from your commodity, which is their servers, we’re lucky enough to have gained real degrees of headroom there for them. And then we adapt it into everything else we do on scale.

Wes Ashworth (26:59)
Yeah, it’s so cool. I think it’s fascinating when lessons from such a specialized environment start showing up in mainstream infrastructure, and you’re talking about high heat, high performance, and those parallels, you translate that across the other areas as well, too. So super cool. you’ve moved into partnerships with major server companies and selected Alps, Alpine for manufacturing. What’s hardest about turning a breakthrough into something thousands of operators can reliably deploy?

Kenneth O’Mahony (27:22)
It is that I sat in my world of EMC, I sat in probably the opposite side of the table so when we were sitting this side I knew what we needed to create, I knew we needed to create assurity, I knew we needed to have reliability and I knew we needed to have scalability. So we put our supply chain together in partnership with Alps Alpine and then we transferred our entire supply chain over to Alps. They went through the green procurement process, one of the most stringent processes in the world for quality reliability.

And of course Alps Alpine are famous for everything to do with automobile, so manufacturing at scale, constant reliability, and it’s a brilliant partnership that we have nurtured and developing, and also scalability globally. So while Alps Alpine have a presence here in Ireland, our partnership with them is truly globally, because some of our client base need built-in jurisdiction, so that gives us that capability as well.

and their name and brand, they’re very excited to be in the ICT industry with us and we’re really excited about the scale and capability of our partnership together.

Wes Ashworth (28:24)
Yeah, it’s incredible. I think having talked to so many companies that have are scaling, have scaled, you know, been on the other side of it one way or the other, how important those kind of key partnerships are and collaborations are.

so it’s always always cool to feature those. And I think, as we know, the market needs more than impressive science, needs supply chains, as you talked about, repeatability, reliability, a product it can live with for years. And those are the that’s the real challenge, right? taking it out and then being able to scale it. so liquid cooling itself is becoming mainstream. I don’t remember hearing too much about liquid cooling, you know, years back, but now it’s a daily thing, right? If if an operator asks why they should choose Nexalus over another strong liquid cooling solution.

Where do you believe the real difference is?

Kenneth O’Mahony (29:04)
Yeah, you’re right in so far as it is. It’s become, when we started in 2018, we were very much pushing, now we’re being very much pulled into conversation. I think there’s three elements to it really. Firstly, it’s the competency of the solution. So you’re not peddling a product. It’s not a specific product. It’s an engineered overall architecture and solution. And then when you focus on three things, you focus on…

the performance, you focus on profitability and you can focus on the planet. And usually in life you sacrifice, there’s something, if we reflect back on something like ESP, electricity I should say, we know we’ve suffered all the time in relation to coal has been the biggest contributor to the electricity. And we know that is something that we’ve engineered, majoring new solutions for to offset that problem. So right now at Nexalus

It’s accepting that the world of data centers, the world of AI is here to stay. And it’s such an advantage to our society and something we wouldn’t want to reverse. But we just don’t have to make the same problem and error that we did do with coal. We have a solution. And ours aligns us to capable of sealing the server, increasing the chip, keeping the chip at optimum output, keeps the performance better. You get more compute out of the chip, more performance engine, the engine of.

the whole infrastructure is in a happier place. You’re consuming a lot less electricity in doing it. Your infrastructure, your design of the buildings, everything changes, and then you can recover. And when you focus on all three of them, it becomes this compelling argument. Some of the solutions out there do one of those, and they do it very, well. And had we not gone through the world of advancement that we have gone through, we never would have ended up where we are. But when you deal with people and the capability of our PhD engineering team,

who look at things as everything is a challenge and a solution that’s ready to be made, and then you get partners such as Alps Alpine and Intel and Nvidia, you then can be happy with your progress and say, that’s why Nexalus is the solution is the one to choose.

Wes Ashworth (31:06)
Yeah, incredibly well said there. And I think covering all of those areas that you know, the advantages are really clear. and I love that I love that thought process of just one, I think learning from the past and

getting ahead of it, right? Addressing those concerns from the from the start and kind of innovating and creating and coming up with those solutions. It’s remarkable to me. To me, I think it’s hard to stop progress, right? Especially technological progress. We think about, you know, go back and stop electricity or stop the internet or you know those kind of things.

you probably weren’t gonna do it. But how can you how can you then think about it through a lens of okay, it’s this is happening, how can we make this the best possible solution? How can we you know make the best impact on the planet?

On people on the world as it as it be. And I think that’s the right mindset. I mean, not everybody agrees with me on that, but I do think you’re looking for those solutions rather than just harping on the problems, but how do we how do we improve it? How do we make it better? So I love that a lot. Love that answer a whole lot as well, too. so much of that decision, you know, kind of like happens inside a data center, you know, outside the fence line, communities are asking whether the benefits of this infrastructure justify the local costs. And we started talking about that.

And a Gallup survey found that 71% of Americans opposed building an AI data center near them. More than oppose a nuclear power plant in terms of that stat. So what are communities telling the industry that developers still aren’t hearing? And what’s your perspective on that?

Kenneth O’Mahony (32:30)
Yeah, I think they’re valid, know, I really do. And I think there’s uncertainty in relation to maybe the industry because it’s always been this way. There’s the bias and the unconscious bias. I do truly believe it’s like when there is an advancement, when there’s when something new is created and we didn’t know before, if we stay in where we were, we would never advance. And I think that’s what this is at the moment. The industry acknowledges that there’s a lot of challenges. The perception of water.

utilization, the space being used up, electricity being done. And so I think we’ve got to embrace those conversations rather than we maybe we’ve approached the industry saying, look, it’s you need this and we’re going to deliver it at all costs. What I do believe that there’s been, as I said before, there has been advancements and credit to that. But this one is a leap. This is the dream, because if you could design it, if you asked the society, what would they want?

They would want a commodity, they would want to consume less electricity, they would want an offshoot of a waste that’s converted into an outcome that’s beneficial to the community. No one would want their hot, their water, natural water taken into an industry and distributed back out at a higher degree. We know that’s got a negative effect on society. So you don’t want any of those things and they no longer lead to happen, which is good. And then you also want to accept the fact that the outcome…

can be efficient from all levels. So less CO2 because your construction reduces and an understanding that these concerns are legit, but there is a legitimate answer to them and there is a legitimate benefit to this. So I think it’s a combination of open communication, but pride in the fact that engineering has advanced beyond what maybe is understood as maybe some of the problems, but acknowledging they’re there and then treating them with respect by showing them that the solutions are there and they can be embraced.

Wes Ashworth (34:15)
Yeah, great answer there. And those concerns do deserve to be taken seriously even in even when people value everything data centers make possible. And as we’ve seen this in a lot of different technologies, having come compelling technology story isn’t the same as earning a community’s trust. And I think meeting people where they are

Helping be in the community inform, educate. they’re often making assumptions. Even within the industry, we do it based on, something that was true five years ago. it we’re we’re moving forward pretty quickly in this technology advancing.

So it may not be true today, but who’s in there telling that story and meeting these people and educating and helping them understand that? So I think that’s so important.

So another kind piece of this.

If cooling can reduce electricity overhead and avoid ongoing evaporative water use, what environmental problems does that solve and what important problems still remain?

Kenneth O’Mahony (35:02)
Yeah, mean, we’ve gone from, we were accepting that the big energy, the big data center operators produce green energy and invest it hugely. So that’s one end and it’s good. But at same time, then if they consume all of it and then they produce a waste that’s now negative, you’re still on the net negative in my opinion. So the strong way is that you have to accept the fact that

that you don’t need to use evaporative cooling anymore, do not need to use water in any way, you can do what they call, if you don’t have an end-use, as we discussed earlier, we think there are lots of them, but maybe you don’t, you have dry coolers. And dry coolers are not, and don’t touch anything to do with the district water system, or a metropolitan water system, so that should remain untouched, that’s a given, and it should be adopted. And the second definitely is the recovery. So once you have the recovery,

and then you really take a third end user off the grid and they’re using your waste as one of their key commodities. You really have an ecosystem that everyone can value and evaluate and be happy to be farming in. And you then, then I think the embracing of the data center in its form factor of core big data centers as well as edge data centers deployed around the place, they will become much more acceptable. less smaller and impactful on a negative way and much more positive impact on society.

Wes Ashworth (36:20)
Absolutely. I think it’s helpful to separate a meaningful improvement, the cooling system, from a broader footprint of the data center. And I think a clear-eyed view of both the benefits and the remaining constraints makes the bigger vision easier to trust. And so I think that’s important to go through. I guess, you know, thinking about that, like how should operators sort of measure those trade-offs honestly?

Kenneth O’Mahony (36:39)
You do, I you know, mean, there’s the meters, the commodity that it is, electricity is metered. So we know exactly what we do. We know what we consume and what the output is. And once you can prove that you’re metering, your consumption of electricity is falling for the output that you’re doing, also maybe the physical size. mean, imagine if when we get rid of the hot and cold aisles of a data center, if we get rid of air cooling, you can now put more servers in. Yes, you do need more electricity to run them.

but you’re producing more output. And I do think those very measurable features should be part of planning processes, should be part of engaging with society. as you onboard the community in which you’re going to prosper, that they believe and nurture, they believe that what you’re adding is a net good to their society, I think you’ll have a much more engaged community. But there is that perception.

versus the reality and the adoption needs to be shown to be true.

Wes Ashworth (37:34)
Absolutely. And with that perception, you it’s no easy feat. You’re trying to change an industry filled with experienced engineers, consultants, and operators. Like what’s actually helped you change minds rather than just win arguments?

Kenneth O’Mahony (37:47)
It has to be the engineering. We’ve proven it. We’ve proven that it does do what it says, but it hasn’t been easy. would say that has the biggest challenge. Being honest, it’s the biggest challenge because even so-called experts, Tony Robinson would have discussions with people who still believe that certain form factors of two-phase, single-phase, one’s better, one isn’t. And as Tony said, they’re right, but not right here. Because what we did, when we did our research, is we didn’t come to it with an answer. We came to it as a blue sky research and said,

This is the optimum. We want to optimize the performance of the chip. We want to create profitability and we want to protect the planet. Let’s do it. So we’ve researched every type of liquid cooling that there is. And as Tony said, in different circumstances, jet impingement would be just wrong. But it’s just right here. And it’s very right. And once you start to have those conversations, you prove it out and then we get validation from the Intel papers and also our end users bring it to market and seeing the benefit and the capability of it.

then the rolling ball as we call it is starting to very much roll in our favour. We’re having much more conversations about it in the public domain.

what we decided to do with your approach was crawl, walk, run. And so you prove it out in the lab, test it to the highest level. We had good financial support and support from Trinity University. Then you bring it to test beds and prove them out. Then actually do the white papers, then take it out to the industry. You prove out the capability and then now what we’re doing is scaling it. And luckily that’s the trajectory we’re under.

Wes Ashworth (39:18)
Yeah, it’s awesome. I love that quote too in the perspective of they’re right, just not right here.

that’s a very cool way to think about it. And I think a breakthrough is only useful if people feel confident enough to adopt it in a system they already depend on. And again, no easy feat there, but that education, proving it out, validation relationships can become as important to adoption as the technical invention itself.

we’ve talked through this, like there’s the engineering work and then there’s a human work of just changing an established industry is what we’re starting to talk on there. that bigger division I think depends on both and kind of bring those together. So looking back, you know, over the journey from finance executive to building Nexalus what’s one belief you had to change or one mistake that made you a better founder?

Kenneth O’Mahony (39:59)
I would say it was Tony Robinson at the start when he presented us at the end that day and I was telling him, this is the way it’s done and it’s always this way. I remember him shaking his head and he said that can’t be true. And I said, but it is. He said, this is done in other industries Ken, so why isn’t it? And that was the reality check is understanding and knowing how difficult it was going to be that

You as was very successful and as was still and is still to a degree growing in trajectory It’s just that the touch points were starting to come and you could see it Intel AMD Nvidia They all told the industry many years ago You won’t be able to cool this with air and the industry kept going we will we will and then eventually you can see that It’s starting to go this way then being Tony and the team that we’ve hired that wasn’t enough They wanted to push it more and at the time that was

I had to take a step and say, right, I’ve got to accept this. So there’s more to this than the one that was ready to go, the one that was marketable ready, hence why we crawled, walked, and run. And I was probably would have pushed a bit sooner to bring what we had to market. And ultimately, we were right in waiting because the Seal server, getting it to the optimum capability and having the biggest impact is the one that we’re proudest of and the one that we’re now running with. So definitely, superb engineers.

challenge you from a commercial perspective as well as what’s capable out there in the world today. They’ve changed the world in a very positive way we’re very proud of it.

Wes Ashworth (41:29)
Yeah, I love that a lot. And It’s the part that people don’t always see, how building a company just changes the way the founder thinks too and that development that happens along the way in your own life. And I think the lessons behind the technology can be every bit as valuable as as the final design. anything like thinking back, if you went back, anything you would do differently today?

Kenneth O’Mahony (41:48)
Yes, I think I would have measured the as is with a much better vigor and we kind of thought that the industry had the as is as in their heads and we would just show the difference and Sometimes the difference is changing so fast the as is was adapting very quickly as well at the same time so it was probably better for us to have that or origin and then show configurations and also The bigger one was probably yes, we can do that, you know

The ask was, yes, yes, yes. And that enthusiasm is still there and it’s vital to have it. But probably, maybe a little bit along the way, we might have said yes on version two. So you could have got there earlier with some things and then said, look, that’ll come in version two. We’ve got a very good balance of it now between the team because the challenges that we have with the client base are really exciting. So everyone’s excited to deliver it. But at the same time, we realized that if you keep on…

saying yes and keep on waiting for the next version You never really commercialize and you get stuck in the R&D stage for too long So we segregated ourselves very very well in their location as well as her strategy still very focused on R&D It’s a huge part of what we do and our partnership with Trinity and but at the same time We know when something’s ready and we have great partners as I said with Alps to bring that to market now Which is which I would have I probably would have done that a little bit different on reflection for sure

Wes Ashworth (43:07)
Yeah, absolutely. the journey that takes us there, right? it’s the great wisdom there and the lessons learned and going through it and developing and I always love that as well, too, and putting it out there to other people that are listening can learn from that perspective as well. so we’ll end on this. So picture a city ten years from now where the data center down the street helps provide heat for homes or businesses or industry. what has to change between now and then for that to feel just ordinary instead of extraordinary?

Kenneth O’Mahony (43:33)
and it’s a strong question. in my research on it, when I was thinking, I have to say, it’s universal globally as well. It’s one of these things that the ideas are there, but there’s an awful lot of administration. And I kind of break it down to three. There’s planning, infrastructure, and business models. So the planning, without a shadow of a doubt, it just needs to understand what’s available. And it must be hard when city planners

don’t understand what’s already available. They don’t want to hinder progression in their city or state, so they want to be friendly to the environment of any industry that they’re welcoming into their environment. But at the same time they need to understand that what’s able. So when you plan something, what might be perceived as restriction is no longer as restriction because it is able. Then that helps the infrastructure. So you align it with infrastructure like a gas line or a network or networks.

and you just align that a data center is embedded in the infrastructure story and then you do need the model. So how is a server data center that generates the heat commoditized inside in the market where it’s a new source of power? And we do know those markets, commodity markets are exceptionally available, but no one sees that the data center is the energy source embedded inside and how do you trade that?

But it would be a trade because it’s 24/7, 365 it never turns off so using it option options available for it and Maximum utilization for it is all needed for the future, and I do think the heads are there It’s just some small tweaks of understanding and I think we can get there really well

Wes Ashworth (45:08)
Yeah, absolutely. Exciting vision of future. And I think, you know, a very different future if useful heat becomes something we design around rather than throw away.

And I love the possibility of just judging infrastructure not only by what it consumes but what it gives back and that being sort of a key thought here. But Kenneth, you know, we’ll wrap there. I think it’s a beautiful thought to wrap on. Thank you so much for being here. I appreciate how you’ve challenged us to look at something the industry has treated as a cost and ask whether it could become a resource. To everyone listening, thanks for spending time with us on Green Giants. If this conversation gave you a different perspective, share it with someone working in energy, technology, or community development. You’ll find links and resources.

in the show notes, be sure to check those out. Subscribe to the show, leave us a rating or review, and follow Green Giants on social media for more of these conversations. With that, we’ll see you next time.