Ep. #40 Mark Bjornsgaard, Founder & CEO @ Deep Green Technologies | Data Center Go-to-Market Podcast
Mark Bjornsgaard: We're in a very different world. And we're in a world where the world of AI has emerged and very power hungry, very, very, very heat generative GPUs, which, of course, because they're so expensive, people want them on all the time, which is brilliant for heat reuse, because then you can offer a consistent, consistent and ongoing heat source. Hi. This is Mark bionchko from deep green you're listening to the data center go to market podcast, hello data Joshua Feinberg: center, friends. It's Joshua Feinberg from the data center. Go to Market podcast, I'm thrilled to have a very special guest joining me here. Today. I'm
welcoming Mark bianchkort, who is CEO and founder of deep green technologies based in the UK, Mark. Welcome to the podcast. Mark Bjornsgaard: Thank you very much for having me. I'm happy to be here. Thank you. Joshua Feinberg: It's my pleasure. I'm thrilled as well. I think the first place that would be super interesting to find out is you're doing some really cool stuff around heat reuse and sustainability in the data center industry. I'm really curious to know what was the path that led you to this? Was this something that you've been working on throughout your whole career? Is it something that you discovered way back when you were young, or earlier on, when you're going through formal education, like, what was the path that got you to this place where you're founding and leading such a transformative kind of business for digital infrastructure. Well,
Mark Bjornsgaard: it's like, like many people in the data center industry, I think it's somewhat circular in nature. We don't we, I didn't come here. Sort of Off, off the bat, if you like. We developed the business eight years ago with a company called British Gas in the UK, and then we kept it on the shelf for many years. The data center industry really wasn't ready for the whole kind of concept of heat reuse, and in any case, we were always standing on the shoulders of a vagina. In this
case, case Microsoft, who had written a white paper in 2011 outlining how the industry could think of at some point in the future, capturing heat from data centers and from computers so and I wasn't I, but even back eight years ago, I certainly wasn't the brains of the operation. We set the business up with a good friend of mine who had read the white paper and we built the first prototypes. But it's taken us, yeah, taken us something like eight years to get to to the point where we are now, and and now, I'm pleased to say the industry is ready for it. And in many ways, it's, well, it is certainly we, and you'd expect us to say it, but it's, it's, it's certainly an idea where that where, whose time has come. Not least
because, as we know, the data center industry is struggling to find power. And if you in, if you, if you, if you reuse heat, as we are fond of saying, you can use the electrons twice, so you tend to get your your power connections faster. All sorts of good, great things happen. I'm sure we'll, we'll, we'll
hopefully dig in some of those details. Joshua Feinberg: So it was that foundational, uh, positioning paper that Microsoft published in the early 2010s that was the seed of the idea that this is possible. This is something that could be possible. So,
Mark Bjornsgaard: yes, so, so that was it. Yeah, that was and, and it was in that paper that Microsoft coined the phrase data furnace. So it was actually Microsoft who came up with with the words. The truth of the matter is, as we know, workloads in data centers have been data centers themselves, the utilization rates even of the hyperscaler data centers, the relatively low, you can be as low as 20, 30% and and on that basis, it was always quite a struggle to imagine a world where you could provide continuous, decent heat from a data center. So there was, there was clearly kind of major
operational and technical issues around doing this. So 15 years ago. But now, of course, we're in a very different world. And we're in a world where, well, we've certainly the the world of AI has emerged and very power hungry, very, very, very heat generative GPUs, which, of course, because they're so expensive, people want them on all the time, which is brilliant for heat reuse, because then you can offer a consistent, consistent and ongoing heat source. Joshua Feinberg: Yeah, it's actually really fascinating to think that Microsoft had the forethought to be projecting this as an issue, because I don't remember exactly when Azure had its big run up in market share, but I don't think they were growing like gangbusters in the early 2010s matter of fact, the part of Microsoft that I was involved in around that time was still struggling with a lot of the channel partners trying to reinvent themselves from premise based client server to SaaS and cloud, Mark Bjornsgaard: right? And we are an we, I sort of. Again, font of, fond of, we love our hyperscale cousins, and we are we fond of being rude about the experiments that Microsoft particularly have been doing around, you know, dumping data centers in the sea to keep them cool. Because I, you know, just in case the seas aren't warm enough or warming enough. But
actually, to be fair, completely fair to Microsoft. They have, yeah, they, they, they, they championed this idea, and they have been not all experiments are good. That's the whole point of experimentation, right? So, and you know, we've got Microsoft, Google, meta, they're all now experimenting with direct air capture from their data center. So whilst, I think for obvious reasons, the hyperscalers provide somewhat of a of an easy sort of target for people. And there's, obviously,
there's a lot in there, in the green reporting and the electron use, which, which you definitely question whether it's been completely straight down the line. Reality is, yeah, they've been leading the charge on this. So, you know, kudos, credit, credit, where credit's due.
Joshua Feinberg: Yeah. And this topic is certainly capturing a massive amount of attention in the mainstream business media and even in the digital infrastructure media. Just before we hopped on to record this podcast, I saw a headline from one of the trade pubs. I'm thinking it was DCD that I think it was 43% of the energy capacity globally is now within the control of the hyperscalers that the data center industry is going to be able to utilize. So, yeah, it's very high. It's the opposite of distributed, right? The Mark Bjornsgaard: opposite of distributed. And then we will, I'm sure, again, we'll probably talk about it later on. But this
whole notion of the trillion dollar cluster, you know, the idea of of the cluster we need to build to a birth artificial super intelligence. That's 100 gigawatts of power. That's 20% of the entire electricity generated in the US now that we need, we need 20% more of the entire US electricity budget to then run the trillion dollar cluster so and, and it's really worth unpacking why decentralization is so good in this way, because when you decentralize, you can, you can, you then give yourself the optionality to generate and store energy on site, which has, certainly in Europe, has all sorts of very good benefits in terms of pricing and availability of electrons. So for example, we've got, we, we have 6.5 trillion, sorry, 6.5 terawatt hours of curtailed wind in the UK, which means we waste 750 megawatts of electricity all the time in the UK. So if you can, if you, if you have to land
that amount of lithium anywhere in one single place. That's a heck of a lot of lithium batteries to land. If you do it in a decentralized manner, and you then allow yourself to put a load of a lot more solar around a smaller data center. You can run more and more the data center off grid, which means you, which means you put less stress on the grid, which means you get, you get to grow AI clusters faster, essentially, Joshua Feinberg: less pushback from the community, exactly, less Mark Bjornsgaard: push. And that's really important because, because everyone now is in a race to the electrons. So if a
data center says, Well, I want, I need 100 megawatts for a data center and a district heating system, a very large energy center that's going to heat flats and offices, says, I want 50 megawatts. Then the 50 megawatts is going to in the UK and the Europe at least, you're going to be way ahead of the queue, because you're perceived to be, you're perceived to be in kind of from a social good point of view, and from an economic good, you you're you're ahead of the you ahead of the data center. So in our case, when we build a data center with a district heating system. So when we build a 20 megawatt data center in a district heating system, that all we're doing is just getting ahead of the queue, really, in terms of power allocation. Joshua Feinberg: So switching gears a little bit by the nature of what deep green is working on, the initiatives and the transformative impact that your company could have in the coming years. It really crosses all different facets of the
ecosystem, the operators. It impacts the technology companies and the facility companies, for sure, construction, real estate, how capital is deployed, given that when you're approached by people that are very, very early on in their career. Perhaps they're still in university figuring out what they want to do in their first job to graduate in a teenage son or daughter of a neighbor or friend or family member approached you and said, Mark, I've heard that there's a lot of great opportunities in digital infrastructure. You seem to have a pretty good pulse on what's going on in the space. What
would you advise to someone that's interested in this kind of career, but not so much on the technical side. They don't see themselves as developer, engineer. They're interested more in client facing roles. What kind of advice would you
offer to someone at that stage in their Mark Bjornsgaard: career? Yeah, I've always very hesitant about offering advice, but as they talked about, certainly my. Experience in it with in the data industry specifically, I think it is no different from other businesses that we've invested in and other sectors, which is, I'm, I'm fond of sort of saying you can't, it's very difficult to stay in your lane now. So I think a lot of people coming to any industry, whether so. Take autos, for example. You know, it was really instructive
to see the head of Ford come out, where it was kind of six months ago, and say, we're actually a software business. And if you're, if you're an autos manufacturer, you're now a software business, you're now a you're now an energy business as well. You're an entertainment business. And I think that's the same with the data center industry, but I think it's even, perhaps even more acute, because it's a relatively new sector, and there are so many aspects of the industry which are which, which are coming together and changing the nature of what the data industry does. And we have, we talk about the seas of change so that there are, there are, there are, there is kind of, at least 6c around that kind of, are changing industry. And those are, you know, you've got to understand energy trading.
You've got to, you've got to understand, you've got to understand the ESG impact and the kind of circular economy of it. You've got to you've got to understand sort of energy generation, then you've got to understand the existing real estate version of the business. But then on top of that, you've got to understand what's going on in the software layers. Because, as we talked about 2015, years ago, some data centers were only kind of utilized 10% of the time. But what's happening in the software layers in terms of this migration from kind of mainframes all the way through kind of virtualization, and then Kubernetes, and now we've got WebAssembly and rust and all these decent these zero trust protocols. This means that the software layers are going to
fundamentally change the asset utilization of the real estate because instead of a data center being on 20 30% of time, it's going to be on 99% of the time, even non AI workloads. So I think anyone coming in the joy of, frankly, the joy of, of working in the world as we know it is that it's no longer, you know, you get into the data center industry, and you kind of understand it as a real estate game, or it's really, it's so much more. And I think, I think for us oldies unfortunate enough to have turned 50 this this week, it's, it's, it's, it's for somebody new coming into any industry, again, it's where everyone is always so anchored in their own reality of a business. And I, I've invested in lots of different sectors in my career, and I've it never fails to amaze me how those who've been working on an industry are are, will tell you, and will be so fixed of a fixed mindset. So I often say it to
youngsters coming into any industry, your ignorance is a superpower, because you can just ask dumb questions. You just, you can, you can just ask, and often those dumb questions of things that the traditional, traditional industry doesn't really want to acknowledge. So for example, in the autos industry, five years ago, when we were set up an EV subscription business, we would ask people, How do people want to buy a car? And everyone would go, sit down, I'm going to tell you, it's a five step process. They want to come into, they want to come into a car dealership, then they want a cup of coffee, then their kids want to play in the play area, and then they know they don't. They want to buy online, you know. And so it's as a young person, you often sort of defer to authority or age or somebody with, you know, some gray in their hair. Actually don't respect no authority. You can
say it in a respectful way, but I think it's very dangerous to respect any orthodoxy. Joshua Feinberg: Yeah, so for sure, it's very possible that formal higher education is giving people the exact opposite of what they need to be successful in this space. One is because if they're talking about digital infrastructure at all, it's probably from 10 years ago. If they're talking about ESG, it's probably from five years ago. And most importantly, they're told that to never disagree with the professor yes or their employer, Mark Bjornsgaard: yes 100% and and if you you know it takes a lot, it's really difficult on the ego and the psycho when somebody tells you you're wrong all the time. I mean, as an entrepreneur, you're a lovely podcast. Stephen Bartlett, the
guy, one of the guys, set up Netflix, and it just says, There are no good ideas as an entrepreneur, there are only bad ideas. It's just how you how you shuffle yourself to a less, less bad idea and eventually to a good idea. So that's psychologically hard, and actually a lot of the time, everyone come everyone in the world of work now is we that we have to be entrepreneurs. The world's changing fast enough
that most of us have no choice. Joshua Feinberg: So that's fantastic. Fantastic advice for someone eventually in their their 20s, early stage of their career, on how to approach these opportunities and most of all, the appreciation of the multidisciplinary need to really be the renaissance person that's understands a lot about a lot of different areas to be effective and the importance of being able to motivate change. What about someone that you've known maybe earlier on in your career, worked at it with a different company 1015, years ago? You bump into them at a conference, and they hear about the great work that you're doing with deep green, but they're not quite as locked into what they're doing with their career. They may be feeling a little bit of sense of
sense of burnout. They've been through a lot of companies that have been through large layoffs or something like that, and they're looking to you for advice. And they say, Mark, I want to stay doing something with digital infrastructure. I'm really inspired to lead change around the sustainability this heat reuse movement seems like it's got a tremendous amount of value for on so many different levels. What would advice would you give to someone like that in a client facing kind of role that's looking for little mini career reboots? Mark Bjornsgaard: It's, it's, it's difficult. We have a we have a saying where we used to have a saying in our venture capital company, that when we're on, we're unplugging. We, I
guess, rather rudely, but we, I guess we didn't mean it like this, but corporate refugees, right? So I think, I think whenever we all get anchored where we are, we all get anchored where we are. So, so some sort of resilience and flexibility around mindset is, is always going to be important. It's very easy to say that, I think, and part of that is being when we only read our own when we're in our own filter bubble, and we only read what we want to hear, or what we what we're familiar with. I think it's it's increasingly difficult, isn't it, in this day and age, to actually read the other side of the of the other argument, whether that's in politics or wherever it is. I'm I still, yeah, it what we're doing is
certainly, certainly not a it's not received wisdom as we, as we know, by any stretched imagination. And I think, and when I would say, kind of when, if you've, if you've, if anyone's been in a large organization or been somewhere for a long time, then stepping into, it's often the case that stepping into a small organization is where you're going to find, where you're going to find a different way of doing something or or perhaps more social, perhaps more sort of anchoring on, on what we think of the Great Transition and what's coming in the in as the climate changes. So, so you're going to, if you, if you want to change, you're going to find yourself in a smaller organization. And I think that then has its own charms and and, you know, up, downsides and upsides. We're so give you a we start a senior leadership team meeting every Friday with with group therapy, so we all check in. Now you wouldn't, you probably wouldn't do that in a
large organization, but in a small organization, because you all need you all need to hear sometimes bad news, and you will need to, you will need to have a lot of faith and a lot of trust in each other, you need a lot of psychological safety. So so to any way, any, any, any way towards psychological safety is beneficial, and part of that we found over the years is is checking in emotionally, genuinely checking emotionally. I'm not doing so well this week because of x. So it's often in
what we would what you'd hope to be kind of relatively well run smaller businesses. There are very, very different ways of working that if you have been in a larger organization, or if you've been used to working in a certain way, I think a good small organizations are quite different. And I think a lot of people are quite surprised at how different that is. If that makes any sense, that's probably slightly rambling answer, but Joshua Feinberg: yeah, actually brings up a really interesting point from both the VC side as well as startups versus enterprises. I see a lot of times as companies go through a
successful funding round, advancing from seed to series A or series A to Series B. There's usually one or two people on the board that love the idea of hiring for logos, mainly because they love the idea of having logos on their home page trusted by and we think about what we just went through a couple weeks ago with CrowdStrike. I'm sure CrowdStrike had some really impressive, trusted buys on their homepage. Every company
gets motivated by a certain element of social proof. The challenge that I found with a lot of small companies is the sales hire, the marketing hire, the product hire that was at snowflake or Salesforce or Microsoft in the past five or 10 years was there when there was over. 1000s or 10s of 1000s of employees, and they're on playbook iteration number 137 they weren't there back in the early days when there was one person running marketing, doing 47 different jobs. And it's just such a different context shift, and it gives people a tremendous opportunity to make impact, but it's a very, very different kind of adjustment process for many people with versatility, I often use that, especially in professional sports. In the US, there's this concept of versatility among players. It's
even very, very well pronounced in baseball, but to a lesser degree and and a lot of other team based professional sports where they like the general manager, especially on a company, on a team that doesn't have a massive payroll, likes versatility. It's the same thing with startups. They love the idea that a salesperson can hop in and run something that may have been a product marketing function in another company, or that they're down an account executive and a marketing person can hop in and take some demo requests and things like that. So yeah, it's a very, very different kind of context, and I totally see that when it comes to mental health, when it comes to psychological safety, being a company, being culturally open enough to even have these conversations and vulnerability probably runs all over the spectrum as to whether it's accepted, encouraged or looked at as a sign of weakness. Mark Bjornsgaard: I Yeah, we, we are we really, I really encourage sort of truth to power. We are brutal with each
other. But then we most of the deep green management team I've worked with for more than 10 years. They're my my wife calls us the genius Donkey Sanctuary, because it's, it's, it's, it's, all of us have worked with each other in different businesses for more than, sometimes more than 10 years, but we all know each other. So there's a there's a respect, of course, there's a respect, but there's that there's a there's a brutal clarity that it's really useful when you know you're not you're you're communicating an idea in words of lesson, two syllables, but there's no ego in there. You're not being horrible. You're just trying to get to a truth as quickly as possible. I
love the book. The hard thing about hard things, I don't know whether I'm sure most, a lot of your audience will know it, but it's, it's Andreessen book about and it's just, it's so good, and it's all about hiring yourselves, your head of sales, and also that evolution through that I gave it to my, one of my good friend who's I've known for 15 years. He's now our CMO, and I gave it to him to read before he started it at deep green. And it says, how you how to fire your best mate you've got in this at this stage business, and then you have to, kind of, you have to eject them, because they can't grow into the next phase of the business. And he sort of but it is, it is it is a Yeah, sales particularly is. And the book has got a whole appendix about how you hire, how you hire sales specifically, it is in a startup. But so much for sales in a startup is because you
actually don't know what you're trying to sell to who the reality is. Lots of people are hiring a sales, somebody in sales and marketing, when the really, really the question they need to ask. Ask is, is, what am I trying to sell and to whom? Because once you really know that, once you really, really know that, then then it's yeah, that it's an easier game. Joshua Feinberg: Yeah, that's why the early sales hires, to the extent that they can wear the hat of product part time product part time product manager to understand the feature request that they've heard three or four times already that they're not closing deals because of they really need to push back to the CTO, the technical founder, the CEO the company, and say, hey, look, there's a million and a half in pipeline that's not closing because we're missing this. What
do we have to do to make the security to deliver this? Because they've told me not only are they going to jump on board, but they said they've, each of them told me they have two or three friends and similar companies. So yeah, definitely a very being able to cross, be multi, cross disciplinary, multifunctionary, across a full go to market function that the smaller the company, the more vital it is, the Mark Bjornsgaard: easier is, yeah, yeah, yeah, yeah, yes. I, yeah, I, yeah. It's, it's the role of sales. I was very keen to hire a salesman, a head of sales, CRO in deep green right beginning. And I'm glad we didn't, because I think, as you say, you learn. Well, I'm a big fan of that rework principle as well, which is, you learn your you learn your craft, and then you hand it over and and I think, I think, yeah, smaller teams learning their craft and then really knowing what they what they doing before they then hand over. And scale is part of
the joy of this phase of the business. Yeah, yeah. The Joshua Feinberg: challenge is, too, is that there's not wide agreement on what product market fit and what go to. Market Fit really looks like. There's actually a lot of disagreement among how it's actually interpreted. I always tell people that look at your case studies. Can you build an ideal
client profile off like that? Were they all over the place? Mark Bjornsgaard: Yes, yeah, yes, and yeah. And I mean, one of the joys, I guess, one of the joys of the data center industry, is that, having built software businesses and other businesses, at least in the data center industry, it's only power and and build cost. You know, it is a relatively at its most pragmatic. If you, if you, if you distill it down, it is, it is relatively. There are, there are fewer, but there are fewer levers to pull, I would say, than some businesses in our in our we, we think, I always like to think very simply about about anything and and in our world, it really is how much you can build a megawatt of a data center for, and how much is your power cost. And those are, of course, there's a million other
variables, but those really two, are the two levers that you've you've got to pull Joshua Feinberg: the data center go to market. Podcast is sponsored by dcsmi, elevating the role of sales, marketing and go to market GTM professionals in the growth of the data center industry. To stay informed about upcoming episodes. Make sure that you subscribe to the data center sales and marketing newsletter on the dcsmi website that can be found at www dcsmi.com Again, that URL is www dcsmi.com you you're spending a lot of time with leaders in digital infrastructure companies, advancing some ideas that to many may seem a little bit disruptive. You're getting, I'm sure, getting quite a bit of pushback. And beyond that, there's a lot of definitions as to what success looks like with sustainability and ESG and reuse and the whole idea of greenwashing, with all of that in mind, what do you think is the biggest mistake that you see being made nationally, throughout Europe, throughout globally, when it comes to how to approach these issues of understanding sustainability with the implications, with power, with water, with food, with carbon emissions. I know
you talk about these four ideas quite often. When you wrap that all together, what do you think the central root cause of all of it is? Mark Bjornsgaard: I think from a sustainability point of view. The mistake is, is assumed that anyone is cares as much as we might. And I think it's, it's really good when we can frame sustainability in the in the in the in the, not in the guys, but in around the fact that it's just a better solution. So we're, we are laser focused on that. You know, whatever we've
done in terms of drawing attention to deep green, dressing up as lifeguards, and showing all the social good that free heat can do we the reality is, deep green as a data center company just provides a better product than everyone else out there. Now you'd, of course, you'd expect me to say that, because I'm chief executive green, but, but just on every metric, it's better. So I, I guess the, I guess the mistakes we see a lot of are people assuming that in the nicest possible way, not that still, not that many people, or very, very few people are going to make a purchase decision in A, B to B situation based only on your sustainability credentials. It's it's got to be better. And so many, so many sustainability options probably are better, but they're not framing it somehow it's perceived to be dirty to frame it in economic terms. The reality is, if you use a deep green data center, it's just so much cheaper for you as a client and and your and you're getting the social benefit. But we can't. We couldn't possibly
justify doing what we're doing. I don't think I mean, we could justify it emotionally into ourselves, but we wouldn't be able to build a business if we were asking people to pay a premium for green now, regulation and where the world's going. That might all change down the line and and that that that might be well and good in three, 510, years time. But at the moment, the reality is, if you build data centers in a decentralized way, you're just rich. It's just faster, richer, better and and it's better for the environment.
Joshua Feinberg: So from a practical business model standpoint, do you see going forward because of proximity issues that you're going to see a lot of data center developers building health clubs with indoor swimming pools adjacent to the data center. Do you see them building office buildings that could be heated from the heat that's generated adjacent to the right next door to the data center, high rise apartment buildings? Do you see much of that going on? Do you anticipate more of that going on? Yeah? Mark Bjornsgaard: So we we don't see much of it going on now, but we imagine a future where absolutely that will go on. And if we take, we take this kind of notion of the trillion dollar cluster, this idea that we need 100 gigawatts of power to birth super intelligence, it's already exists. We've built it all. It's
just in the fabric of our communities and our towns and cities. So we, and we've got a we've got an event with Nokia in two weeks time where we're showing how you can network clusters of AI factories together to build superclusters. So we think of London as a 1.6 gigawatt data center, which just
hasn't been built yet. And then when you see, when you think, well, I need 100 gigawatts of power to power a supercluster, well, actually, you only need 70 cities in the Democratic in the democratic world to to to fit data centers, retrofit data centers to where heat is required. And then, not only can you take all the free electrons, which are currently from Cattell power, then you can land a whole load more solar than you otherwise would. And and then also, of course, brilliantly, you can provide billions of pounds of free heat into all of our communities and and towns and cities that that you know, many of our towns and cities desperately need that that money. So if we built, if we if we retrofitted all the district
heating systems, all the hospitals, all the swimming pools, all the distilleries, all the laundries in London or LA or New York, we would, we would be delivering, in the case of London, 600 million pounds worth of free heat back into our communities every year. So now this is a paradigm change. You're not anchored here again to our earlier conversation, right? We don't. No one thinks of the data center industry like that, like this, but how else are we going to do it? You know that the if you, if you reverse the question, you say, Well, if we need that much power, how else are we going to do it? Are we really going to build 100 nuclear reactors in one place to build a cluster that's going to defeat CCP and other bad actors in the race to artificial superintelligence. There's just no chance of that happening in the foreseeable future. The only
way you're going to do it, and again, you'd expect us to say it, but the only way you're going to do it is by, is by taking all of the compute that we can land, reusing the heat. And a brilliant benefit of that as well, of course, is that it gets all of our populations on side. Because instead of saying, What the hell are you doing? Our government, our government building all of this, this AI capacity, what is it for? You don't have to worry about that because you're heating people's homes and offices. So there's, yeah, there's, we see the world will emerge. How quickly that happens? Who knows? Who knows? What's Joshua Feinberg: super interesting for that also is when you think about the scarce grid capacity. And I've been
talking for six months or a year about what happens when, right now maybe 10% of your neighbors are driving EVs. What happens when 50% of your neighbors are driving EVs and you're on a collisions course? But if, all of a sudden, you don't need grid capacity anymore for heats, because essentially, these data centers aren't just ai ai factories, they're heat factories. It's almost like having a generator, power utility, generator substation next door to, adjacent to your residential, commercial entity that doesn't need to draw any power that's heating anymore. You're Mark Bjornsgaard: using the electron size, 97% the energy that goes into computer comes out as heat. And we capture 97% of that heat. So even direct to chip, you're capturing 80 85% of
it. So yeah, as you say, if you look at it, if you, if you frame the problem in this way, it becomes this wonderful solution. Now it's extremely hard for people to get their heads around that and and from a networking point of view, we're going to, we're going to show with Nokia in a couple of weeks that we this is technically possible. I saw this week. Meta said that night, they losing 10% of each training run for llama. They're losing on on hardware failures and GPU failures and and storage failures. The degradation you get from cluster, from from, from clustering decentralized sites, is way less than 10% so when you actually frame the problem of train of AI training runs really decentralized. Doing it in decentralized manner is
really not that it's not even that controversial, actually, in terms of, in terms of the degradation you'll get you're getting from from doing it. So, yeah, it's a, it's a, it's a fascinating, I find it a fascinating way to look at the world, and a lot of time, I think, particularly on your side of the pond, us. You know, we might think as us long launching Europeans, we might think of our US colleagues as you know, ESG really isn't on the agenda, but, but of course, that's not true in many, many areas, the US and and, and I think. Is also the assumption that doing doing heat reuse in the US isn't really worth it. But there's 9000 megawatts of district heating systems in the US already, and you guys haven't even begun. And
when you guys get going, you get shit done really quickly, you know, so, so, so there's a huge opportunity, even now, to kind of get ahead of this, to get ahead of this problem, and what a wonderful world it would be where, where we as an industry, instead of being the pariahs, instead of all these people throwing bottles at us at the moment, saying, you're using electrons, you know, we don't want you to be here. You're sucking up grid capacity, yeah, we could be the heroes. Joshua Feinberg: So there's already, there's because of where solar is in the 2020s there's already a lot of people that get the idea that you can put panels on your roof and you can store in batteries. And I know you've talked about the idea of homes even being able to have a trench down to the street, and transmission is how close are we? Is it decades away to the point where we have these transmission lines and storage capacity with the heat, where adjacencies, or is just, they're just such an enormous opportunity with locating heat key consumers adjacent to the data centers that have, even if we did that, it's a massive, oh Mark Bjornsgaard: yeah, even if we did that, it's, it's three we think it's, we reckon it's a 3.4 trillion retrofit opportunity just in Europe alone. So, yeah, we don't even need to think
about what's next. All we need to do is, is, is, do what we've got and and expand what we've got. Yeah, we there are three, 3% of the UK's heating is district heat, centralized heating sources. It will be about 15% in however, however
fast we can get there. So no, we can. Yeah, we can. We can. We don't actually need to do anything different. We just need to accelerate what we're doing. Yeah, Joshua Feinberg: and presumably, the trend that we're on now where a lot of hyper scale facilities are being built in secondary and tertiary markets, because that's where utility power is. The problem is, there's nothing adjacent to those facilities that can capitalize on that. And then you have the other side of the trend edge that's being built for dealing with mostly latency issues, and they're not just not going to generate quite enough heat for it to be that meaningful. So yeah, there's, there's there's a lot of
education that needs to go into these conversations started, Mark Bjornsgaard: yes and, and just from this kind of strategic point of view, I speak, we were speaking to different, different sort of military, those sort of representing different military and governments that the idea of decentralizing this critical infrastructure is, is, is a good thing. You know, we are. We're presenting a much larger attack surface in terms of what building this stuff, building it in a centralized manner, actually having it in the fabric of our towns and cities, distributing it is, if, even from that point of view, it's a good it's a good thing to do. And then, as you say, from in terms of the grid constraints. The reality is, even if you
believe that electrons, all of those, that 20% entire you, more of the entire US, electricity supply is going to be available in the next 10 years, even if you believe that, as you say, you just can't the grid, there are so many other blockers to it that the idea of looking at the through the other lens of the telescope is is hopefully starting to take hold. And I don't it's obviously not just us. Now, you know, if you've got Equinix who have announced that they're integrating with district heating systems, it seems to me, like every week now there's an announcement of a data center integrating into another district heating system. It's, it's becoming a norm, but it's, it's, it's much more normal than it used to Joshua Feinberg: be. I saw one of the channel publications I
forgot, which one is now tracking the hiring of Chief Sustainability officers and technology companies. Just the fact that there's interest in the media and events companies in carving out a publishing spot for that tells you that there's at least more awareness of this. Mark Bjornsgaard: We have a we have our B to B buying kind of matrix, classic three triangle, influencer, decision maker by we, we in. We always land best when we have been weirdly being bought in by the sustainability sort of the CSO. That seems to for obvious reasons, right? Given what we're doing in Europe. Of course, we have the benefit. Well, in the US as
well, we have a 20 a lot of people have got 2030 targets. And whereas even three years ago, you could say 2030 is, no, it's 2030 is now, oh god, it's my, you know, it's within the buying cycle of this hardware that I've got to worry about this. So I think in a in a good, bad way, whichever way you look at it, regulation and commitments and targets, the corporate machines are starting to, kind of to churn in a good way, and corporates do whatever. Bad you might think about them. What they do is they action the plans that they've got in place, that that's, that's how corporates work, right? They, they stick to the plan. So I think we can be quite heartened
by that, that, yeah, I think, I think things will change. Hopefully the speed of change we've seen has, has accelerated, as you say it. Hopefully it increases Joshua Feinberg: given all of this change and just how multidisciplinary the nature of what you're doing is, do you have any favorite resources that you use to stay up to date on what's going on in this industry? Any favorite podcasts, any favorite YouTubers or similar kinds of outlets that you depend on to truly get a pulse for what's going on in this very vast industry, Mark Bjornsgaard: in all of the above, I think, to that discipline of reading outside of what you do. So for, since I was, since I was almost my first sort of years in work, I would, I've had a Feedly, I've, read, sort of probably, I probably read three, 400 blogs a day, but I do it in such a way that I've trained myself to flick through my RSS reader, and I kind of trust my brain to read it before I'm conscious of it. So I, I
mean, all of the usual kind of suspects in the industry are great and and the wonderful kind of podcasts, both this your side and our side of the pond. But also, I think there's, there's, there's, you know, energy and battery tech and energy tech podcasts, I think are really, really important. And then actually, the wider, the wider ecosystem of technology, again, this kind of software layers, if you like, I think there's there's there's, you can't be too wide, if you like, when you're when you're when you're approaching it, when you when you're approaching this. The industry we have. I there are certain, I have to say, I do love the Stephen Bartlett podcast. He's Diary of a CEO. Is brilliant, not necessarily, not necessarily the specifics of our industry, but just in terms of kind of the mindset of running a business. There's, there's a lot
of health on there, but there's also a lot of, there's a lot of business and and, you know, entrepreneurial and sort of next stage business. So whether that's the Chief Exec of Amazon or, you know, a boss, a famous basketball coach, or whoever it is, there's learnings are often, often applied, right, aren't they? It's often, often the great in one sector. You can, you can very easily port their learnings and their thinking into into what we're doing. Joshua Feinberg: Yeah, that's truly fascinating with how many discrete pieces of information you're consuming on a daily basis, to synthesize all of this and figure out the pulse of where this is going, in addition to, I think, even more valuable part of having dozens of conversations every month with stakeholders to get a feel for how they're approaching this on a local I Mark Bjornsgaard: test myself. So I if I've got, if I've got 400 to read, I'll flick through. I'll flick through, and if I notice something, I'll stop. And then I have gone back over those
400 and thought, right, I'm going to read them and see what I've missed, what I would have tagged that otherwise. And it's amazing, but 20 years of your brain just being kind of, I mean, God knows what it's done to my god knows what it's done to distraction. Or, actually, I'm not sure it has actually affected the ability to kind of, actually focus on on one but, yeah, breadth, I think, is really, really important. Yeah, that's Joshua Feinberg: the premise of all these feed based platforms now, where like YouTube and LinkedIn. 10 years ago, in order to see content consistently from the same individuals or companies, you had to follow or subscribe or show some signs of engagement. They've all gotten smart enough now to know that if
you stop and you're hovering over a post for 3040, 50 seconds, they that's a micro engagement, and they know they're going to keep showing you that creator over and over again to increase your time on a platform. So yeah, and then there's this whole phenomenon the last couple of years, the Tick Tock vacation of the world, where everything's getting broken down into 59 seconds or less. Mark Bjornsgaard: And that is brutal, and we're all, we all suffer from that. I've just finished a brilliant book called The anxious generation, which is, which is, Jonathan hate, talking about, talking about the the effect of social media, and particularly on our kids brains. I think there is a, yeah, there's, I think, I think I know I'm putting you in my age bracket. Joshua, but, but, but, I think we perhaps we were
fortunate. We were fortunate enough to live in a world before, before, perhaps some of our ability to keep, to keep concentrating more than 10 minutes, but, yeah, no, no doubt we are. Our brains are being rewired. That's that's for sure. Yeah, Joshua Feinberg: it looks like even LinkedIn has very recently succumbed. Into this, because in addition to supporting the short
form videos, it's not that front and center, but there's a way in which it senses that you're watching these shorts and opens up something that almost looks like the YouTube mobile app experience already. So coming to a B to B, professional network platform near you in the coming weeks Mark Bjornsgaard: and months, yeah, yeah. I mean, I am guilty now of putting a longer wired or some other sort of lengthy but I want putting it to chat GPT and saying condense, I have to say I am guilty of that. You know, spitting out, I did that on the trillion dollar the the that white paper that Leopold wrote about situational awareness. It was called a brilliant white
paper, but it was very long, so I put it into chatgpt, and that part of the utility is, yeah, part of the utility is that distillation I put perhaps might. I'm sure I've missed some of it in the non reading of it, but you get the bones of it Joshua Feinberg: absolutely. And I think for anyone that needs to be able to advance ideas that are anything but a commodity, where they're many times disruptive solutions, I talk to a lot of Gen X founders and sales executives who've been into the space for a while, and they don't understand why it's not as easy to get attention of prospects as it was 10 years ago, dude. At the risk of sounding really meta, what we're
doing is how you need to show up. They need to see you as an expert and not just someone that's focused on the transaction. That's the key unlock. You don't necessarily need to go immediately change your job title on LinkedIn, your email signature business card, but it's not doing you any favors, no.
Mark Bjornsgaard: And Stephen Bartlett, actually, he wrote his book, which I have to say, I don't know why I wasn't it's really good book, but, but that disrupting. We did that at Deep Green, more by luck than judgment. But it turns out, instead of talking about edge data centers, about the quality of the immersion tank, if you, if you dress up as a lifeguard, and then you, you build a supercomputer in a swimming pool and show how much that that story is such a cross. It's cross. It cuts across so people's perception of what's normal, of what they're seeing in their LinkedIn feed, that that that disruptive, that that was the thing, that that sort of elevated it into kind of a kind of a crossover story, if you like, yeah, it was, it was just something that most people in the data center industry just weren't used to reading. So the unfamiliar got, got the attention. Yeah, now I was
Joshua Feinberg: gonna say, even when I saw that video, the idea of the the child recounting how they missed the idea of the ritual of having soup and hot chocolate after they went for a swim. I didn't even know that was a thing, Mark Bjornsgaard: no, and that was not again, not made. That was Matt our CMO, who, who framed that like that, but yeah, who, who was framed a data center company in that, in that childhood story of hanging out with your dad in the 70s, who probably didn't really know to hang out with, you know, wasn't, wasn't particularly emotionally aware dad in the 70s, and and linking that story into a data, into story, suddenly is, is, is the unlock? Yeah, it's Yeah, telling, I guess just telling a more human story. I think that's what we we haven't talked a lot about that, but I think that's a lot of particularly our industry, because it's so technical, and we love talking about pues and immersion tanks, but not so much for outside Jesus, yeah, Joshua Feinberg: yeah. I know the big data center operators for sure in the US get this because they invest quite a bit on community outreach. They do community fairs. I know one sponsoring a concert series so that they get that. They need to
spend a lot of time, not just with the lobbyists and government regulators, but it's definitely influencing how they're they're the taxpayers, and they're actually see their the value of their company. But Mark Bjornsgaard: a lot of that is badging. It's not, it's not deeply integrated into the product. You know to mean it's, it's, it's, it's the belief that I can somehow, I can just somehow badge something, and it will be all right. I think, I
think we're all in, I guess it's perhaps case of through Crosby to be marketing. But I talk a lot about we mentioned a couple of times in other scenarios where, where I think it was Google that was getting a hard time about building a data center in the UK, and that picture in the Daily Mail, one of our one of our newspapers, was the data center, and then in front of it was the housing estate. And as we said before, it would have been such a small, relatively small lift to dig and provide that housing estate with, with free or subsidized heat, to get that to get that whole community online, you know, mean, that was that, yes, they could go and they could, they could sponsor the local football team, and that would be great. But there's something that you can come. Go beyond just cons and into kind of product, if you like, just mean, you can actually, you can actually integrate yourself. I think now, again, you'd expect me to say it, but yeah, Joshua Feinberg: and that's the challenge is it's really understanding who the different stakeholders are, understanding what's top of mind for them, and segmenting, contextualizing, personalizing it in a way that people understand what truly is in it for them, and so much of it is that classic of the job to be done and figuring out the different value, the value canvas, and all different other ways of explaining the essentially what's in it for me, yeah, Mark Bjornsgaard: courage, and it's courage, right? I mean, we, we didn't have much to lose, so it was but, but it would be very You need courage to do something different. I think you know, particularly as when you when you've got a large, established business. But I think again, what our marketing taught us is
that no one makes a judgment about the quality of the data center you might build just because you framed it as doing social good. I think a lot of people would think our B to B, it's got to be very serious and boring and numbers and rational, but I don't think that's the world we're in anymore. Joshua Feinberg: Reminds me a little of this book and movie that was really popular in the US and early 2000s Moneyball, and it talks about professional stories. It Mark Bjornsgaard: was my first business. I built a business called artist first, and we, we built music metrics. So we took data from artists in the music business, and we told record companies whether they would be successful. I'm, I'm a, I'm a
massive Moneyball fan. Yeah, we and sabermetrics. We called it museum metrics, after sav matches and my set one of my favorite movies. I I'm Yeah, I Michael Lewis is a is is a is a God, in my opinion, yeah, yeah. I just Joshua Feinberg: literally saw our hometown baseball team go through this in the last six or eight weeks, where they hired away someone who looks like depodesta, who looks like the Jonah Hill character in that movie to run the Miami Marlins. They literally traded away 10s of millions of dollars of payroll to try to find the undervalued rookies on a team that's actually sustainable and affordable for a team that hasn't traditionally been able to compete with the massive payrolls from the teams in Boston and New York and LA. We Mark Bjornsgaard: were in my first business. We did it. We're
doing this 2023, years ago. And we and to go right back to the beginning of our conversation and talk about people being anchored where they are. So we would we walked into the ministry of sound, which was one of the coolest record labels in this country, right? We walked into the ministry stand and we said, we've got all this data, and we can show you which of your artists is going to be successful or not. We were literally, physically ejected from the building. It's like it was, I mean, to say that that was not you're telling me, you can tell me better than my ears can tell me. No. So yeah, it's,
it's, it's, well, when we're Yeah, we are right back to the beginning of the conversation. It's, it's, it's the data sometimes tells you stuff that you either don't want to know or can't hear. Yeah. Joshua Feinberg: It's like that silly scene in Bohemian Rhapsody where the he ends up on the street and he throws the rock through the window and said, You're going to be forever known as the producer who passed on Queen, Mark Bjornsgaard: yeah, yeah. You just, so you just, you just don't Joshua Feinberg: recognition, yeah, telling evaluator, Mark Bjornsgaard: yeah, exactly, yeah. Mark, Joshua Feinberg: this has been terrifically fascinating at the intersection of digital infrastructure, energy storage policy around this, reframing, understanding the benefits, evangelizing, getting a movement started to see the value of this and unseating kind of incumbent ideas that feel that they have this all figured out. There's so much work to do, but it's so
much promise and so important. No, Mark Bjornsgaard: well, thank you for having me on Yeah, no, we're happy to be part. We have to be part of the movement, and it is certainly not just us. There's lots of people talking
about this now, so no, we're happy to be a part of it, and we're loving it. So thank you for thanking Thank you for having me. Joshua Feinberg: Mark, what's the best way for someone to keep up with what you're working on or connect with you? Is LinkedIn a good channel LinkedIn Mark Bjornsgaard: is LinkedIn. Is my channel of choice? Yes. So find me on LinkedIn, please. Yeah, connect with me on LinkedIn, and then we can connect you into or complete the contact form on the deep green website. If you Google deep green, you should find us and and on that form, if whether you're a host, so whether you're a district heating system or another off taker, you know somewhere where we can build a data center, or you want to use the data that will get filtered to the right teams.
Joshua Feinberg: So when you say host, it's a host of heating, as opposed to, like a colocation company, hosting. Yes, sorry, Mark Bjornsgaard: yes. I nomenclature. We divide between demand and host. So we, when we say hosts, we mean somebody where we meet, that place where we can build a data center, where we can offtake the heat. Yeah. Joshua Feinberg: Awesome. So. Much work to do, but so
important and there's a lot of urgency. Mark Bjornsgaard: Thank you so much for having me keep Joshua Feinberg: up the great work. Thanks, Mark. Really appreciate it. The data center go to market podcast is sponsored by dcsmi, elevating the role of sales, marketing and go to market GTM professionals in the growth of the data center industry. To stay informed about upcoming episodes. Make sure that you subscribe to the data center sales and marketing newsletter on the dcsmi website that can be found at www dcsmi.com Again, that URL is www dcsmi.com you
2024-11-05 10:59