 Hello everyone and welcome to JSA TV and JSA Europe, the newsroom for telecom and data center professionals across the continent. I'm Jean-Marc Lim and on behalf of the team here at JSA, thank you for tuning in to the third and final part of our end of the year special broadcasting series. Today, we will explore how private equity backed platforms, Greenfield and Brownfield developments, real estate and M&A will shape Europe's digital infrastructure deployments in 2024. And we're also delving to how the demand for AI could flip the table on investment portfolios. But with that out of the way, it is now my pleasure to introduce to you our exceptional executive lineup, which includes Sika Buzma, managing director for TMTMEA at Ingbank, Sebastian Dulley, senior fund manager for principal asset management, and Dr. Attif Ansar, executive chairman for Foresight Works. Guys, thanks so much for joining me. Before we jump into this very hot topic, which is AI and investment and what's going to happen in 2024, because every year we like a good prediction, let's just do a very quick around the clock kind of 30 second elevator pitch so you can explain what you do and what the company, what the business does as well. So maybe I'll start with you, Sebastian, going from top to bottom. Sure. So I'm at principal asset management. We're a global investment company and have around 100 billion in investments in real estate. I'm part of the data center team. So we've got a team investing in the US and more recently in Europe, we've done around 2.75 billion of equity committed to data centers over the last few years and over a billion in debt and we're sort of continuing to build our platform. Small amount. Attif. Thank you, Drams. Pleasure to be here. So my name is Attif Ansar. I work two heads. I'm an academic at the University of Oxford focusing on major and mega projects and I co-founded Foresight Works about five years ago. We are an AI company that supports projects in their construction phase to get built on time and I'll look forward to telling you more about that. We specifically work a lot with digital infrastructure, so both data centers as well as semiconductor parts. It is very interesting, especially with the mega side of things where things are going. It's quite interesting to hear your perspective on the market and what's going on. What will have to actually change to deliver those mega projects? And Siko, in bank big name, we know you're pretty much everywhere and everywhere. So Siko Bomsma, I'm an MD in ING's TMT team. Within TMT we focus on digital infrastructure which is fiber, data centers, mobile towers and satellites. I focus on the data center side of things. We have a very wide geographic mandate, so I cover wider EMEA. We have a team in New York and a team in Singapore that can service our clients. We can do various structures on a senior debt basis ranging from single-tenant greenfield to multi-tenant projects to portfolio-based financing structures. What we try to do is support our clients when they start with an idea and work together to bring that idea to fruition. Interesting. Bringing ideas to fruition is what this industry is all about, both for the industry and for the rest of the world, because the rest of the world is built on what this industry does with all these real states being deployed across the continent. I guess the first question, the question in order really is what do you think would be the main trends around investment next year? And maybe we can also push more the private equity button here because that's kind of the buzzword everyone has at the moment. Who wants to start talking about what the trend for 2024 will be? Do you want to go around the clock? Maybe, Sebastian, you want to start? Sure. So as you said at the beginning, AI is a very big buzzword at the moment. I know you say you want to focus on private equity. I look at things much more from a real estate perspective, so I think I'll stick on that. In Europe, we've been, from our perspective, we've fully raised our dedicated funds, so our focus is on executing our business plan around that, which is aggregating together a portfolio of standing assets. And I think there is sort of changes coming through when we look at that from an AI perspective, especially when we look at where that demand is going to go. To be perfectly honest, not many people seem to have a very clear answer to that. There's a lot of different theories going around, but what is clear is it's not so location-sensitive as, say, cloud deployments. You don't need to be in the availability zones. There's a lot of talk about the Nordics coming up in popularity, although whilst there's a lot of power generation there, could you end up compromising some of the sort of power infrastructure with the 250 megawatt schemes that are supposedly coming to fruition up there or being climbed up there? But again, that leads to, from our perspective, maybe some of the more onto the network and connectivity sites in the Nordics could become very interesting to look into. Just as an example, I think whilst AI is a big buzzword, it's definitely not the only trend that's going on. I think we're still seeing strong growth coming through on the cloud side. That's more of a focus for our US team. We've been partnering with a lot of different operators to the joint venture, and we're continuing to roll that out. At the moment, we're looking to pull together around sort of three to four billion US dollar program to support the existing portfolio. It's just suddenly had so much demand come through from the cloud sort of scale-to-type groups. They've had to, yeah, frontload a lot of their capex that they weren't expecting to originally. That's really the huge amount of extra capital in there. In Europe, I think we still, the demand hasn't quite come through. We still see the edge state as being a very interesting layer in the go forwards. Especially again, AI, I think, will indirectly impact the demand for that as models become, our apps get built. We'll really run on that, and then how is that going to be integrated into the world of individuals, work life, and personal life as well. I think as AI, it's definitely the best word that from our perspective, it's not the only interesting thing around. I mean, you mentioned edge computing that is very intrinsically connected to AI and what the objectives are. But then the question still comes down to what is edge computing? I don't think we want to get into that discussion today because otherwise we'll be here. Just talking about what edge computing, but that is interesting. You mentioned the Nordics, and we're going to locations a bit further down the discussion, but the Nordics are really coming through a second wave now. We've seen the first wave in the mid-2010s, but now it's starting all over again, and we do have some rooms of projects with 1 gigawatt, 2 gigawatt campuses, which I mean, we'll see if they come to fruition or not, but that's just ginormous amounts of power needed to do that. And I guess the geographical spread of these facilities that you were mentioning is shifting. The ideas are shifting on how we're going to build all this and disperse the infrastructure. But Attif, onto you. What's your vision? So let's look at some data from the US, and I think European data kind of mirrors that as well. So just in terms of green field, new projects being built, we were building under 500 megawatts worth of data centers in the US in 2020. Now numbers vary, but depending on the numbers you look at, anywhere between 3,500 to 5,000 megawatts worth of data centers are under construction in the US as we speak. And again, depending on the estimates, because the number don't break down that easily, between 1,200 to 1,800 megawatts of those are AI data centers, and the remaining are sort of hyperspace scale and COLO data centers. So I think kind of going back to your question about what are the trends, the first trend is hypergrowth, particularly for digital infrastructure. So that's not just data centers, the fiber that goes along with it. And if you sort of cast your net more broadly, with the semiconductor fabs and semiconductor plays as well, ultimately these data centers have to run on chips. So there's just a lot going on in this space altogether. Now, your second part of your question, John, was what's the role of private equity here? If you think about on average, per megawatt installed capacity for data center is around $10 million or euros, however you want to look at it. So with 5,000 megawatts being built, that's a $50 billion financing requirement. It's just staggering. And private equity has a massive role to play in this. So funds like digital bridge or variety of them have taken big best, almost specialized digital infrastructure funds out there, like Devonshire, Armour Fidelity, et cetera. And I think part of that is private equity can only really take on so much of that portion. So even with lots of leverage, I think both banks and private equity fund would be plenty busy trying to finance this wave. Yeah, it's just a second. I mean, although those numbers that you mentioned, they are huge compared to what we were talking about five, eight years ago, where $100 million, $200 million was a huge amount. If anyone touched the $1 billion, it was headline news all over the world. Those numbers are quite interesting. And I mean, digital bridge predicts $1.2 trillion of investment into data centers alone over the next five, seven years, something like that. It's potentially going to get over that anyway, because another beauty of the issue is that all the predictions that we make, they're always wrong because it's always more than what predicts will be. It's wrong in a good way because it's always more. But Siko. It's interesting to hear what's being said in relation to the growth profile, right? When we started looking at this as an asset class that was back in 2006, 2007, 100 kilowatt order profile was still something that was a significant order profile. And you were putting your financial walls together on the back of multiple of those contracts. Now, people are mentioning megawatts like the kilowatts back then. And I think that that's a reflection of how this industry has grown. What you see is the underlying fundamentals, they are still strong. You see enterprises that are still looking to decrease their total cost of IT expenditure doing their cloud-based shifts. I think they're becoming more proficient in selecting what type of cloud solution they want to have, be it private cloud or public cloud or a hybrid cloud solution. The development of the telecommunication networks is a very strong driver as well, right? Is that what you see now with certain operators that are actively offering to data center fiber connectivity and creating that ecosystem allows the data center assets to deal with this, computes this storage and this communication flow. And I think that as a consumer, both in the enterprise as well as on the residential side, we see that there is a tremendous evolution in terms of how we use contents and how we use applications. So where in the past we were okay, waiting a bit for mediocre content, now everything has to be content rich in real time. That is a very significant shift that we see in the market and that has its implication from a demand-driver perspective. Underlying that, we see that the hyperscaler demand profile is continuing. As I just said, a few years ago it was in the US 500 megawatts. Now we're 5,000 megawatts is a 10-fold increase and we see that as well in the amount of transaction flow that we have to deal with within our team. And especially if you look at hyperscaler strategies, we see so much evolution there which I think is a reflection of how this market as an asset class is developing. We see them shifting from their initial Tier 1 location into Tier 2, 3 and 4 location. We see, I don't want to mention Edge, but we actually see cloud out movements. We also see enterprise in movements where enterprises want to have their presence on a local basis, but they want to have their compute on a centralized basis. And that already in itself is a significant growth profile. But if you augment that with things that are happening with AI, the hyperscalers need to invest very, very significantly in order to do all the machine learning. So with all that, we expect that there is going to be a tremendous wall of capital that is going to be required. Unfortunately, there are a lot of banks and funds and institutional lenders out there that want to support that along with the traditional infrastructure based private equity players that actually see this as a very promising asset class. Interesting. There's a few things I want to follow up on there. I'll do a very quick question because actually you mentioned that you're seeing this on your site as well in terms of deal overflow. So actually a very quick sideline question. How are you guys managing within your own teams to answer the demands when it comes to actually dealing with the deal overflow? Like how are you dealing with all these demands, with the deals coming through, with all the dual deligency? Are you finding the right talent? Are you struggling to find talent to work within your businesses on the back of all this rush of business coming in? Again, it's a great problem to have, but it's important. Very quick one just on the sideline. Yeah, we work hard. Kitting aside, what we have done is basically we started looking at this as an asset class very, very early on. And I think that allowed us also to work with the juniors and the medias in our team to actually touch a data center transaction very early on. So it's a point of pride for me that everybody within our team speaks data center for a bit and that already saves a lot. If you let's say have the basic tenets of financial structuring, the basic tenets of how a data center works, if you know how to do a contractual review, you can do your instructions towards legal counsel in the right way, towards technical advisors in the right way. It just saves a lot of work if you have trained your team correctly. But yeah, it's quite a lot of work right now. Okay, I'm going to add something around building up the teams because everyone is building. So there's the common ground here. Just to come in here very briefly, so I'll come in from an execution perspective. So those megawatts of data centers have to be built by a team of some kind. And with between seven to 10x growth over last four years, supply chains are not seven or 10x bigger or better. If anything, they are worse off because of COVID in between. And so I think what we are seeing as well is there's serious capacity constraints in the execution side of the industry. And a metric for that is simply time delays. So projects have in general issues with time overruns, but partly owned to really lethal cocktail of COVID, geopolitical tensions, supply chain disruptions, those leak times are prolonging an overall kind of estimation error between when people want to build those data centers and where they actually get built is getting harder. I think that's also exciting because you can now use AI tools to try and win back some of that productivity. So I do think it's both a push and pull. And certainly for data center industry, because they are so keen on building AI data centers, I think they might as well use them in their own execution phase. So that will be a helpful message for them. It's the same old story that when a challenge occurs, there's an opportunity to do some things in a different way in a more efficient way. And I think that's where AI will enable is just becoming more efficient, faster, more efficient. And when you struggle, you always find a solution. And this industry is one of those industries that always does that. And Sebastian, just quickly on that one. I think on the senior end of the team, we started looking really 2019, 2020. And I think we're slightly lucky in that we were able to make a very good hire there for the big rush of a lot of investment firms starting to look into it. And again, we're working since 2007 in the US and having good knowledge there helped a lot. On the junior end, I think exactly the same as what SICO said, every time we go out for a site visit, we try to make sure that the analysts can come along, have a look around and really get to understand the asset class better. Ultimately, for what we do, we don't need a massive team, we're quite targeted, and we can have a smaller team as a result of that. There's nothing better than boots on the ground to actually see the state. One, another satellite question, I promise I'll get back on the horse after this one. Are we worried about the microeconomic environment across the world and how they might impact the European data center market next year? Is there an issue for concern or is just a passing by feeling that the industry will navigate? I think it's obviously a concern, especially given a lot of the geopolitical tension that's going on. But I think the demand driver behind data centers is something that we think can withstand quite a lot. And to be honest, I think a lot of operators are being slightly more squeezed at the moment with debt costs coming up, with inflation and construction costs. Over the last few years, but from our perspective, that adds and leads to a lot of opportunity. People are really looking at their capital stack. It's not as simple as equity from a private equity platform and then loan from IMG in order to be able to get the numbers to work. People are now starting to look at joint ventures. How can they be created and bring in additional capital? So I think it's leading to a lot of opportunity for us. It is very interesting because actually the west of Europe, North America, was very ahead of this whole construction. But when you look at Asia, it's all being done through joint ventures. So now it seems like we're actually taking a bit of the idea of joint ventures from Asia and bringing them into Europe and into the US as well. So it's very interesting how the market is shifting. I'll go around again. So, Attif, what's your micro economy for you? I'm bullish. I think looking at inflation numbers coming out of the US, employment numbers, again, I think the US is in a better, more robust shape than European continent as a whole, at least just the way I assess it at the moment. So I don't, I mean, it's impossible to predict the future, but I don't think the macroeconomic headwinds are the key constraint. The constraint in my view is a bit more micro, which is power constraints, chip constraints, people constraints. I think those are bigger worries in my mind for the data center industry, rather than a sharp demand just disappearing. And we see them, the prices, the price of the data center services have doubled in the last three months. So that all of that is pointing to supply constraint. So I think for the next 12 to 18 months supply just has to catch up with demand. What happens after that? No one can really know. Yeah, we're not there trying to guess. Listen, I agree with that. I agree with that. You know, the thing is, does a behavioral pattern change if there is a geopolitical thing happening? I don't believe that people will stop using the Azure platform if there is something going on in the region. But I think geopolitical tension leads to increased caution. And I believe that leads to some additional vetting of business plans in a way that they're able to withstand a certain systemic geopolitical shock. But that all being said, we see that the large hyperscalers are rolling out their networks across various markets. And that is still ongoing. That is not stopping. And clearly, as Atif referred to earlier, there are some regional elements that need to be taken into consideration. Can you have sustainable growth in a country such as Ireland or in the Frankfurt area? And we still see there that even though there might be certain political aspects that the industry, it's an engineer-led, creative, fast-thinking industry is still able to come up with solutions. Take Frankfurt as an example. The market has been very much overfocused on the definition of the size of the availability zone in Frankfurt. But what we see now is that that availability zone is widening up because the strength of the telco networks allow that. You don't need to be in a specific neighborhood in Frankfurt anymore. You can be in the surrounding area because that's where you still have your land availability and your power availability. We see that operators in unison in discussions with hyperscalers as well as with large deployment enterprises, they become more tailored towards where the demand needs to be and where the most efficient demand should be. Amsterdam, communication-driven market, we know that in the vicinity of Amsterdam there are certain locations where you can do more, that's a latency insensitive compute and storage. That's how the market is evolving as well. No, I just agree with what everyone said really and I do like the view of moving the conversation from micro-economics to micro-economics, micro to micro, sorry, I got my wording wrong there. Again, it's a fascinating market that adapts very quickly. The problem is when policy gets in the way and I guess that's where the interest needs to become a little bit better, get a seat at the table more early on as opposed to waiting to see what's going to happen. We've seen Germany now rolling out probably the most stringent sustainability legislation towards data centers and that's just going to be rolled out to other countries, so there's no way around that we know what's going to happen, so it's a matter of people preparing for that. That's where we need to move on now. It's not trying to cry over spilled milk, as we usually say back home, it's just acting on it and then building the platform for what's coming next. Let's bring the conversation back to the original plan. Let's talk a bit about just brownfield versus greenfield development. Even yesterday I was talking to CEO whose strategy is to acquire businesses and start on the brownfield side and then they will expand the facilities that they are acquiring and then eventually they will move into greenfield once the books allow. What are you guys seeing in terms of brownfield versus brownfield? What strategies are being followed? Is this a common practice, the one I just described for example? Just give us your view of what's happened and how they can change shifts stagnate next year. The first one they say something and goes first. And this one partly because I track greenfield projects, so I don't have a strong view on the brownfield side of things, so I'll defer to colleagues to discuss brownfield. I think on greenfield one of the interesting things that we're seeing particularly again with data centers is this issue of time delays. So if you look at the overall risk profile, the three things you're playing with, your potential revenue that you're going to get once you lease up the whole property, the cost you're going to invest in the capex in the time it's going to take you. Incidentally, cost overruns which are typically a very big issue in the execution phase for new projects is less of a problem for data centers. On average they're relatively low cost overruns partly because of very big contingencies, but time overruns a big issue. So on average we've built the world's largest data set of its kind on data centers. We're seeing 34% average delays. So roughly speaking a two-year project becomes a three-year project. Up until now the hyperscalers have taken a relatively benign view, so they sort of talk a tough talk around penalties, but when the color developers have been unable to meet those deadlines, they've ended up in negotiated settlements and those penalties haven't been extracted. I think that patience is running out. So we're going to see a lot more of these penalties being extracted and a lot more rankling over ready-for-service dates. Partly because the demand imperative is so strong that there isn't that much of a leeway there. So I think from that perspective just without knowing too much about Brownfield myself and just looking at Greenfield, many private equity investors may take the view that where the demand profile is already known they may get away with it, but I think my bigger message to the audience would be take the execution risk whether it's Greenfield or Brownfield extremely seriously. Oftentimes people think become fixated with Excel spreadsheets at the final investment decision and don't bake in any risk for what happens if I'm one-third overrun and my customer actually extracts the penalties, suddenly the private equity IRR is going to disappear. Interesting point here. Spastin? I think it's also a question of span of control. No. As hyperscalers you only have a team that is a certain size and you are dealing with the demand profile for both AI as well as the cloud-based business where you have a team that is working across multiple projects and they need to retain a specific amount of control. If you're working on one-two projects you can still get internal resources to help out in order to meet the RFS. If you manage 20 projects you need to be a little bit more stringent and you need to hold the project company responsible for the agreements that have been made. I think everybody wants to be pragmatic in this industry and they also acknowledge that they need to make it work because they also need to have a 10 or 15-year productive relationship with the project company. But if it really becomes an issue then at a certain point in time you'll see that these things start popping up as in RFS penalties being invoked or additional payment holidays starting after a project commissioning. Now we look a lot at brownfield situations and we support operators that have landed in let's say the tier one markets and have managed to secure good location for certain activities, certain compute mission critical, lazy sensitive compute and communication at data center functionality. Those locations they very quickly transform from a green field situation into a brownfield situation and we have seen over the course of six, seven years in project one evolving into project six because that order profile is just continuing. What you see there is that with such a demand profile and with such a multiplication of projects this is great for the investors because they do have some scale advantages. So it has some pretty beneficial effect on their IRR and investment strategy. Now the hyperscalers also know that and with the environment changing with weighted average cost of that increasing and this is a depressed supply chain they want to be pragmatic but they also know that if they have a repeat order profile in brownfield that you know as the Roman says agreements are there to be respected. Interesting and Sebastian any thoughts to add? I know it's more on the real estate side but oh all right we lost Sebastian there for a second. As you guys were actually talking I was thinking because oh there you are. I was saying if you've got any thoughts to add I know it's more of a real estate kind of question but maybe you have just one or two things to add. John I'm the brownfield versus greenfield. I think apologies for dropping off. From our perspective we definitely have a preference for brownfield and I think you do get more challenges and you do have risks of overruns from taking on existing and the challenge of not everything is exactly how you expect whereas with greenfield is a little bit more of a clean slate but yeah numerous reasons I think brownfield is preferred and I think coming forward is sustainability as well. I mean just looking at sort of the ground of a site. Some people are saying up to 50% of carbon created in bodies during development processes within and especially somewhere like the Netherlands where a lot of regions need a lot of piling due to the ground quality and yeah a lot of the embodied carbon goes into just the concrete base and so if you're able to to sort of continue using what's already there then that's got to have a benefit whilst embodied carbon is not overly tracked at the moment down the line I think that's going to become more important so again. It's about like life cycling the circle economy around the entire project as well which is why the attention is shifting but as you were all talking I was actually remembering after Covid the few months after Covid kind of ended. There was a lot of discussion around do we have oversupply? Did we build too much? Was Microsoft too rushed to put those 100MW outs and the truth is it's not. I mean here we are just a year and a half later and we're talking about we don't even have enough as it is to let alone talk about oversupply. That's how fast the market changes. I'm actually going to bring one question because we do have a question from the audience and I think it's quite quite a good point around modularization building modular units within the data center. So what's your view on using modularity to build the facilities and how can it kind of help with the 30% average delays in the delivery of these projects? Are we moving towards that side of things in a wider picture or is this for a very specific Actually modularization is a phenomenal idea you know and wherever we've seen industries modularize we see those overruns come right down and in fact the absolute cost curves come right down so solar wind, onshore wind doesn't have cost overruns or time overruns same with solar etc. So I think you know in theory really good ideas. For some reason I personally haven't come across lots of scalable deployments of modularization for data centers. So I'd be curious to be educated here by the audience about projects that do use a modular approach and I have managed to actually replicate that. There's obviously always edge kind of case studies that people talk about and but it seems to me more of a marketing ploy than reality. Although there's really in my view no barrier to it. We've seen for example companies like DuPont do this for chemical plants have achieved very high degree of modularization, very similar complexity of you know electromechanical systems. So it's doable but I'd love to learn more about who's actually doing it and doing it at scale and it's not like a one-off project that's something repeatable. I think that that is exactly in line with how we see it. You know we have seen modularization in the industry and it's just like edge it's an often used concept. If you look at all the info memos that we have if you just would do a search function there is always modularization in there. The fact of the matter is we haven't seen it yet in scale in large deployments and the reason is for that is twofold. I think there is a very vast difference between you know standardization that we have seen initiated you know back when we were talking about the Yehu chicken coop design and what Google has been doing versus modularization that's one and the second thing is is there are specific business models where modularization is just not useful. If you have a single tenant data center with an early install profile there is no rationale to do modularization so it's also applicable to specific business models as well. We've seen it sort of the blade room technology and everything like that rolled out one sort of investment we were considering and I think from that perspective again it's definitely on the smaller scale compared with sort of the big hyperscaling rollouts. I mean it is fantastic from an IRR perspective your risk management is a lot easier you know it's you know what the time will be before you need to roll it out and you can go up in what 600 kilowatt chunks so the upfront investment and the potential to invest something and then have it wilting away for a while before actually coming into use is much less so a brilliant idea but yeah that's scale much trickier. One last thing I might add here is it could be just that this is a concept that's in need of an entrepreneur you know so it's like somebody who's able to build a business out of it so I think fundamentally some concept but it probably needs that Tesla moment where you know somebody does the obvious it makes a success out of it and you know navigates all the treacheries that go along along with a new idea like that. Yeah it reminds me of when Cole Crawford from Vaporio created the I can't remember if it's called the chamber it's a ceiling I can't remember if the name is actually the chamber now but he created and they kind of transformed the whole computing I mean it transformed the idea of edge computing it transformed what we called edge and when you see the machine you're like all right now I understand what edge is and I remember seeing it the first time so you probably need something like that someone like that to come up and then create something new but I think we've seen a bit of modernization in some US markets but mostly I think you see it more within emerging markets so the southern hemisphere where you you're building less you're building smaller amount and the Asia's easy to probably test it versus bigger markets like Europe. I know I know you're kind of getting to the last 10 minutes now so I'm going to throw of course the geographic question which is what location I'm going to be looking out for next year in terms of investment because I mean especially already mentioned the Nordics but what locations are we going to be seeing popping up even more next year and beyond the the tier one cities because we kind of know what's happening there but what's the new up and coming and I'll throw in the M&A angle at it as well where we're going to see more M&A geographically speaking and I know you're going to try that one carefully because I'm sure there's NDAs but I'll leave that to you. Listen I think I think what we're going to see is that that shift to the tier three to four markets I think that's going to be very relevant and what we see there is the the growth profile of hyperscalers that is actually going to materialize. We saw that case very clearly when we did the financing for a company called Atlas Edge that was able to bridge the gap between what the hyperscaler demand is and how much capacity they can offer in tier three, tier four markets you know if you look at a hyperscaler team they need to go to their capital communities as well I think the capital education committees you know you can easily request 50 megawatts in Frankfurt but you know do 10 megawatts in the Munich area or Hamburg area it's a different story so those type of operators that are able to bridge the gap between you know smaller projects but still culminating that that growth profile there I think are the winners in the industry in in that respect. What we see as well is you know significant profile in the Middle East there is a lot of enterprise demand there that is not being covered yet by cloud service providers and yeah I think that we'll see more emphasis on yeah as said the Nordics we see high density compute high performance compute applications and we see enterprises becoming more finessed in how they want to run their IT infrastructure right and we have seen early examples of that already with you know BMW doing their compute in Iceland and then when the engineers come back they have the results in Munich we see that there are more specific IT requirements that are being accommodated by specific data center models and that's very exciting to see so it used to be this is the data center sector and now we see very specific bespoke models that in my view are very exciting. It's absolutely no longer one size fits all it's it has to be bespoke to the industry and the application you're trying to to drive the demand for. Mr. Sebastian? Yeah I mean I agree with Sik, I think we've whilst we still got a focus on we want to have a base in the flat markets but our focus is and it always has been very much more on those secondary markets. I'm glad you brought up Atlas edge there are tenent of ours in in our Barcelona facility I think their approach to things is is is fantastic in terms of looking at those markets where there is that under supply there's the fundamental reasons for data centers to be there it won't be a 18 megawatt hyperscale but it will be there there is definitely requirement I think that's a lot of where our focus is and on those sort of connection points between various markets as well when there's more sub-C fiber is interesting I don't think we're going to be going too far off the beaten path in terms of trying to develop something sort of in the the back end of say the Nordics or something like that but again focusing on those major cities where there is that connectivity coming through and where we do see demand building out I think it's going to be a lot of interest for us. Southern Europe is definitely an exciting place to look at with the different use cases and that's it so look I think much like Seiko and Sebastian this is a rising tide so rising tide lifts all but and I was at a data center conference in Zurich last week and I was amused to they kept referring to Switzerland as an emerging market because obviously we're talking about data center as an emerging market so so I think across Europe you know Nordics of course but markets like Switzerland, Estonia you name it I think you know Italy Milan the number of data centers being built around Milan airport at the moment so you know I think all of them is a rising tide but even further field you'd be amazed by number of data centers being built in Jakarta and Kuala Lumpur in Mumbai you know this is so I think you ain't seen nothing yet and as such from a capital allocation perspective you know people need to allocate capital of the area they know about as a macroeconomic trend you know digital infrastructure is here to stay and if anything you know it's sort of just climbing up to a level of maturity now so it's more about your understanding of the local regulatory market etc to to make those bets whether you know Jakarta or Mumbai or Zurich it's you know I'll drive those dollars you ain't seen nothing yet I like that I think that kind of actually sums up what's going on in the world but I think that all ties up very well as well to do the the earlier point that we you guys made around the different types of buildings that are going to need to be built from the edge building to the the mothership data center some people call it and that's where then the diversification within the metros will accentuate even more and that's where the knowledge are coming into the second stage but it's not going to be a stage where you're going to build small facilities it's going to be the big best they're going to go into the Nordics because of cooling in the cheap electricity then I was going to say as well it's just a very quick question to end I don't have 50 billion dollars to give each one of you like like you gave an example early on if I gave you one billion dollars where would you go next year what what city what locality municipality something specific where would you go to with a billion dollars to build something you can't be a hundred million you can't do anything with a hundred million nowadays yeah there is a question who wants to you know wants to throw a name into the table I would I would do several strategies first of all I think the market is very much focused on hyperscalers and with a lot of investor demand you also tend to get a little bit of an over focus there is still a very significant enterprise demand profile and that commands also a different price profile than what the hyperscalers get so I would put a significant portion in in the enterprise demand because what we see there is that the order profile is going up we see that the contract tenders are being stretched and we see that those parties still become very nice return profile for for projects so that's one second one is you know the tier three tier four markets clearly we see there are a lot of development Milan yeah I think we all see the activity there but there is a lot of Milan's out there across Europe right so if you do the math where is the enterprise where is the tele-density and where is the capacity you know where to invest so I would put one third of that billion there and the other one third I would put into let's say the long-term next generation type of technologies right there is a lot of things happening in semiconductors so you know auto electronics is a very exciting area edge applications call will definitely get part of my investment there so yeah that's how I would split it yeah there's a few r&d aspects that I would definitely bet some money in as well from the next 30 40 years like DNA storage I think that's going to be quite exciting and very disruptive Sebastian where would you go it's difficult following second a very good I just wanted to geography um I mean I'm slightly biased we've got our strategy of building out in with some focus on the core markets but um sort of acquiring assets across and I think very much shadows against the first part of his answer um sort of that enterprise demand in locations which are more sort of secondary and what tertiary and third and fourth tier struggling with my English um and I think that that for us is a very interesting subset of the market that isn't really getting too much of the attention at the moment and will sort of remain sort of strong from a demand perspective given the business the people the wealth that is existing in a lot of these locations across Europe okay and the final words um so I'd certainly give some of it to SICO because I think his answer is fantastic but you've got an angel yes I think if one had that there's an element this is a market of immense information barriers so it's not for the faint hearted and if you've not been doing data centers for 10 15 years um you know down just march out with a billion dollars you're you're very likely to lose it and and if you look at private equity returns um you know even private equity funds um it's almost like a venture plane in many ways so that some people have made a huge amount of money but there's also some big losses there as well so um so I think from a purely capital allocation perspective my primary strategy would be passive and go with people uh like SICO who you know and Sebastian who know what they're doing and have been doing it and park the money there one final thing I would say uh you know there's an element of some venture style investments that are right here and they're actually not in data centers they are in power um so in onsite power in basically allied services so not not sort of the gold rush itself but but selling selling the Levi's to the to the people so I think that that would be an area where I'd I'd put some allocation in is be it uh onsite LNG or or you know gas or things like that and and and you know things about nature right well I really loved the three answers that you gave because you actually flip the question around um usually people struggle not to say the areas where they're actually going to invest and then they have to run away from the question so there was a very good flip on how to answer it so thank you so much Attif, Sebastian and SICO I found that to be extremely insightful and that we could be here talking for a lot longer because I know we didn't even get through half of the questions we had lined up for this chat so thank you so much uh and on behalf of the speakers I'd like to thank everyone for tuning in either here on LinkedIn YouTube X or JSA.net you can also watch our two previous round tables on JSA's LinkedIn and YouTube pages one discussing the future powerhouses of Europe data centers 2024 and the other looking at the financial frontier of Europe and leashing sustainable investments in 2024 with that we wrap up today's conversation look out for the playback of this session coming soon to JSA TV, YouTube, LinkedIn, X and more from me and the team at JSA Europe all that's left to say is to wish everyone a jolly holiday season and a prosperous 2024 see you next year and as always happy networking. Thank you very much everybody. Thank you very much. Bye now.