← All episodes
S2E14

The Grid-Positive Data Center: AI, Energy & Infrastructure | DAG | S2E14

with Christian Diendorfer, Heinz Nusser· Decision Advisory Group· 64m

TL;DR

Data centers are evolving into energy-dense infrastructure assets whose deployment is bottlenecked by grid connection timelines and mandatory waste heat utilization laws in Europe. Co-locating on-site renewables and battery storage offers a viable pathway to bypass congested grids, while heat pumps are essential to upgrade residual heat for district heating networks. Successful development hinges on coordinated multi-stakeholder business models rather than treating waste heat or site selection in isolation.

Synopsis
As rack power densities surge toward 100 kW under AI workloads, data centers are shifting from hidden IT facilities into major infrastructure nodes that demand rapid power access and proactive grid integration. In this episode, Christian Diendorfer and Heinz Nusser from Decision Advisory Group examine the regulatory, spatial, and technical challenges of deploying data centers across Europe. They discuss the impact of mandatory waste heat integration policies, the economics of heat pump utilization, grid queue dynamics, and why battery storage is outcompeting hydrogen for short-term data center resilience. The conversation provides actionable perspectives on site selection, European data sovereignty, and the critical need to align grid operators, municipalities, and hyperscalers.

Key metrics

by the numbers · 6
  • 3 to 4%
    current share of global electricity consumption by data centers
  • 25 and 40°
    typical temperature range in degrees Celsius of data center waste heat
  • 50 100 kW
    projected rack power density scaling from 5 to 10 kW
  • 10%
    projected share of US electricity consumed by data centers by 2030
  • 80%
    share of Austrian data center capacity concentrated in Vienna
  • 15 to 20 years
    construction time for a nuclear plant making it unviable for immediate data center power needs

Topics

6 tags
data centerswaste heatgrid integrationheat pumpsbattery storagedistrict heating
Stats
Duration
1h 03m
Words
9.7k
Questions
56
Speakers
4

Daniel Schaub · Rahul Mishra · Christian Diendorfer · Heinz Nusser

Timeline

9 chapters

  1. Data Center Power Projections

    The hosts outline soaring rack power densities and projected data center electricity demand growth through 2030.

  2. Guest Introductions and Decision Advisory Group

    Diendorfer and Nusser introduce their engineering and project finance backgrounds and DAG's advisory focus.

  3. Site Selection and Capacity Complexities

    Nusser details why site selection must weigh local power costs, scale, and stakeholder coordination.

  4. Regulatory Mandates for Waste Heat

    The discussion covers European requirements making building permits contingent on residual heat utilization concepts.

  5. Bypassing Grid Congestion

    Diendorfer outlines how on-site solar, wind, and battery storage allow facilities to bypass lengthy grid connection queues.

  6. Data Center Typologies and Data Sovereignty

    Nusser categorizes data centers by latency needs and explains why European digital sovereignty is driving local siting.

  7. Upgrading Heat with Industrial Heat Pumps

    Diendorfer explains how industrial heat pumps bridge low-grade cooling runoff to municipal district heating temperatures.

  8. Regional Deployment in Central and Eastern Europe

    The guests evaluate data center economics in CEE, highlighting that waste heat utilization is a regulatory necessity, not the core business case.

  9. Quick-Fire Round and Nuclear Realism

    Diendorfer and Nusser answer rapid-fire questions covering SMRs, hydrogen versus batteries, and capital allocation.

Key insights

4 takeaways
  • 01

    Waste heat recovery is a permitting hurdle rather than the primary business driver

    While EU directives and national transposition laws in Germany and Austria make construction permits contingent on waste heat utilization plans, heat sales do not create the core economic case. Operators must proactively coordinate with municipal utilities and district heat networks purely to ensure their facilities can legally be built.

  • 02

    Islanded renewables and battery storage outpace traditional grid expansion queues

    Data center operators prioritizing time-to-market face grid bottlenecks governed by rigid first-come, first-served connection queues. Co-locating dedicated renewables with rapidly cheapening battery storage allows projects to bypass grid constraints and deploy capacity significantly faster.

  • 03

    Battery cost declines have squeezed hydrogen out of short-term data center resilience

    Rapid capital expenditure drops in battery energy storage systems have relegated hydrogen to long-term seasonal storage rather than primary backup power. High-density battery storage paired with intermittent generation provides sufficient operational reliability for data center demands without the round-trip efficiency losses of hydrogen conversion.

  • 04

    Nuclear power timelines fail to match the data center sector's speed requirements

    Despite growing corporate interest in nuclear and small modular reactors, 15-to-20-year construction schedules are incompatible with data center operators demanding 50 MW ramp-ups immediately. Shorter-term deployment needs keep the near-term focus squarely on renewables, storage, and grid-friendly load management.

Pull quotes

6 quotes
  • Consultancy is a wide field, yeah? And and it's a long way from PowerPoint to power plant, yeah? That's That's what we learned over the last decades.
    Heinz Nusser
  • It's a power-to-heat cube, and it makes no sense to to blow that heat, yeah, in in into the environment
    Heinz Nusser
  • The The first is the business of business is business. Right. Don't expect businesses to be sustainable for whatever reason.
    Heinz Nusser
  • The waste heat utilization or residual heat utilization never is the business case. It It doesn't even make the project economical or not. It's smart to use the heat and and it's regulatory requirement for a good reason to use the heat.
    Heinz Nusser
  • You see the challenge in in in one KPI, and that's time to market. When When are you able to go online with the next 50 MW? 50, 5-0.
    Heinz Nusser
  • And you could go with nuclear if you're willing to accept a building time, construction time of the nuclear power plant for 15 to 20 years, Wow. but this is not what we're seeing what data center operators need, they want energy now.
    Christian Diendorfer
Transcript114 cuesClick a timestamp to jump
  1. Daniel Schaub: Welcome to a new episode here in The Energy Bridge. Um Today, we have a special episode because we are going to talk about data centers with two experts. And just as a why we are talking about this today: We all know that the data centers um consume a lot of energy. So currently is near to 3 to 4% of the global uh electrical consumption. You have a significant of um waste heat between 25 and 40°, Uh there are opportunities that can be used as with together with heat pumps, and also uh some figures for the future as well. We saw here that the power density it's going from racks from 5 to 10 kW to 50 100 kW. And also the projection for 2030, it's expected to consume the data centers consuming 10% of all electricity in the US and near to 5 to 7 5 to 7 in Europe. Also, the European growth European expect to grow 160% by 2030. And the context in Vienna 80% of the data center capacity is concentrated in Vienna. Also, as a metric it's um in 2026, the whole data center consumes as much electricity as 100,000 households. So that's why we're here, one of the biggest electricity consumption. So thank you very much for for coming today. And yeah, looking forward to learn more.
  2. Rahul Mishra: Hey, welcome, Heinz and Christian. We uh we welcome our guests with some introduction, but data centers and the problem needs no no other introduction. But we want Before we dive deep into the data center topic, please tell us about yourself and also about Decision Advisory Group.
  3. Christian Diendorfer: My name is Christian Diendorfer, working for the Decision Advisory Group. My background is basically rooted in engineering and energy economics. I have a PhD in mechanical engineering, an MBA in energy management. And I special specialized during my career into energy systems, thermal systems, and in the past, I've applied my knowledge as a facility and concept engineer in the oil and gas industry. I was a research engineer at the Austrian Institute of Technology, where I did a lot of work in the area of heat pumps, heat integration, focusing also on hydrogen and decarbonization of energy-intensive industry. Then I expanded my knowledge and my experience more on the high-level policy Mhm. level, working for OPEC as a research senior research analyst, participating in UN climate and environmental negotiations, but also working on the long-term scenarios for the World Oil Outlook. And in addition, I act as a senior advisor to the World Energy Council in Austria. And now I'm working for the Decision Advisory Group since about half a year, and I made this step because we are focusing very much on data centers, and I see that see data centers are at the epicenter of the energy transition, requiring a lot of energy, but also providing a lot of potential in regards of excess heat. Mhm. And this makes it really interesting and highly rel- relevant to make good and well- founded decisions for data centers, because this will pave the way into our energy future.
  4. Heinz Nusser: I'm founder and CEO of Decision Advisory Group. We We might abbreviate DAG, it's it's faster and easier. Um and was originally educated to be a tax advisor 25 years ago, and that was so boring and I was not accurate enough, so I I um soon um changed into project management and project development, and made my years in in the energy industry, so real oil and gas upstream Mhm. projects and gas storages, and everything you did back then. Um and 10 years ago, I I studied uh at Hochschule St. Gallen for 1 year sustainable business. And from that moment on, it was uh it was clear to me to only develop sustainable infrastructure and uh work on sustainable infrastructure projects. And well, now with with DAG, we are at the at the junction of decarbonization, digitalization, and and energy transition. And, that's So we are.
  5. Daniel Schaub: Yeah. And and what is your background? Because you are expert like managing stakeholders, kind of coordinating. Yeah, yeah. Yeah. How
  6. Heinz Nusser: Well, I I I talk to everyone. Important. Yeah, yeah. No, it's um the I mean, the the the focus of our work is the business case, uh? But the business case is not is not only the return on investment. A business case, if you see it holistically, is the value of a project, uh? And that that is the public value, it's a sustainable value. So it's really about people, planet, and and profit, uh? Mhm. And talking about um about the planet, you are in renewables, you are in sustainability or in in ecological projects, yeah? But talking about people, you are in stakeholder management, you are in leaving no one behind, you are in engaging society to be part of the of the story of digitalization, for example. Yeah. And when we start with figures, it's it's not only about terawatt hours or uh or um whatever energy flows, it's about it's about the the the nerves and the blood that that's that's pushed through through society. I mean, it's all data, uh? Mhm. And that's for me so interesting to be in this data center industry, because it's power to heat cubes? No, it's a it's actually the the heartbeat of of modern life on that planet, uh?
  7. Rahul Mishra: And and the last time, Christian, we saw you uh you also moderated a panel for our Wave First Forum. Mhm. And and then you went on to uh uh start working for DAG. And coming from your background that you come from more technical, hardcore technical background, getting into a consulting firm, what was the trigger, and uh uh what was the missing link that And also for you, Heinz, like to to get together and work on the same problem?
  8. Christian Diendorfer: During my career, I realized that from a technical perspective, Mhm. there's so much good work available, and if you talk about energy storage, heat pumps, and in particular in Austria with the AIT, there's so much ex- excellent work done. But the missing link to bring it into action is somehow the decision process, and Mhm. bringing together technology, economics, policies, and also breaking up outdated perceptions on on how the the energy world evolved. So when I was a student 20 years ago, solar energy and PV, that was rocket science and only used for satellites because it was so expensive. And only in a few years in the last few years, the price has declined so much that it's now one of the cheapest energies. And we are seeing the same trend for for batteries, and I believe a lot of people are still living in this past world, thinking about how expensive renewables are and all the risks, and not seeing what what we did in terms of technology development, and so we need this decision processes, this social this techno-economic modeling, showing that there are actual business cases and there are actual benefits, and I believe that these are the key topics we as consultants are answering and dealing with, and trying to to move to move forward.
  9. Rahul Mishra: Heinz, um uh since I also had the uh pleasure of working with in consulting uh back back in Netherlands, and consultant as as the word is It's it's like a stereotypical word in terms of where you have classic engineers, but it's it's an important tool. And when you have uh someone like uh Christian in your team, tell us about the dynamics of your team as well. And and how do you drive a process? Mhm. Like a decision-making process?
  10. Heinz Nusser: Consultancy is a wide field, yeah? And um and it's a long way from PowerPoint to power plant, yeah? That's That's what we learned over the last decades. In a team like ours, we we have to clearly understand who's an expert, uh, who is a manager, and who is a consultant. So, the- those are different profiles, yeah? And due to the to the the size or the limited size of the company, we we have to Everybody has to be sort of a a a best of all, yeah? Uh and still, we have we all have our strengths. I'm not an engineer, uh but I'm I'm brilliant in in in facilitating multi-stakeholder decision processes. The The core dynamic is is curiosity and the will to bring expertise into the real world, uh? Mhm. And the moment we see this in in in someone and and that person fits into the into the um competence of the company, that's that's the moment when we see the company grow. It's it's not It's not a numbers game, it's a it's a it's a skills and attitude game, uh? And that's that's why we are, for example, with Christian happy to have him on board. Brilliant engineer with the will to bring things in into life, uh? And the decision process, I mean, it's it's Decision Advisory Group, so uh there is an underlying process called decision quality. Mhm. And uh we strictly follow that process. I I I never deviate, uh but not always our clients see that process. It's sometimes the questions we ask um and that that the the lack of assumptions and our our hunger for data and and reliable information that that brings us to a high quality decision basis, which is super crucial, especially nowadays when when the scene is flooded with assumptions and and guesses, and yeah, and everything.
  11. Daniel Schaub: That's it. Especially now that uh yeah, we are now in beginning of March, everything in Iran, it's happening, so uncertainty everywhere.
  12. Heinz Nusser: Yeah, and and and so many things are post-factual, uh? I mean, AI is has the potential to create that that parallel world right in front of our eyes.
  13. Daniel Schaub: I mean, just so understand because we talk before recording what you are doing, and and that's why we were interested in in having you here, that you are thinking from the designing point of view, where to place a data center, where it's more um cost-effective to do it, to help your clients to get the better business case, where to place it, what are the ecosystem around, so um just in a few words, um how how do you help your your client? What is the process?
  14. Heinz Nusser: One step back. It It's not We're not only doing data centers, it's data center and industry. Uh so it's it's We also help industrial clients. But data centers are the the ones that are allocated right now, so that's why why it's such such an important topic. Um Actually, we start one step earlier. It's The question is: Where does which data center make sense? Uh? So it's it's there is no sense for hyperscaler in a in a high energy or high energy cost area, uh, just because you have that many acres of land available. It's
  15. Daniel Schaub: Yeah. What are the sizes of the data centers, just to have a
  16. Heinz Nusser: In in Yeah, in megawatt.
  17. Daniel Schaub: Yeah, in in consumption.
  18. Heinz Nusser: Well, st- starting from 2. I mean, it's it's depending on the type of the data center. It's It's 2 to 200, if you say Mega or Mega megawatts. Mega. Okay. Yeah. Yeah. So that's But in you you have modular solutions that might be smaller, and you might be a part of a hyperscaler park where where 200 MW is one one step. It's, you know, one module. It's I also hear it's smaller. But um the uh the I think the the the core added value is to help our clients to maneuver through this complexity in terms of energy, business case development, regulatory questions, uh? It's so much to consider, uh, and the data center operators are coming from a world nobody knows, uh nobody knew, actually. So they were plug and play in hidden places, they have fences around their data centers, and their core staff was not on LinkedIn because you should not uh address them, and so on. Um and the power grid was developed almost almost the same way, uh? So it it grew, yeah, by driven by politics, yeah? Now they have to talk with each other, yeah? And now and and and complexity is still growing, so uh the moment you have enough power and access to the grid for your data center, yeah, the the next hurdle is, "Oh, you're creating waste heat, you have to make use of it." So now the data center operator has to talk to municipalities Yeah. to heat heating grid operators. Yeah. The Those people never met before in history. Mhm. Yeah, and And and now we we we have to bring them maybe not on a table, but we we have to to fit them into one business case, yeah? The overall business case must fly. And then we can cut it, yeah, and have energy crown contractors here, investors there, operators here. But if the total business case is not valid, stop the project.
  19. Daniel Schaub: Yeah, also in in that point is how data center has uh evolved, because they started just like uh clouds with storage, now they have computing, so Mhm. we we are going to talk more about uh business cases, but now that you mentioned this new stakeholders that they didn't knew e- each other, and and now they are they have to talk, data centers can be seen as a new best friend or a new enemy, [laughter] right? So, so how do you address, for example, a mayor to to start talking about allocate a data center with a Mhm. convincing that you can use the waste waste heat, or what is the approach with a mayor, for example?
  20. Heinz Nusser: I mean, um coming back to your sentence, uh data centers usually are primadonnas. Yeah, so they they they want it cool, they want it quiet, A primadonna, like. yeah, no shaking, no they don't disturb me, safety, a lot of safety. Yeah. So that's that's more a data center, yeah? They They have special needs. And when you when you want to build a data center, you have to ask for a construction permit. Mhm. That's the latest moment you might meet the mayor or one of his or her um employees dealing with your request for a construction permit. Uh the regulatory framework in the European Union, yeah, says you have to make use of your waste heat or residual heat, uh? So Germany started uh with the Energy Efficiency Law a few years ago. Um Austria now um with mid of this year uh will or made it uh obligatory to make use of your waste heat. So now whoever, the project developer, that whoever this is, an engineering company, has to to to negotiate or to find common sense with the municipality, yeah, because without a valid concept for waste heat utilization, you don't have the construction permit.
  21. Daniel Schaub: That's in all Europe.
  22. Heinz Nusser: Well, that that that the uh that that the European uh guideline is published I think 2 or 3 years ago, uh and ratification of the in in national law is an ongoing process. As I said, Germany made the start, Austria uh followed, and and all the other countries will have to follow, yeah. Mhm. And it makes sense, I mean, it's it's It's a power-to-heat cube, and it makes no sense um to uh to blow that heat, yeah, in in into the environment, uh, when
  23. Daniel Schaub: So even if the city doesn't have a district heating, they have to find a way.
  24. Heinz Nusser: They have to They have to They have to develop a concept for that. And then, yes, there are some some um ways out of that if you if you have a concept, but uh you did not find a partner although you tried, and so on and so on, yeah. And the same moment, uh a lot of energy goes into cooling of data centers, and it's interesting that the discussion is rare that somebody says, "Hey, I sell you cooling services by taking off your heat, uh, and not paying for heat, but getting paid for cooling, yeah, which in a certain environment might absolutely be that the business case. Yeah. And if you see it holistically, the the whole project, why not? Depends on the on the on the on the project, on on the type of the um of the data center, um and and I mean, there are so many open questions. If you ask a data center operator to tell you that the the the amount of waste heat that will be created over the over the next years, Mhm. it's like asking them to open their P&L, because they would tell you their utilization, because it's a simple ratio, power in, heat out, yeah? So if they don't create heat, they're not up and running, yeah? So you you immediately could see their utilization, yeah? Well, companies don't want to show you their Mhm. workload and utilization. So there is a lot of um of uh hurdles you you have to take, but but it's simply it's possible and it Yeah, absolutely.
  25. Rahul Mishra: Mhm. And coming to uh a bit more technical part of the data centers, especially with the grid constraint that we already uh facing in Europe with a lot of decentralized assets being uh getting inclined in the grids. Where do you see data centers coming in the picture as well and already constrained grid. Uh How do you see the picture? What will be the technical route for it?
  26. Christian Diendorfer: There are several options on the table, and one would be to simply bypass the grid by co-locating renewable energy sources and energy storage close to the site, near the site, so simply getting data centers out of the grid. This would allow us to provide the necessary energy faster, and than waiting for for the grid expansions. Then we could also transform the way how we operate data centers, and make them more friendly to the grid, Mhm. more communication with the grid, and data centers are mainly used for AI services, and AI services are a great source for improving the operation of the grid, Mhm. how we use energy. So we shouldn't forget that data centers provide a lot of options and opportunities to improve our entire energy system and increase the efficiency we have there. So it's not only a huge energy consumer, it's also a service provider for efficiency improvement. And in addition, we have the excess heat or should also integrate the excess heat into existing existing infrastructure, Mhm. extend the infrastructure in a way that we can use this excess heat. So if we do the correct integration, which requires modeling, policy discussions, and and the right common direction all the stakehol- holders are going in, then they are, I think, a great opportunity for for the development of our energy systems.
  27. Rahul Mishra: And just on that note, also uh with you, Heinz, you said uh you were talking about different stakeholders, right? So, what is the perception of uh the grid operators for data centers and around data centers?
  28. Heinz Nusser: I mean, grid operators, um they are by definition, not the most flexible organizations.
  29. Daniel Schaub: Conservative, let's say.
  30. Heinz Nusser: Well, it's I don't know if conservative, but it's a grid, yeah? That That it's It's our bones aren't the most flexible part of our body, and that's how how a grid has to has to work. So my full my full appreciation for their work and full understanding for for their um stability, let's put it this way. As Christian mentioned uh earlier, that the speed for decentralized power sources are getting more and more economic, yeah, by prices for photovoltaics simply dropping, yeah? Mhm. Um and uh and and windmills becoming more and more uh not popular, but but feasi- feasible. [laughter] Um The uh the the the grid is simply facing a complete change of frame conditions, uh? So what if I was a grid operator, I would love to have a national or regional, because we're in Austria, we have, you know, nine regions, and um I would love to have a strategy for for locations, uh? It's like mushroom picking. If you know where they are, it's it's easy to pick them, yeah? And um if if we would know where where data centers will grow, it would be easier for the grid to handle it, yeah? So we should we should think about having data center friendly zones, yeah? Also with with with full implications on uh construction permitting, etc. etc. Um and on the other hand, we have um it's not only data center operators or developers who are asking for access to the grid. At the same moment, it's batteries, Mhm. it's um electrolysis, less at the moment because the hydrogen business case is not that strong. Um We have landowners hoping for the big shot with the data center Mhm. investor, yeah, simply asking for access to the grid. In addition, classical industrial companies who make the fuel switch into power, coming from gas usually, or other fossil fuels, and they also say, "Okay, if if we use large-scale heat pumps, we need more power, and that exceeds the existing access to our grid, at our transformer station, so we need more." And now you're sitting at a at a grid operator's desk and say, "Okay, all those requests, first come, first serve principle." And then you you have to deal with them, one after another, because that's simply how the rule is. And you might change that rule and say, "Okay, there is something like a qualified request, Mhm. uh? And then you you simply have to show up not only asking for capacity, but showing when, how much. There a data center has a ramp up. If you say, "200 MW," it's not 200 MW, it's growing, growing, growing.
  31. Rahul Mishra: Hey, you touched upon uh about like the location where exactly the data center should be, and a special zone for making for them. So that zone uh would be closer to the city where the or closer to the where the energy is produced, or uh how would you define that?
  32. Heinz Nusser: Depending on the on the type of data center. There are data centers um or [laughter] Let's put it the other way around. There are clients of data centers who want their data center right in front of them, actually on-premise. If you're a pharmaceutical industry player, and you decode the human DNA, and make real fancy research, blah, blah, blah, yeah, you want to have that data center right on-premise in your safeguarded uh area. The next step would be there is a data center with super low latency. Imagine uh autonomous driving. There If there is a latency, yeah, in in in my car reacting to the information the sensor is sending, ah not that smart, Mhm. yeah? So uh here we have decentralized data centers, but not huge ones, they're small ones in connection with 5G or in future 6G um um infrastructure. The next one is we have uh decentralized uh cloud computing, yeah? We have co-location closer because still super low latency, closer to the cities, and close to international connection points. And then you you have um like the monks, yeah, in the monasteries back in history, sitting somewhere writing the next version of of the Bible, we have large language models, Mhm. who could sit in in in the Nordics or somewhere, simply doing their job, yeah, but not that latency critical. So it's the the zoning, yeah, is depending on the use case. It's like where where's the best parking lot? Well, it depends if you're a truck, a taxi, or or a private car, or a motorbike, or a bicycle, or a scooter, yeah? It's exactly the same for data centers.
  33. Rahul Mishra: I I just have one more uh curious question around the location of it. Like this this on the on the business side of and also the uh let's say, in terms of how critical they could be. But also um there's a lot of data sover- sovereignty uh question that uh a lot of people want to answer in Europe. Mhm. How critical is that where it should be? Like, some people would say like why even in Europe? Why not just somewhere else? So Mhm. what would that approach be?
  34. Heinz Nusser: Well, I think that that approach changed over the last 12 months. [laughter] Uh quite a lot, actually. I And And I'm serious, yeah. I think uh data sovereignty is something uh each country and and and the European Union or Europe as a community should care about. It simply uh It is critical, it is critical, not only infrastructure, but it's it's vital information, yeah? If um if you turn off Teams, Zoom, and all the other video conferencing tools, that would be a huge traffic jam, because people would have to go to the offices, uh? My My impression is it's the same amount of people in the rush hours in the city, Mhm. but it's less in all the other hours, yeah, and still every- everyone is working. So every Teams call is a non-car ride. So And And and if you if you're able to interrupt that, and you're a hostile, then interrupt it, uh? So in in in a world where where we're not growing to a big friendly family, was that might have been an idea, and I I I I witnessed the year 1989 when when when Eastern Europe was was opened, and the mauer fell, uh the wall fell, and and the Iron Curtain, and so on, the impression was, "Well, now the the ever never-ending summer is starting, and and we all become friends, and it it's a it's a it's a global family and global community." And the the the nerves of consumption and economy will glue us to each other, yeah? And they do, yeah, because we still make business with all the the countries we we we hate in the newspapers, or that are hated in the newspapers. It's still We're still doing business with each other. When it comes to data and seeing this as as a vital element of sovereignty of of uh of not only business, but but actually humanity, then, Mhm. well, take care about it. Get it closer. And it's in my backyard, please. Yeah, we we know this, not in my backyard.
  35. Daniel Schaub: I wanted to come back to the technical part with Christian, uh that we talked before um that the data centers need this reliability of 99, 5 9s or [laughter] of of reliability, and when we talk about distributed distributed energy, it's intermittent, and we talk about also the challenge for the operator. So what how do you address this uh reliability request for data centers being powered with renewables?
  36. Christian Diendorfer: Sure. So yes, you're absolutely correct, the demand for the energy reliability of data centers is simply astonishing, and what they actually need. With renewables, as you said, they're intermittent, but they are also somehow predictable. So it's We have a lot of information on how they provide, we have production profiles, so we can estimate and create models on on how the future energy supply from renewables is going to be. And in addition, what we are currently seeing is an absolute amazing price decrease for battery storage. So the the cost per megawatt hour stored is just going downwards at an unprecedented pace, and combining different sorts of renewable energies, wind, PV, with local energy storages would even make it possible to run data centers as an island. It's probably not the most economic way and not the way to go, because it's If you're connected to a larger grid, it's still cheaper, it's more efficient, and in general, if you have intermittent energy sources, [laughter] the the wider and broader your system is, the more reliable it's going to be. But combining these new technologies to to one holistic system, I think it's it's going to make a huge step forward, and still, you could have some small gas or hydrogen turbines on site or or even diesel engines operated with maybe fossil diesels or even HVO, and more in renewable energy sources or in renewable fuels, which are probably going to be pretty expensive, but for these few hours in the year where you're going to use it, it could make sense to use them as a backup. And with this, I'm I'm pretty confident that it's possible to create very sophisticated and amazing renewable energy supply systems for data centers.
  37. Daniel Schaub: Yeah, so again, battery systems in the table. [laughter] Yeah, and this is something that we have seen also in the US with the mini reactors nuclear, that they are being popular, but we are fostering here more renewables, so yeah, good to know that it's possible, actually. And since you were working in AIT as well, and we worked uh and and we talked about the waste heat, uh how do you see the potential with heat pumps to use the heat weight the waste heat with a heat pump?
  38. Christian Diendorfer: Regarding waste heat in data centers, it's there there are different options and different ways, so if you have a normal air-cooled systems, the excess temperature, or the the temperature of the excess heat, is is comparably low, so there's not really a lot you can do with it. Maybe you can use it to grow tomatoes or Mhm. to to grow fish, but that's somehow not feasible, I would say, in in the the the amount of excess heat you're going to have from a gigawatt data center. Heat pumps provide you the the opportunity to increase the temperature of the excess heat in a way that it's going to be usable for district heating, for industrial processes, and I believe this is going to be the key for excess heat usage, so they're really a critical enabler for excess heat usage. And as you said, AIT is doing a lot of great work on on building large, industrial, high-temperature heat pumps. I think we will see in the future ways of using excess heat we haven't thought of before and thought were not feasible, not economically viable in any way, but I believe there's a lot of stuff going on, and wha- what we need to do, and this is what we are focusing on in the Decision Advisory Group, is bringing stakeholders together, and make them talk to each other, understand each other, what what are the requirements, what are the expectations, the the options, to make the right decisions, and bring different companies and industries together, and I believe with this such a process, we can really make excess heat usage a very efficient, cost-efficient way of Mhm. of using and providing heat.
  39. Daniel Schaub: Have you seen in your work now with uh uh with data centers, some use case for heat waste? For example, Heinz mentioned pharmaceutical, that they need to they need cooling and heating in in their processes. Have you seen some synergy with some industry?
  40. Christian Diendorfer: Sure. There are a lot of of cases going for that. For example, in Vienna, the the largest data center in Vienna is connected to a hospital, providing excess heat from the data center for the data for the hospital, where the the hospital is is a very good consumer because they don't only need heat for heating in in the winter, they have more a constant heat demand for the washing machines, for all the equipment they need. We have it right in Vienna, and it's it's working, it's working very well, and Since when? I don't know the exact date, but It might be 10 years now, I guess. Wow. But it was uh kind of the old the 2017 Old data centers, like before AI exploded.
  41. Heinz Nusser: Yeah, yeah.
  42. Christian Diendorfer: It It's not an AI data center, but it it's working, and and we have it, and it's probably not that well known, Mhm. and I mean, it should be more known, it's super good. Not communicated, the thing we actually reached, and and how amazing it is to have this connection, and that is is a pilot case that is it's feasible, it's working, and on a daily basis. So we have it, and we we just need more of it.
  43. Rahul Mishra: Just a curious question again, what's the difference between the data center that were built 10 years ago and the AI data centers that going to be built now? Is there like more I mean, just a general difference in how how the technology of making data center might have changed as well?
  44. Christian Diendorfer: Looking at it from a pure energy perspective, what has increased is the energy density. Mhm. So each stack now uses a lot more energy than before, which brings us again back to the topic of of excess heat, so with this increased energy density, air cooling might not be feasible anymore, so you need liquid cooling. With liquid cooling, you can increase the temperature again, which makes heat pumps even more viable. So we are bringing a lot more energy into a much smaller space, Mhm. and this is m- the the main change in in data centers from a data center perspective. It, of course, it means we have a lot of lot more calculating power in smaller spaces. So this is what you also see on on your smartphone, on your personal computer or laptop, the calculating power of of each device is constantly increasing, and with this, also energy density is increasing.
  45. Rahul Mishra: And Heinz, uh just again about the location questions with what uh what Christian just mentioned. Did it change uh in terms of last 10 years, how the location with with uh certain parameters that have changed in last couple of years, how much of that had also affected the location? Mhm. Scouting.
  46. Heinz Nusser: You have to access your data center every now and then. In a colocation data center more often, of course, because there are your your clients in there. Mhm. So you don't want to have this in the middle of somewhere after a 3 hours car ride, that would be unfeasible. Uh Back then, it was simply possible to have those data centers in the cities, yeah, because there was power, there was access to the grid, we had more space, and now the uh as cities grow and grow, and get become more denser, yeah, um there is simply no not much not enough space anymore for data centers, yeah? Mhm. We We would still love to have it right in front of us, I guess.
  47. Daniel Schaub: Yeah. Now Now that you mentioned the location, we were talking about uh Elon Musk uh statements in World Economic Forum. So what is What are your thoughts about data centers in the space, or underwater, or in in in the poles, I don't know, this out of the box ideas?
  48. Heinz Nusser: Yeah, yeah. Well, if the business case flies and and and it's sustainable, [laughter] I mean, there are business cases on that planet that are business cases on a return on on a financial return on investment basis, but they are a disaster to the planet and to society, and um knowing that data and data processing and AI, which is more or less more data and more data processing, there is nothing beyond that, honestly, it's becoming even more important to consider that the boundaries of the planet, uh, and the um and the limitations we have to accept in terms of water usage. We had it in our discussion to discussion today. No adiabatic cooling anymore, closed-loop circuits, and so on and so on. Well, we often say the total cost of ownership. Well, let's consider the total CO2 footprint of ownership. And I don't know if a rocket start with with a few satellites and data centers on board uh is has a lower footprint than running the same data center 3 years on a diesel engine, um doesn't happen in Europe, but uh in the US, um that's in a holistic business case. And of course, with PV and and so on, a permanent day, and the cool side, the the dark side of the data center, out of a sudden. Um Well, if that design makes sense um and and the and the primadonna likes it up there, [laughter] yeah, because it's safe, and it's cool, and it's power-supplied, and it's no vibrations, why not? Think it through.
  49. Daniel Schaub: And since you mentioned um the sustainable approach, you were I don't know if you still are professor in Innsbruck?
  50. Heinz Nusser: No, unfortunately not anymore for for 3 years not, yeah.
  51. Daniel Schaub: Okay. So but Yeah. but you were professor of sustainable project management? Tutor, yeah. Yeah. Tutor. So uh could you tell us one point that the student missed when they have to uh address or analyze a sustainable project?
  52. Heinz Nusser: I mean, there are two recommendations to them which one I started usually. The The first is the business of business is business. [laughter] Right. Don't expect businesses to be sustainable for whatever reason.
  53. Daniel Schaub: I mean, we always say that it's economical, social, and environmental.
  54. Heinz Nusser: Yeah, yeah, yeah. But it's business.
  55. Daniel Schaub: It has to be.
  56. Heinz Nusser: It has to be business. So um if you want to make a sustainability project fly, make it the heartbeat of someone's business. And that business might be a lot of things, yeah? So it might be taking off the burden uh for the social insurance system, and by doing so, having the a client who is willing to pay something for you because you you just saved a lot of money somewhere else. So that's a typical indirect or triangular uh business case. So that was something that um that that was important, and it connects the the competence people who study economics, yeah, microeconomics. Uh it it connects their competence to sustainability immediately, because it's the same tool set. We We do SWOT analysis everyday, yeah? It's not outdated, it's still there, but we do it in a sustainable way. And the second uh one um I was taught 10 years ago and I I didn't get the point uh right on spot, but it's: Don't die as a sustainability hero. [laughter] So you you have to feed your children and yourself, you have to pay your mortgage. Um it's nice to be sustainable, yeah, but don't die. [laughter] I mean, stay alive. And that doesn't mean
  57. Rahul Mishra: Sustain yourself first.
  58. Heinz Nusser: shoot shoot people and and rob a bank, [laughter] yeah, but um be as sustainable as you as you can suffer. [laughter]
  59. Rahul Mishra: And and on that note, also Christian, you have a uh research background. And do you feel that the research world is somehow kind of disconnected with the ROI requirement of the of the industry?
  60. Christian Diendorfer: By now. [laughter] Well, of course, there's a certain amount of disconnection, because I I think scientists have a different mindset than business people, and engineers or scientists, they're focused on or they they enjoy solutions that are almost perfect from a from a technical from an engineering perspective, while business people enjoy solutions that are perfect from a business and balance sheet perspective. And of course, it's it's important to to bring this together, and we have a lot of initiatives going into this direction of of bringing research and science more into the market. What we need is more understanding of each other's perspective, so businesses need to understand that the value of research, of science, and how this could actually help, and maybe even being more willing to take risks in in research, so install components that are in an early research phase to Mhm. do some testing, to bring them into market, to just follow econ- economies of scale, which is often a an issue for for research projects. So you build one demonstration or pilot plant, and of course, it's expensive, because it's the only one you have built, but if you build several hundreds or thousands of such components, and increase them in size, they get more efficient, they get cheaper, and I I believe a lot of very good ideas are still very well-researched, Mhm. but are lacking the support for scaling up and and becoming more efficient. And as I said, I already mentioned the the topic of of PV cells, when PV cells, they were built for satellites and space stations, and that was basically the the use case for PV cells, and now you can go to every supermarket and they will sell you PV cells for a few euros, and it it's amazing how not only that the research, but the scaling up made the the technology feasible.
  61. Rahul Mishra: Yeah. And and uh Heinz, we we are very much focused on Central and Eastern Europe, so the also the question arise uh again coming back to the business case for data centers in this region as a block, what do you think of the potential of building data centers in in this region? What is the potential for this region? And also uh Christian mentioned about uh the uh the ro- kind of role model of of using data center heat for a hospital. And this region is also very dependent and relying on the gas. Uh And there's a I mean, we from the outside perspective, we think there's a case, but how do you Mhm. uh see that?
  62. Heinz Nusser: The waste heat utilization or residual heat utilization never is the business case. It It doesn't even make the project economical or not. It's smart to use the heat and and it's regulatory requirement for a good reason to use the heat. Mhm. And that's it. So you would never build a data center to to heat something. The first question, well, invite the primadonna to one of those countries, and look where where where she would love to stay, depending on on who is uh who is paying for the the the the computing capacity in that data center. And if you, for example, have super low energy costs in a cool region, walk around the map you just opened. Um Well, that might be a a good hyperscaler location if it's politically stable, Mhm. if the regulatory framework is stable, and so on and so on. And I'm with you, I think the European Union, for especially uh especially, has not fully identified its potential when it comes to the data center locations. Mhm. Abs- Absolutely not. Um The same moment, if if you rely on gas or coal-fired power plants, well, you you have to switch to renewables. What is the the the the the cost guarantee you have? Same moment, renewables are that the the have the lowest volatility in in in costs if you directly uh purchase their energy and and and make use of it. So for me, the solution is obvious. Yeah, the market is is partially there, um but still, you you you need the the customers more or less on site, otherwise you would go into another place that is more obvious.
  63. Daniel Schaub: And and what is the feeling of the industry in terms of the region? Are there some plans to install data centers in some of these countries?
  64. Heinz Nusser: I think the pipeline is not fully filled in more obvious locations, because you need the uh uh you also need the uh the the the the terrestrial grid, I mean, the the fiber cables, yeah, that that a data center makes sense. Um and I think we are beyond FLAP. I mean, Frankfurt, London, Amsterdam, Paris. Mhm. Yeah, so that's Yeah, that's tier one, fine. Mhm. And we are beyond that, and we we open new areas uh for for large-scale investments. Rheinisches Revier in Nordrhein-Westfalen, for example, was a blank spot on the data center um map a few years ago. Started 2 years ago with a study and with Microsoft being interested in in investing there, and that's boom town at the moment, Where is it again? uh it's left of Cologne. Okay. Düsseldorf, Cologne, Aachen, in the middle of that triangle, former lignite uh power station. Okay. But uh it would be super interesting to to to coming from a uh from a uh meta perspective to develop such clusters proactively, yeah? Mhm. And the if the need for for new locations is rising, then I think that those locations might be the next tier that's that's interesting for investors, yeah.
  65. Rahul Mishra: And on um just to understand the size of the problem that we are facing as well, what is the size of what do you see approaching in next 5 years?
  66. Heinz Nusser: You see the challenge in in in one KPI, and that's time to market. When When are you able to go online with the next 50 MW? 50, 5-0. We don't talk about a gigawatt, Mhm. um because we have the ramp-up, yeah. But if you say, "Okay, give me 50 with the potential to increase to 400. Where is it?" The 50 I need now, actually. That's And we need it in Europe, honestly.
  67. Daniel Schaub: Yeah. And also for Christian again, the we have heard some hydrogen projects in the region, in Central and Eastern Europe. Uh Do you think some relationship between hydrogen and data centers, that they could be powered with them somehow, that increase the hydrogen demand somehow or what is your vision?
  68. Christian Diendorfer: I think we are seeing is that battery storage is eating into the use case of hydrogen, as batteries are getting cheaper and cheaper. So for short-term energy storage, I believe batteries are the way to go. There are use cases for hydrogen, in particular, long-term energy storage, use cases where the molecules are actually needed like for the steel industry and and other use cases for for petrochemicals, and so on. So yes, hi- hydrogen is going to play a role in the future, but we're seeing that the role is is changing, creating electricity out of hydrogen, which was created out of elec- electricity, really only makes sense if the storage component is make makes the business case. So if you have need long-term energy storage, and you store electricity generated during summer, store it as hydrogen, and use it in winter, and the business case makes it possible, then I I see a use case for hydrogen. But as as a fuel for daily operations, I'm I'm not so sure if if hydrogen is is actually go is the way to go.
  69. Daniel Schaub: Interesting. Yeah, short-term storage. Chemical, long-term, molecules. Exactly. Kind of.
  70. Rahul Mishra: Super interesting conversation, and we can keep on going. I have multiple other questions, but uh getting to the last part of our uh uh our interview, we call it the quick-fire round, asking you some on-the-spot question. Putting you on spot, Heinz. So [laughter] what one word you would describe a stakeholder who says no to a green project?
  71. Heinz Nusser: Interesting.
  72. Rahul Mishra: Oh, you find it interesting when they say no?
  73. Heinz Nusser: Yeah, and it would trigger a discussion. But First of all, I would say, "Interesting."
  74. Daniel Schaub: "Tell me more." [laughter]
  75. Heinz Nusser: "Explain yourself."
  76. Daniel Schaub: Yeah. Um Another short question, it's uh which business case do you see now more profitable for data centers?
  77. Heinz Nusser: I think the AI- related non-large language model applications will be the next wave that really makes sense and, thus, will build a business case.
  78. Daniel Schaub: Well, will which Uh I mean, such an No, but it's it's it's not building another LLM, well, which is done, yeah, here and there, but um making um the a a very specific use case for the utilization of AI uh in a certain industry, or translated on in the old world, we don't uh write the next encyclo- Encyclopedia Britannica, but we write the handbook for the heart surgeon. That would make sense, and that can be I don't know, forecasting of grid blackouts, uh preventive maintenance models, things like that, very specific, high-density, smart use cases, huh?
  79. Daniel Schaub: Yeah, so uh you also mentioned uh before like computing as a service, kind of like to rent part of the computing um How How do you call it? The
  80. Heinz Nusser: Mhm. Well, well it's Capacity as a service. capacity as a service, yeah. Um the Yeah, of course, I mean, that's that's simply the the the the machine, uh? I would rather see a model that is uh that is so secure that a company allows itself to upload their own data Mhm. to get the perfect response that really boosts their business case. And that's why we need that model up and running in Europe and not somewhere else, because you will not upload your company data to I think an an an appropriate regulatory framework for that, yeah.
  81. Rahul Mishra: Nuclear energy for data center? Asking Austrians about nuclear energy is always like a [laughter] But
  82. Christian Diendorfer: Okay, so so you're with nuclear, you're opening a huge box, where we could discuss hours about, I I would say, but as we're talking about data centers and from the current perspective, and as Heinz said, the most critical issues for data center operators is time to market. And you could go with nuclear if you're willing to accept a building time, construction time of the nuclear power plant for 15 to 20 years, Wow. but this is not what we're seeing what data center operators need, they want energy now. And as the sooner the better, and I don't see nuclear being an option for providing energy on in this short time range.
  83. Daniel Schaub: Which has been the most difficult city to work with? In Europe.
  84. Heinz Nusser: In Europe? Yeah. Oh, that that that's a that's a restriction, huh? [laughter] Yeah, the most difficult city in Europe. Oh, that's difficult, because I cannot travel there anymore, I can tell you. [laughter]
  85. Daniel Schaub: And worldwide?
  86. Heinz Nusser: Yeah, worldwide, I think uh for me, the Kathmandu was was a disaster.
  87. Daniel Schaub: That's in Nepal?
  88. Heinz Nusser: Yeah. Wow. And it was for we we we had it's wasn't a data center, but a cable car project. But it was simply by it it was physically demanding, yeah, after the big earthquake, it was dusty and and And I guess I would not like to work with the city of Brussels. [laughter] Not because of of the European Union, but the the the sheer existence of that many districts that are completely independent from each other, that why the construction the road for construction is so complex in Brussels. I think that would be a city um I would think twice to to [laughter]
  89. Rahul Mishra: There There's something in common between Kathmandu and Vienna, you know? Yeah? It's both are uh called as city of spies. [laughter]
  90. Heinz Nusser: Yeah. [laughter]
  91. Rahul Mishra: Yeah, yeah.
  92. Daniel Schaub: Yeah. Um the next question is about if you had 1 billion euros, how would you invest it to accelerate energy transition or data center deployment? You can split if if you want.
  93. Christian Diendorfer: So I would start, so
  94. Heinz Nusser: You take 500 million and tell us now. [laughter]
  95. Christian Diendorfer: I wouldn't invest it in in technology, actually, I would invest it in education and energy literacy. What I see there's so much perceptions on how the energy systems work, how the economics behind it work. We have a lot of discussion now about the merit order, and I'm not really sure if everyone understood how the system works, and I believe we have a lot of lack of of understanding of this It's a really complex system, so I'm not I really want to say it's it's complex, there are so many different stakeholders, so many different perspectives, positions, and I believe we need a better understanding on that, and if we educate the people, stakeholders, the companies, I think this would make a greater benefit than just So, so basically, teaching people on on how to use energy, rather than investing in systems providing them with energy. I think this would make more sense, and this is where I would invest the money.
  96. Rahul Mishra: And also, uh Heinz, you do you think like the policymakers usually that you you come across, they're energy literate?
  97. Heinz Nusser: Well, [laughter] some of them are, yeah, some of them are. But it's not their job to be experts, uh, it's their job to translate expertise into digestible uh mess- digestible messages, and that's unfortunately where they very often fail. Mhm.
  98. Daniel Schaub: Yeah. Actually, this is something that we are addressing in our forums, that's why we are inviting policymakers into the discussion, because it's a super-regulated sector, so they have to be where the discussion are are happening, so we hope to help you in your mission, Christian.
  99. Christian Diendorfer: We tend to simplify things and assume that things are easy, but in reality, they aren't, and this is why such forums are a great opportunity to exchange information, to demonstrate how how complex the systems are, and what solutions can be reached if the system is is well-understood.
  100. Rahul Mishra: You can give uh part of your 500 million to us as well.
  101. Christian Diendorfer: Definitely. [laughter]
  102. Daniel Schaub: And yeah, we were told that you are a wine expert. So, which kind of wine would you say that is energy transition, like a long-aging wine or more sparkling with innovation? Energy transition. Yes.
  103. Heinz Nusser: Well, it's an interesting comparison with wine. Huh? I think if you if you take champagne, um that's that is sparkling, it has a very very complex It's actually a bad wine, yeah, champagne, because it's it's a cold wine-growing region, you shouldn't grow wine in the in the Champagne. So what they do is they make something special with it, add add some more yeast and then the bubbles created in the bottle, and and so on and so on. And then it's durable, yeah, and it's sparkling even after years or decades, and it and the good ones are getting better. So for me, the the energy transition is so super complex, so if I if I would compare it, I would rather say it's a it's like a a brilliant bottle of champagne. [laughter]
  104. Daniel Schaub: Nice one.
  105. Heinz Nusser: Gives me a hint for tonight. [laughter]
  106. Daniel Schaub: And for you?
  107. Christian Diendorfer: And, well, I I'm not a wine expert by far, but I I I would say it's and and we are in Vienna, so it's probably the Wiener Gemischter Satz, [laughter] which is it's it's So basically, what you do is you you mix grapes of different Sorts. Assemblage, I think, in Yeah, but and and and it's basically a mixture of of different types of of grapes, and and I think this is what we need for the energy transition, so different technologies, different business models, different approaches, and we mix it together, well balanced. I I want to add, so you cannot simply mix I wouldn't mix Coke Mhm. into the wine like some people do to make a high quality wine, but really think about what what needs to be in this mixture, and when put all these components together, and then it's going to be a a great experience.
  108. Daniel Schaub: Wow.
  109. Rahul Mishra: Amazing. On that high note, [laughter] Yeah. we uh thank you very much, Heinz and Christian. It's been a pleasure. I think it's a very persistent topic, and I think uh we need to go deep into it. It's it was very insightful for us as well. And I believe uh for community to also understand what is happening with data center, and how it's going to in the coming years uh change our energy system as well, because it's going to be a persistent part of it.
  110. Daniel Schaub: Yeah, actually, we're super happy to work with you together for our next forum, 2nd and 3rd of June, that we're going to have a session especially for um data centers, that are we were surprised that how much attention it's having, and interest that people are asking for more. So yeah, I think that we can do something to also um increase the knowledge about the topic with the community and and different stakeholders. So thank you very much for coming today, very insightful.
  111. Heinz Nusser: Thank you, it was a pleasure.
  112. Christian Diendorfer: It was a pleasure.
  113. Daniel Schaub: And thanks for the community for watching the episode, and stay tuned for next interview here in The Energy Bridge. Ciao, ciao.
  114. Rahul Mishra: Bye.