clean-IT: Towards Sustainable Digital Technologies

1. 00:00With this, I will now open the first panel: "Climate Impact of Digitalization" and as the first impulse, as the first panelist, I would like to welcome
2. 00:12and ask for his input:
3. 00:14Professor Doctor Dirk Messner. He is the President,since 2020, the President of the Federal Environmental Agency. He was previously director
4. 00:28of the UN Institute for Environment and Human Security, Vice Director of the UN University and Director of the German Institute.
5. 00:37for Development Policy, thus
6. 00:39he knows what he is talking about. We're glad to have them.
7. 00:42Maybe we'll do it that way after their impulse, if they're already going to the podium then, that we'll then follow up with
8. 00:50when everyone has given their feedback, we can then start directly with the discussion. Thank you very much, Mr. Meinel, my very
9. 00:57Ladies and gentlemen, dear Franziska, it's nice to see you live again, that's great.
10. 01:02I'm speaking from three perspectives on the topic of digitization and sustainability: For a long time, I was a member of the scientific advisory board "globale
11. 01:08Environmental Change", for ten years, and the last major report for which I was jointly responsible, that was on the
12. 01:14digitization and sustainability, because we found that the topic was not being dealt with at all, that these research
13. 01:19communities don't stand together at all, that we don't even know
14. 01:22on the sustainability side where the potentials are, there is a lot to do. That's my first perspective of more scientific
15. 01:27Advisory Council "global environmental change", then in the Federal Environment Agency the State Secretary has just indicated that is digitization
16. 01:34is an important topic for us.
17. 01:35I've made that more important again, sort of in the last two years since I've been there now, because in this study we're
18. 01:40of the WBGU have shown that there is a lot of potential, generally not mobilized so I'm talking about missing links, we
19. 01:47have to create a lot of missing links between these potentials and what we actually want to solve.
20. 01:52The third perspective I'm speaking from is an international perspective.
21. 01:56We at the Federal Environment Agency have formed an international alliance with a number of other partners.
22. 02:00Mr. Meinel, we should also talk about that, I think that's very good, where we try to address this issue also on the global level
23. 02:06moving forward, because you talked about the SDGs.
24. 02:08The interesting thing is yes, when the SDGs were adopted in 2015 or, when we were discussing Paris
25. 02:15then also adopted this this autumn of multilateralism 2015,
26. 02:20digitization didn't play a role at all.
27. 02:23In this SDGs, you don't find this digitization mentioned at all, and we are arguing today, after all, with good reasons: "We
28. 02:29will not be able to solve the SDGs without digitization."
29. 02:32However, digitization, even if we don't align it properly, can lead to accelerating
30. 02:37of the unsustainable growth patterns, so with even more force, so to speak, digitization can lead to growth spurts
31. 02:44produce, which produce us even more climate problems, even more resource problems. Creating these links
32. 02:50is very important to us.
33. 02:51We work at the global level with UNDP together with UNEP together in the International Science Council.
34. 02:56the Future Earth Alliance, is indeed the global sustainability research and in this context we advise the Secretary General
35. 03:04of the United Nations to reinterpret and realign the SDGs from a digitization perspective.
36. 03:11There is a process called Global Digital Cooperation, and we are involved in it with our network
37. 03:17to redefine the environmental dimension of digitalization, so to speak.
38. 03:21So we should definitely talk to each other because they are working on a lot of questions, a lot of issues that we are also
39. 03:27play a role, and there is worthwhile, I think
40. 03:28Cooperation. I would like to say something about three aspects and I think this fits well with what Mr. Meinel and the Secretary of State have already
41. 03:35have stated. From my perspective, from our perspective, on the one hand, we have to put "greening IT" first, which is basically
42. 03:44the first necessary step that we have to take, so that digitization itself does not become a problem.
43. 03:51There you have the relevant numbers,
44. 03:52Mr. Meinel, called. "Greening IT," I'll say 2-3 sentences on that in a minute. The second point then is we have to try to better
45. 03:59understand where digital tools and digital innovations can become enablers to solve the problems that we face through the SDGs
46. 04:07described by the target system, i.e., "digital as an enabler, as a problem solver," and
47. 04:14there's still a lot of work to be done on how we can best use that so that we can be more effective in climate change mitigation in the
48. 04:19resource issues and in the circularity of the economy, use these digital tools in a way that we can move faster.
49. 04:25That's the second dimension, "As an enabler." And the third dimension is,
50. 04:29if I look quite optimistically at the development, maybe that's the disruption that we're looking at in climate protection, Franziska,
51. 04:37have been waiting for.
52. 04:38You pointed out: We are in a "decade.
53. 04:41Factor Three." We've done 2% reduction in greenhouse gas emissions per year over the last 10-15 years, 2%! We
54. 04:49but now need 6%, towards 2030. tripling!
55. 04:53We will need to expand renewable factor 3 in this decade and continue like this in the next decade,
56. 05:00Factor 3. If you look into building retrofits, also factor 3. So these technologies are the disruption and the
57. 05:08potential that we can leverage to really move forward quickly here,
58. 05:12and I would like to give you 2-3 thoughts on that as well. That's the basic structure and now just a few thoughts on each.
59. 05:19The first point "Greening IT". There, in principle, I can only subscribe to what we have already heard.
60. 05:24At UBA, we are working very specifically on three issues at the moment: The first issue is the data centers, about that
61. 05:30we have already talked about, the data centers. We did a study, from UBA where we looked at.
62. 05:35The first important finding is: we don't know anything about the energy efficiency of the data centers, because we don't have any
63. 05:39standards, we don't have monitoring systems, so that's where standard meeting and regulation and benchmarking is needed,
64. 05:45big field.
65. 05:47They also executed that
66. 05:48Mr. Meinel. We have found data centers with high energy efficiency, we have data centers
67. 05:54found with incredibly poor energy efficiency, but there is no transparency in the market, we have no
68. 05:59mechanisms there to provide competition, and we need those. The second field we're looking at, it's also already been addressed,
69. 06:04"Software," Franziska you brought that up, we work closely with you guys, that's where the problem is just like Mr. Meinel brought up the
70. 06:10describes, from our perspective. We don't have heuristics at all yet to measure the energy efficiency or the energy consumption of algorithms.
71. 06:18and software in a really consistent way.
72. 06:22That's why a number of research projects, one.
73. 06:24you just mentioned, Franziska, where we are looking at these issues of methodologies and heuristics for efficient
74. 06:31energy efficiency consideration of software. Then we'll look at digital tools and infrastructures and products.
75. 06:39The observation there is: a lot of the products that are digitally driven have very short life cycles and then you quickly end up
76. 06:46with e-waste. Lifecycles,
77. 06:48and extending those life cycles on the one hand, and how do we advance circular economy so that we can
78. 06:54the resources that are consumed here can actually be recycled back into the digital business.
79. 07:00Those are the big challenges that we have to deal with.
80. 07:03And that's the UBA package that's going on there right now, so greening IT, that's the first dimension if we don't get that done,
81. 07:10then we won't even get to the second step, which is "Digital as an enabler", irony of history, so to speak.
82. 07:17We have great hope that these tools will help us unlock the potential in our economy that we need for
83. 07:23Climate action and circular economy.
84. 07:25But first, we need to see the digital processes themselves become sustainable.
85. 07:29So the potentials in the enabling area. Here I am very optimistic, I have to tell you, because almost all of the
86. 07:36studies that are available show that the potentials, I'll sign off, potentials, not what we realize, the potentials
87. 07:43of digital innovations and tools for decarbonization, for circular economy, for dematerialization, for monitoring.
88. 07:51of ecosystem and therefore protection of ecosystems and then optimization in climate research itself.
89. 07:57As far as our models are concerned all these potentials are underlined in majority of the studies.
90. 08:03So we have to lift those now.
91. 08:05I am optimistic in this dimension, we need incentive structures for this.
92. 08:08As long as we have a tax system that essentially taxes labor and hardly taxes resources, we should not be surprised.
93. 08:15that digital innovation is being used to save labor, not resources.
94. 08:19This is about appropriate incentive systems so that we can mobilize this potential. That's the second dimension, optimism
95. 08:25from my side and now the third dimension comes briefly and that brings me to the end,
96. 08:30is that the disruption we need, so this can be a game changer, is it not just about containment
97. 08:39of risk if we don't create Green IT, but is there a potential here that really sustainability transformation
98. 08:45that we're all talking about with a new perspective.
99. 08:49I think there's a great opportunity there, and I just want to outline that with two examples.
100. 08:54The first example is, we have, if we look at the large companies around the world, studies that show if you look at the
101. 09:00largest 500,000 companies look.
102. 09:0240% of these companies now have a science-based climate strategy.
103. 09:07That's good, it should be 100%, we're at 40.
104. 09:10Of that 40%, only 5% are taking advantage of the opportunities of their climate strategies, with their AI strategies, which they are also pursuing, with a
105. 09:20To combine with each other.
106. 09:21That is in the companies themselves, even in the large companies around the world, have a parallelism of AI processes on the one hand
107. 09:28and digital innovation and climate protection on the other hand, and so they can do what I just mentioned under enabling
108. 09:33mobilize at all because of that. We had
109. 09:35a few days ago, Franziska, with Robert Habeck and the Minister of Science and the Industry 4.0 platform,
110. 09:41about this, the discussion: how we in the companies can do this merging of climate orientation, climate strategy on the
111. 09:47one side and digitalization, how we can better combine these. There's still a lot of potential there for Game
112. 09:54Changing, because the proportion of companies that have already systematically linked their digital strategies with climate strategies
113. 10:01Is vanishingly small.
114. 10:03There's still a lot of room to go, there's a lot we can support from science.
115. 10:07Politics can set the incentive systems, the companies themselves have to put this on the agenda.
116. 10:12And the second point I would like to conclude with now, this auditorium is exactly right and Mr. Meinel,
117. 10:17You have said everything that is relevant there, and I would just want to underscore it: We need to do research on the
118. 10:23digitization and research on climate neutrality, circularity, and sustainability together.
119. 10:30And there's an incredible amount of room to go up there as well because both of these research communities, you pointed out,
120. 10:35there's not even a journal for it, they're barely networked together yet.
121. 10:39Now if I had some slides with me, at this point I usually show a slide with the highlights of the AI research
122. 10:45of 2020/2021 from the international congresses of AI research. Then you can hardly believe when you see the top ten
123. 10:52Topics:
124. 10:53But there's not a single one in the top 10 either,
125. 10:56that with our questions that we are discussing today "Greening IT", sustainability and digitalization, there is not a single one
126. 11:02among them. So we still need to merge these two things also in the research area, and then we have great opportunities
127. 11:07not only to narrow down problems, "Greening IT", but to raise the enabeling potentials and maybe even to Game Changing
128. 11:15Situations to come.
129. 11:16Thank you very much for your attention,
130. 11:24Mr. Messner, thank you very much, it is always a pleasure when a scientist also heads a federal agency, for the fireworks of ideas.
131. 11:32May I ask the next panelist, Dr.
132. 11:36Bernhard Rohleder here on stage, he the was a founding member of Bitkom Verband der IT-Industrie in 99,
133. 11:49has been the chief executive officer since then.
134. 11:52Some people say Bitkom
135. 11:54be
136. 11:54Rohleder, was previously in senior positions at the VDMA,
137. 12:02in the Brandenburg Ministry of Economics at ZF Friedrichshafen,
138. 12:06please give their input. Mr. Meinel thank you very much, at this point I'm always almost under so much pressure to justify why I'm doing the
139. 12:13been doing this for so long, but I can say it's a job that totally excites.
140. 12:18After all, we all don't just have one life, we have several lives, I also have two lives and the first life I worked on in the
141. 12:23Environmental research. I did my PhD on a topic that we're also talking about here 25 years ago and I was talking almost daily about
142. 12:31deja vu every year. At the Environmental Policy Research Unit, first reviewer at the time, was the Chairman of the International
143. 12:37Panel on Climate Change, so issues that moved me at the time and that are now intertwined with this second passion, the
144. 12:43digitization, can be brought together quite wonderfully. When I say deja vu, I mean it quite seriously, not only in my
145. 12:50own academic training and now in my profession, but also in the topic of green IT, they just
146. 12:56addressed Mr. Messner. We have in 2007 for the Cebit, that used to be the world's largest trade fair, a few will be
147. 13:05know them.
148. 13:05For the year 2008, the core theme sought the guiding theme, the guiding theme was "Green IT". 2008, so it's been almost 15 years now
149. 13:14and we built up a traditional office there as Bitkom, so there was really someone sitting there and typing, and we
150. 13:20set up a digital office and ran electricity meters on the side for seven days over the duration of CeBit.
151. 13:27how much energy is consumed there.
152. 13:29It was quite impressive numbers on that practice example. Did that really make a lot of difference?
153. 13:35I would say no. We are discussing very similar issues today as we were 15 years ago, but with a very different drama
154. 13:43Than we did back then.
155. 13:45And if you ask yourself, "What is the reason that less has moved than could have moved and should have moved?",
156. 13:51then it certainly has to do with the speechlessness between these two worlds.
157. 13:55I still experienced this in the old federal government, at a round with Svenja Schulze it was about digitization of agriculture.
158. 14:03We were the only business association sitting at the table. It was a small
159. 14:07Round, all the environmental associations were there.
160. 14:09The farmer's association was not there and what I already noticed there was a lot of reservation from the area
161. 14:17of NGOs towards digitization.
162. 14:19Not everybody really liked that at all when I talked about smart farming and talking about precision
163. 14:24farming can reduce nitrate inputs to groundwater, because in many cases it's not just your own passion that's going to be reduced.
164. 14:32on climate and environmental protection, but because other lifestyles and other economic designs are also related to it.
165. 14:38connect.
166. 14:38I think we're in a situation now where we have to be extremely pragmatic about these issues and how we can
167. 14:44need to think about: How do we manage to solve this most pressing problem that humanity has ever had, this most pressing
168. 14:50problem, and digitalization can perhaps also make a decisive contribution to this and, by the way, it can also help to solve this problem.
169. 14:57without leading to dangerous discussions about abandonment.
170. 15:02We can make it in many areas, not everywhere, but with digitalization, that we maintain our standard of living
171. 15:09possibly even improve it and still and protect the climate at the same time.
172. 15:14You have quoted the study, Ms. Brandner, that we as Bitkom have produced. We are also criticized for this study
173. 15:20been.
174. 15:20That was so a little bit nicely calculated etc, so we have
175. 15:23there I can say yes, us as Bitkom very far taken back, we have only in order given, have the results
176. 15:30looked at. They were also partly surprising for us, I may also say that.
177. 15:34We found out some impressive numbers.
178. 15:36Accelerated digitization 151 megatons of savings by 2030. 41% contribution to more ambitious climate
179. 15:45and savings targets that we have in Germany, that's impressive.
180. 15:48Of course, there is also a separate footprint of digitization.
181. 15:52There is also a leverage, which is a factor of 5, so 5 times more CO2 or CO2 equivalent savings than own consumption.
182. 16:00There's one area where, in terms of footprint, digital can't really do much.
183. 16:06We have to use the energy that's generated from resources as it's just available in the marketplace, and we can
184. 16:12reduce the footprint where we buy other energy.
185. 16:17You just described that, Mr. Meinel, renewable energy sources, are not so heavily developed right now that
186. 16:23we would have 100% corresponding electricity and as long as that is the case, there is this competition for the resource regenerative
187. 16:31generated energy, and if that's where the big cash-rich companies come in, then of course those are the first ones,
188. 16:37that pull there with an energy provider, and it's a renewable power that's missing elsewhere.
189. 16:42But the footprint, I will say, can be significantly, can be significantly reduced by also changing the energy mix
190. 16:49in Germany.
191. 16:51Which fills me with, for all my optimism, some concern,
192. 16:56that is the strong effort of policy makers to intervene in a very deep regulatory way in processes of the economy.
193. 17:06I don't think it's going to work because we don't have the competence to really go in very deep at that point, in
194. 17:12the algorithms and then to make a reasonable decision, which algorithm is the right one in a
195. 17:19overall perspective really
196. 17:20the more climate- or resource-friendly algorithm compared to a corresponding competing algorithm.
197. 17:28And to that extent, at this point, my plea would be for policy makers to focus, to focus on what they're particularly good at, the
198. 17:34framework and give very strong price signals to the market that the strongest signals that companies are receiving are
199. 17:42can they are much stronger than regulatory intervention, and when we set price signals, so when primarily resource consumption is
200. 17:49becomes more expensive, then that automatically means that digitization becomes more competitive.
201. 17:55Small example fiber optic networks, copper networks.
202. 17:58We have relied on vectoring in Germany for a long time. Vectoring means I have to use an extremely large amount of energy to get through the same
203. 18:03copper cable to bring a larger amount of bits and bytes.
204. 18:06Fiber is the much more environmentally friendly technology.
205. 18:10There were wrong signals given by policy makers, in regulation, and at the same time energy was just way too cheap.
206. 18:18So
207. 18:18I personally look with some skepticism at the current discussion about heating costs, subsidies and about too high energy prices.
208. 18:26It's actually exactly what we need.
209. 18:28We need authentic prices on the primary
210. 18:31resources, that's what we need.
211. 18:33And then the economy is oriented out of self-interest, because companies always behave opportunistically, out of self-interest
212. 18:40to these price signals and develops technologies that are particularly resource-efficient.
213. 18:45The biggest opportunities we have are in the area of substituting digital power for physical power.
214. 18:52What we're doing here is a typical
215. 18:53physical meeting.
216. 18:54I actually drove here in the car, I drive almost everything else
217. 18:57by bicycle. Got a little physical problem,
218. 18:59that's why it didn't work today.
219. 19:01But of course it is much more resource consuming and intensive here than a video conference in the appropriate
220. 19:09framework. Where we can substitute, that's where we're strongest, that's when we're most effective as a digital economy
221. 19:17And make the biggest contribution,
222. 19:19in terms of climate protection and resource conservation.
223. 19:21That would be my appeal also to future work, that we look very strongly,
224. 19:25Digital twin is one such example, we have calculated that as well, that we look at where we can leverage physical power
225. 19:32substitute with digital power and focus quite strongly on that.
226. 19:36That should be it.
227. 19:37For now I look forward to the discussion
228. 19:45Many thanks
229. 19:45Mr. Rohleder. Next, I would like to welcome a colleague from Potsdam from the GFZ, the GFZ Earth System Modeling Section.
230. 19:56here from the large-scale research institution Geo-Sciences, Professor Doktor Maik Thomas, please come in
231. 20:03You. He has been leading this GFZ Earth System Modeling Section since 2007, and since 2013, he has also been a professor of fluid dynamics in Earth
232. 20:15system at the FU Berlin.
233. 20:17We look forward to your input.
234. 20:20Thank you very much for the introduction. Good morning to you. When I was kindly asked, here is something about the conference.
235. 20:27to contribute, which of course I find from that exciting,
236. 20:30initially thought well, can't say anything at all. I'm an oceanographer, I run mathematical models on observational data.
237. 20:42So I actually can't say much about the technical things, but what I might be able to present to you is why
238. 20:49do we need digitization in the geosciences, specifically why do we need high performance there
239. 20:56Computing? Is that usable tool or is it more gimmicky?
240. 21:02So two examples that you all know so weather forecasting, there is
241. 21:08the meaningfulness, usually nobody questions, is based on models. With climate models
242. 21:15well, there the opinion is not so very clear, but there are already discussions, but there are many other possible applications
243. 21:24for numerical models.
244. 21:25These are only the two
245. 21:26first ones, the most prominent ones.
246. 21:27And I would like to show them where we use essentially numerical models at the Geoforschungszentrum in Potsdam.
247. 21:36So what we do, we confront our models with reality, with measured data. There are definitely modeling branches
248. 21:47that don't need any reality at all.
249. 21:50And now say that quite self-critical. But what we do,
250. 21:53we used numerical models to evaluate these various geodetic data and geophysical data. What they
251. 22:02are seeing up there,
252. 22:04what they see up there are different measurement systems: terrestrial, satellite based, sometimes airplane based
253. 22:11supported so the diversity is very high.
254. 22:15These measurement systems are extremely precise, but the problem is they always measure the sum of an almost infinite number
255. 22:25of processes acting simultaneously.
256. 22:29So we say we have an integral signal, if they have an altimeter, that measures sea level, there they have
257. 22:34although a distance actually anrollende running time rather yes not and from that you say
258. 22:39the sea level is so and so high. So until you get there it's already a long way, but even if they know that, they know
259. 22:45still don't know, is there just more water in the ocean or it's gotten warmer or none of the above and just the bottom
260. 22:51has moved. And that's the same with many other systems, for example, the gravity field.
261. 22:55You may know the Potsdam potato from GFZ, so based on gravity field measurement for example with the satellite
262. 23:02Grace mission, which we run together with our American colleagues.
263. 23:06But it's not until you unleash complementary methods on these observational data, that is, you first have to get the raw data
264. 23:14process, and you have to somehow develop methods to interpret this integral data, then they often come
265. 23:20the quantities that we're actually interested in. Sea level change, groundwater storage changes, or what about
266. 23:26Ice mass changes?
267. 23:27They need models for that.
268. 23:28So if you abstract it a little bit, this is how you can represent it.
269. 23:32So we are in the process of contributing with models throughout the value chain.
270. 23:38So starting with processing and cleaning observational data and ending with effect separation.
271. 23:44How do you get from knowing the sum the individual
272. 23:47summands?
273. 23:48Then, in terms of the causal processes, we can interpret the signal. Once we understand that, we can
274. 23:54by assimilation improve the models and then we have better production, better prediction and this
275. 24:01prediction, which we can also use, among other things, to optimize, for example, future monitoring systems,
276. 24:09to optimize economically or in terms of output. So the applicability of numerical models is much more diverse.
277. 24:16than is widely perceived by the public.
278. 24:19What is a numerical model, anyway?
279. 24:22Can be explained quite simply, so first of all we summarize all that we think we have understood of the earth system
280. 24:30or of climate dynamics in the form of equations.
281. 24:33These then become so complicated so quickly that a closed analytical solution is no longer possible at all?.
282. 24:39We then call these data of these equation monsters nonlinear partial differential equations, which per se are not
283. 24:44solvable, so what do you have to do?
284. 24:47You just have to discretize them in time and space, and then you calculate for certain lattice-
285. 24:52points the solution for certain parameters that are of interest at the moment. That's the basic approach, that is, they have
286. 24:58in space the variables and as a function of time, so four-dimensional. And it's clear that the quality of a
287. 25:07of such a model of such a prediction depends on the model construct.
288. 25:11How complex is the model, what processes are carried along, how it space-time resolution and how is the scalability
289. 25:17In terms of computer architecture.
290. 25:20With all these parameters, of course, the resource requirements increase, in terms of CPU processing time and memory,
291. 25:31all depending on that.
292. 25:32So we're using these computer architectures to tickle the processes out of the Earth observation data that we've
293. 25:41of interest for climate research.
294. 25:44If we times it so few numbers here opposite is completely correct, us is not at all conscious:
295. 25:52What do we actually consume with such a model?
296. 25:54Nobody asks that and it would be very difficult to know because it's very sensitive.
297. 26:00Regarding the model design. I've included two examples here. It's the same American Earth system model.
298. 26:07All configurations are the same, resolutions time step and so on.
299. 26:12And the only thing that differentiates those is how you take atmospheric chemistry that just a tiny component and they're
300. 26:18see immediately there's computational time going through the roof.
301. 26:23The whole model performance is different because you can compute on other kernels,
302. 26:26and so on. That is, we can't even keep track of that as geoscientists, that is extremely sensitive, but what we
303. 26:33perceive, of course, is the demands, are getting bigger, despite all the development also in terms of hardware computing time-
304. 26:40demands are getting bigger, but also the memory space is getting bigger, if there is for example Climate Model Intercomparing Project
305. 26:46stand out, so a single example and look at the development over time these are typically, also
306. 26:51if these numbers terabytes are not so high nowadays, but that's just one model, we're talking about multi model ensembles,
307. 26:56then that also goes up exponentially.
308. 27:00I have also collected a few headlines to get a sense of how much is the resource requirement.
309. 27:07I don't even want to read them out one by one.
310. 27:09Many examples Mr. Meienl has already brought much more concisely, just want to repeat This example with the data centers
311. 27:15the demand for energy that it is comparable or the emission of CO2 is comparable to the entire aviation industry.
312. 27:22But, and this is the last point, listed on this slide, the climate and earth sciences are massively dependent on
313. 27:30On computing and storage capacity.
314. 27:32Storage capacity particularly in the context of Earth observation data. And you can now ask yourself, is the cost of all this
315. 27:38worth it?
316. 27:39That's why I want to show 2-3 example just now where we apply these model methods to tease out relevant quantities of the climate system.
317. 27:48The first example is for example a gravity field measurement over Antarctica, you think there is
318. 27:56not much, there is no vegetation, hardly any life and so on.
319. 27:59And you now measure this trend in the gravity field that means there is a change in mass and in the early days of the Grace satellite there was
320. 28:06Publications in Nature, in Science and so on.
321. 28:08Where it was simply said: It's quite clear, you see a deficit of ice masses here in the western region
322. 28:13in the eastern area an accumulation of ice masses, but one has not considered that one also has the dynamics of the solid earth
323. 28:20and if you do that, then you see that this positive fraction in the eastern area is
324. 28:27Land elevation.
325. 28:28And there's really only one deficit in the Amundsen Sea sector in the western area. so this same correction, that you have all the
326. 28:35effects first, you can also do that in Greenland and then just look at what the signal of the ice looks like and
327. 28:42this is shown here once for Greenland. And here you can see over the years how the ice masses change over time.
328. 28:49In particular, deficits in the southeast area of Greenland and so on.
329. 28:53They can then integrate all of that and they can then separate the total mass trend.
330. 28:57But they can't do that if they don't apply complementary methods to the Earth observation data, because then they're going to have to
331. 29:04interpret things as ice mass change that are not ice mass change at all, and that's the kind of thing that you can in principle
332. 29:12do with the water mass changes for the whole earth.
333. 29:16We see, for example, what a signal in terms of the deformation of the time-varying water content on the continents
334. 29:25so the deformation causes this time variable water masses on the earth you can also convert that
335. 29:32of course, shortly comes an example for Europe for Central Europe, in deficits of water masses and there you can visualize that
336. 29:40what we all have experienced in the last summers, this water mass-
337. 29:44Deficit in central Europe.
338. 29:47Now you can say, that's where I come to my last example.
339. 29:50Well, the resolution is nice for application purposes that you did that, but we can't do that much with it there
340. 29:57These are not even in the catchment area of water basins and for that we are now just using methods of artificial
341. 30:04intelligence. Not as complex as the examples they gave.
342. 30:08We also work in completely different levels, always depends: what kind of uncertainties do you allow?
343. 30:13How exactly do you want it?
344. 30:14But that's a typical Grace signal,
345. 30:17What you see on the left side. You can't see anything there that actually the average person can do anything with.
346. 30:21But if they unleash neural network artificial intelligence methods on this field, then they have a field like this
347. 30:28is on the right-hand side, and of course you can do that for the whole time domain.
348. 30:32At the top is, for example, something like the observation, which you can then obfuscate, and then you look at how does that react
349. 30:39System out of that with the artificial intelligence and that's the bottom and out of the match See, that works
350. 30:45quite well, although we only have one resolution like that.
351. 30:47On the left, you can work out very fine structures from it and that's just an economic aspect again because they then
352. 30:52don't have to invest too much in sensor technology and in satellites, but also with such a coarse signal relatively much
353. 30:58with such a coarse signal.
354. 30:59So what I just want to say is that models are not only used exclusively for prediction but we need them in all steps of the
355. 31:06value chain: to process, to interpret Earth observation data, but also to estimate uncertainties.
356. 31:17Since the requirement for computation time and storage space depend very much on the complexity of a
357. 31:24Model, on the resolution and many other parameters.
358. 31:27And I leave you now: Is it worth it?
359. 31:32Thank you very much.
360. 31:39Thank you very much, dear Mr. Thomas, great mixture, which we now want to complete with Oliver Süme.
361. 31:48He is the Chairman of the Board of the eco Association for the Internet Economy, run the world's largest Internet node and look after
362. 31:58take care of all issues concerning the Internet economy.
363. 32:02He himself is a lawyer with Fieldfisher in Hamburg and was previously president of the provider association
364. 32:10EuroISPA to mention that as well. Mr. Süme please your input.
365. 32:15Thank you very much for the kind introduction,
366. 32:16Prof meinel. I am pleased to be here today.
367. 32:19You have already briefly introduced the Eco Association, for which I am speaking here, and you have mentioned DE-CIX in particular.
368. 32:25That is particularly a special feature for us as an association, that we are not only stakeholders of companies,
369. 32:31but with DE-CIX as a wholly owned subsidiary, we ourselves are also a central component of the Internet infrastructure.
370. 32:38not only in Germany, but worldwide.
371. 32:41DE-CIX now operates such Internet exchange points at 22 locations worldwide.
372. 32:48And to that extent, we are particularly, of course, a welcome customer of data centers, data center operators.
373. 32:54That's why the data center industry is particularly close to our hearts, and it's already the subject of various comments today.
374. 33:01have been.
375. 33:02By the way, we can read quite well at DE-CIX what is happening in terms of digitization, because of course we see the data throughput there,
376. 33:10that is happening in Germany in general. We assume that about 95% of the total traffic that happens in Germany is
377. 33:18is coming through this DE-CIX node in Frankfurt am Main, distributed to different data centers there, that that's going through there.
378. 33:26if you look at what happens especially in the pandemic years in the lock down and home office phases at the DE-CIX
379. 33:33is, then this speaks a quite clear language.
380. 33:35We have seen 20% more data throughput, data growth at DE-CIX in 2021, and of course that's due to
381. 33:45to the framework conditions, which we all know, digital infrastructures, digital applications were used intensively.
382. 33:52That has yes also a very positive effect in digitization booster, which we like to talk about, and it shows just
383. 33:59then also in the data throughput at DE-CIX.
384. 34:01We just detected a new all-time peak now in February, we measure the data throughput there per
385. 34:08second and in February we measured 11 terabytes per second data throughput for the first time. To give you an idea of that
386. 34:17that corresponds to about 2.4 billion written DIN A 4 sheets per second, mind you, if you were to stack it up
387. 34:24they would come to a height of 240 kilometers, that happens there per second.
388. 34:29And that shows of course digitization is data intensive and digitization and especially data centers need energy
389. 34:37that's clear, but digitization above all also contributes to less CO2 emissions in many areas.
390. 34:48energy consumption.
391. 34:49The overall life cycle assessment, I think it's important to emphasize, of digitalization is quite positive, there are numerous
392. 34:57examples.
393. 34:58Let's take the one that has probably affected all of us: Working from home, in other words, the failure to require
394. 35:05to drive or take public transportation to the offices to work has had a
395. 35:12enormous CO2 emission saved.
396. 35:15Just one day of home office per day, there's a study by Greenpeace on this, saves about 1.6 or 1.8 million metric tons
397. 35:23CO2 emissions per year.
398. 35:26These are things that we can save on, but also especially when we look at digital applications in industry
399. 35:32particularly in manufacturing also there are studies that say by consistently using digital
400. 35:39applications in production, we can create resource savings of up to ten billion euros per year.
401. 35:46And of course these are also processes where we can become more efficient in production, in the maintenance of production processes.
402. 35:54and that also saves energy and that can of course be expanded and that must also be the goal.
403. 36:00Crucially, of course, the digital infrastructures themselves also have energy saving potential, and considerable potential.
404. 36:08Mr. Rohleder mentioned it earlier, comparing copper technology to fiber. That is true to a similar degree, by the way.
405. 36:16for the cellular standards as well.
406. 36:18Switching from 5G to 4G saves 80% energy compared to the old standard when they compare it to 3G
407. 36:2598% so also in the infrastructure application itself through innovation and latest technologies in digital infrastructure
408. 36:35we have quite considerable savings potential and that is something that we should make much more use of in the future. by the way
409. 36:43also the move of many companies to the cloud, which is still far from complete.
410. 36:47Many companies, especially in the midmarket, are still working on their local, self-operated infrastructures.
411. 36:54Is actually an absolute madness in terms of energy.
412. 36:57When you move to the cloud, no matter which major provider, you save up to 80% of the energy that you would normally spend on
413. 37:03costs on the clock if you run that as an enterprise yourself, so significant potential there as well. if we look at
414. 37:12take a closer look at the data center industry, then of course, Ms. Brandner
415. 37:18they also said it,
416. 37:19of course this industry needs energy, but you also have to know that fortunately in Germany we already have
417. 37:25in a global comparison, especially in a European comparison, we are among the most energy-efficient data center operators.
418. 37:31include.
419. 37:32We are very far ahead there, of course there is still room for improvement, and above all we are also very far ahead in the
420. 37:39Innovation of the computing processes that take place there. If we look at the development since 2015
421. 37:45then we've used 12 times less energy per workload since then than we did in 2010, because of course what's going on there in
422. 37:55the data centers, the computing processes, which are of course indispensable for digitization, are being optimized further and further.
423. 38:03are becoming more and more efficient.
424. 38:05And that's a path that we want to continue to follow, of course, and to that extent we're very excited about the legislative announcement,
425. 38:11Ms. Brantner, outlined this earlier, we are looking forward to being able to contribute with that.
426. 38:18Keyword waste heat has been dropped.
427. 38:20I think that in particular is a great potential that this industry has the data center industry, and we
428. 38:28already have incredibly great best practices for what's possible.
429. 38:32We know because we are located there just as DE-CIX, especially in the Frankfurt market very well, that is the largest data center
430. 38:39density that you can find in Germany and there are various best practices there including the building complex
431. 38:48Eurotheum, which is already saving up to 40% in energy costs today with the help of data center waste heat.
432. 38:5670% of the waste heat from the data center is used to heat this building complex.
433. 39:04And if we look at the potential of the data center industry as a whole in Germany, it shows how
434. 39:09high the potential for this waste heat utilization is.
435. 39:12We have about 50,000 data centers, in Germany there are also many small ones, but also large ones, as we have them
436. 39:18in Frankfurt.
437. 39:20If we were to consistently use this potential in Frankfurt alone as waste heat by 2030, then theoretically they could be
438. 39:27heat all office and private buildings in Frankfurt with data center waste heat.
439. 39:32That shows the potential of that.
440. 39:36And to that extent, consistent waste heat recovery is indeed a very crucial key and a contribution that this industry
441. 39:45can make to operate energy efficiently and with less CO2 emissions.
442. 39:51The big question now, of course, is: How do we do that, which set screws do we need to turn?
443. 39:55Ms. Brantner, you talked about a registry and transparency for the energy efficiency of the data center industry.
444. 40:03make.
445. 40:06I think first and foremost, we have to realize that if we're going to use this consistently, of course, it's going to require the
446. 40:13Data Center industry can't do it alone.
447. 40:15We're data center operators, we're not energy suppliers by design, so that means we're obviously focused on the
448. 40:20exchange and cooperation also on a technical level with the network operators, with the municipalities.
449. 40:27We can tell them what's coming out in terms of waste heat.
450. 40:30But how it is then technically fed in, into the local supply or even into the district heating which technical prerequisites
451. 40:37are required within the grid companies and the grid infrastructure in the energy sector,
452. 40:42that is, of course, beyond our expertise, so it's very important that we get the key players here early on
453. 40:48together and think about it and discuss it.
454. 40:51Where are there obstacles, where do we need to work out solutions together and to that joint dialogue,
455. 40:58in which we must above all also involve the municipalities, because after all, they are the ones who decide on this later:
456. 41:03What is being fed into my network in terms of district heating or data center waste heat?
457. 41:09There will also be a need for dialogue in that as a federal legislature they can't regulate that at all
458. 41:15How the municipalities feed their energy into the grids there.
459. 41:19In that respect, we hope to have early multi-stakeholder dialogue across the board that all stakeholders can engage in
460. 41:28I think that's an important key to success in this area.
461. 41:32What else is important?
462. 41:33Standards have been addressed.
463. 41:35Of course, standards play a crucial role in this area as well.
464. 41:41I advocate that and I think that's quite an important criterion for success:
465. 41:45Let's think about this in a European way, please. There are good standards, there are approaches where we have standards also in the energy efficiency
466. 41:52Area also in waste heat utilization can be expanded.
467. 41:56But let's not make the mistake of retreating to German standards.
468. 42:01The competition for digitization and also for digital infrastructure is a European one and we are not doing a
469. 42:08favor by chaining the German data center industry to a German island standard, and the competition
470. 42:15in nearby foreign countries can operate on completely different terms.
471. 42:19In this respect, this has to be a European approach, and in this context, I would also like to mention normally the GAIAX project
472. 42:27which can perhaps also be an approach to be able to say here I am not only establishing data protection and security standards
473. 42:34and standards for the secure exchange of data between different industry sectors, but possibly
474. 42:40also standards for energy-efficient use of data sharing.
475. 42:44In building data spaces.
476. 42:46So that can also be an approach and it would be a good one in that it's a European one.
477. 42:53Yes, these are, I think, adjusting screws that can be turned and that should be it.
478. 42:58Now I'm looking forward to the discussion.
479. 43:00Thank you very much.
480. 43:07Thank you very much Mr. Süme, thank you very much gentlemen, that was a firework of different ideas, approaches and possibilities,
481. 43:14that we have.
482. 43:18Thank you very much, mrs. brantner, for being there.
483. 43:24take care
484. 43:30to start in a very concrete way
485. 43:32You mentioned that little bit in code, Mr. Messner with certification software, "blue angel for software" might be
486. 43:43that you explain that a little bit and then I would be interested in the reactions also especially from the from the economic sectors like
487. 43:51that is seen, what do they have before that or how far along is that?
488. 43:55We have shown that for an enormous potentiality in the digitalization see in the energy efficiency area to move forward
489. 44:03in terms of climate neutrality to race ahead, circularity to progress and make and so on.
490. 44:08And at the same time, we have to observe that the potentials are not self-perpetuating, so the question is at which set screws
491. 44:14can we turn to move forward?
492. 44:16Have mentioned a number of adjusting screws.
493. 44:18That starts with research.
494. 44:19You also addressed that Mr. Meinel.
495. 44:21That goes on with the companies, the companies need to develop an awareness of that.
496. 44:25And I had brought, yes, the examples of the big companies around the world that don't even see AI and climate together yet
497. 44:31The companies themselves need to look in new directions.
498. 44:33And then, of course, at the end of the day, the question is: What can we do about political
499. 44:41framework conditions to create in order to support the processes, and
500. 44:46what we are doing at UBA in terms of the software issue in terms of
501. 44:51...
502. 44:55Is that we develop heuristics to make data centers comparable with each other, so you create transparency
503. 45:03and by creating transparency, you also create competition. When you have created such transparency
504. 45:08you can then also think about energy efficiency standards. We can send companies on certain paths, so to speak.
505. 45:15in order to then get a corresponding orientation here.
506. 45:17And then I found the point Mr. Rohleder, which you have addressed very important, about which of course in our house also
507. 45:22a lot of thought is being given to it.
508. 45:23What are the monetary incentives to make sure that we're using the technologies in the right direction, and that's where I completely underscore
509. 45:29the point that they made.
510. 45:30If we enacted tax reform that adequately taxed resources and was based on taxing the labor
511. 45:39side reduced, then of course we would get a corresponding direction of technological development, a use of these
512. 45:44technological potentials so what we're doing at our UBA is we're trying to make the regulatory and the standard
513. 45:51oriented instruments that we can use to actually mobilize the potentials that we agree on
514. 45:56Can through appropriate frameworks.
515. 46:00Mr. Rohleder, what would you like to see?
516. 46:03Glad to follow up on the last point.
517. 46:07I think it is quite difficult.
518. 46:10There are studies, ecological rucksack of a yogurt-.
519. 46:13Cup, that's been researched, that was hard enough.
520. 46:16Or
521. 46:16what are the most environmentally and resource friendly milk packaging?
522. 46:19But now they're trying to actually calculate the environmental impact of software.
523. 46:24I think it's important to create as much transparency as possible and generate strong price signals, and when they create transparency
524. 46:31have, that is, as a commercial customer, as an administration, as a residential consumer, I know what costs are
525. 46:37what consequential costs are generated
526. 46:38the use of a product of a solution, a software.
527. 46:41I know that, then I know what does it cost me in the sense what does it cost me for energy or what does it cost me for disposal
528. 46:49And then I behave opportunistically ecologically reasonable and what is missing is the transparency we don't have.
529. 46:56And what is lost by too much intervention in the market is authentic price signals.
530. 47:04So I share that, I'm not in favor of as fine a regulation as possible, I'm in favor of as much as possible
531. 47:11transparent.
532. 47:12We need transparency on the side of the polluters, have talked about data centers, where we are at the moment at all
533. 47:17don't have an overview of what the efficiency standards are across the market, that's where we need transparency, because then we need
534. 47:22we need clear price signals.
535. 47:23We should not make the mistake now in this sector that we made when we invented the circular economy. At
536. 47:28the invention of the circular economy, we started the waste management sector.
537. 47:32Have built a very expensive waste disposal and recycling system, while the mountains of waste in front and the material flows further
538. 47:39have grown. So instead of starting at the front and setting the price signals there, we've started from the back.
539. 47:45So we should see that we develop lean regulatory systems, but so far we have a high level of non-transparency.
540. 47:51in these markets,
541. 47:52that we're talking about and we need to eliminate.
542. 47:56If I times this problem, if we say to our young people: here think please in the development of software and
543. 48:01IT system also this issue of energy consumption with, they actually have no real chance to see what they do
544. 48:11have to or what's good and what's bad?
545. 48:14That's a very big problem.
546. 48:15Of course software on hardware, operating systems, the technology is right on top of it, then it runs and that abstracting
547. 48:22from what's under there. Do you see any chances there with any labs with any approaches?
548. 48:30We also did yesterday with, coming up later as a speaker from VMWare also with the virtual machines where you might be
549. 48:39then also such things can test, that the developers can really try hands-on and somehow also structured say:
550. 48:49yes, that's with these two ways of doing something, the better one to achieve efficiency in terms of energy demand.
551. 48:57We're doing studies right now where we're looking at whether there can be metrics, heuristics that we can use to guide together
552. 49:03CAN.
553. 49:04You can be sure, there is not something given, which is based on a study of the UBA, but at the end of the day
554. 49:09we want to throw a stone in the water to be able to think about this question systematically so that companies can
555. 49:15or developers, so enterprise level and the developers level have any orientation at all in which they can
556. 49:21can align.
557. 49:23And that is, after all, the basis for subsequently making good corporate decisions good investment decisions. For
558. 49:28who give another example: We are
559. 49:30For 2-3 years in the very structured dialogue with the industry-
560. 49:324.0 platform and have tried to understand where the disruptive potentials are via use cases?
561. 49:38And here's the emphasis again.
562. 49:41What's important is that the companies themselves start looking this way, so the industry-.
563. 49:444.0 platform did not have sustainability and climate neutrality on the screen at all in the first years of its existence, the
564. 49:50was supposed to be a waste product, but digitization doesn't have climate neutrality as a waste product, you have to have climate
565. 49:56Neutrality and energy perspective as part of the business model.
566. 50:00These perspective changes, they are important, they can come from the companies themselves, there has been a change now
567. 50:05there has been, that's very good.
568. 50:07But of course, they can also be supported by policy accordingly by energy standards, by CO2 prices and so forth
569. 50:12can be supported.
570. 50:13And that interaction, we need that now. Now I have described a little bit our needs
571. 50:17in the development of IT systems, then our colleagues from the GFZ are in the use of these IT systems.
572. 50:25So you can't ask them to think about whether this model is more profitable this way or that way, but
573. 50:32they must be able to rely on the fact that these things that are made available to them also have a good balance sheet
574. 50:42How can you help with that?
575. 50:43Because so asking the GFZ to care... they are fully challenged to develop the models in the process
576. 50:52and to verify now down below to take care of the energy efficiency of the data centers,
577. 50:57I think that would be overkill. If I can just say something about that, it would be a misconception to think that we have a lot of commercial
578. 51:06software. Then it would hit exactly that problem, with commercial software that is used operationally
579. 51:13That's where we have to rely on others. All the things that I presented earlier,
580. 51:16But are basically based on software that is developed within the community, so partly in our group
581. 51:22but also in cooperation with international colleagues and there it is of course even more difficult to create a standard,
582. 51:30because these models are always in flux, they always have to be adapted and how to introduce a standard is not clear.
583. 51:37to me is a mystery, especially if you also demand this from your doctoral students, who are only paid for their work.
584. 51:43what comes out and not how it comes out and you need an international solution.
585. 51:51What would certainly be a completely wrong perspective would be to say the end users now for now , because you use
586. 51:58components,
587. 51:58You use software, so now to impose that on the end consumer and end user, the problem of carbon neutrality
588. 52:06at the end of the chain, so to speak, is complete nonsense.
589. 52:10So if we look at that, if we look at that sector as a whole, so the examples that they brought show how important
590. 52:15it is that we harness the potential of AI, machine learning in that context so that we can get better at understanding
591. 52:21of the earth systems and the climate systems in the interaction with the social system, so we should be
592. 52:27absolutely continue there, because they asked: Is it worth it? Is it worth it?
593. 52:31My answer would be: of course it's useful. But we need a framework for the whole sector.
594. 52:38Are three core elements central, the first is: so that the data centers that are at the beginning of the of the data processing.
595. 52:45so that the data centers can become carbon neutral,
596. 52:47first of all, it is important that we are dealing with 100% renewables, that is the basis that the energy,
597. 52:53that is consumed is also carbon neutral. that the first point. the second point.
598. 52:57Franziska Brantner had also rightly addressed that.
599. 52:59Energy efficiency remains an important point, because we have such high requirements for the expansion of renewables, if
600. 53:06they think about all the sector coupling potentials, that it's not arbitrary how much electricity we use, so
601. 53:12in addition to the renewable dimension towards 100, we need to continue to have energy efficiency and in data centers we have seen
602. 53:17we don't even have an overview of how much energy is being consumed there. The data centers have, that partly still themselves
603. 53:22not even developed for themselves.
604. 53:23Energy efficiency, there you can work with standards. Comes a 3rd element to that, if you look at the whole sector and
605. 53:29the data centers first of all, we also have a high resource consumption in the area of
606. 53:35digital innovations, and that's why circularity and circular economy is of great importance, so that we can get through
607. 53:42the efficiency potentials that we tap into, in terms of energy, we don't subsequently have rebound effects via the increased
608. 53:48resource consumption, which in turn leads to energy consumption, so renewable energy, energy efficiency, and circuli-
609. 53:55tat in terms of resources that's the guiding principle that we should be moving towards. maybe once I can go to the first one
610. 54:01point briefly, because of course that is indeed one of the biggest challenges for the industry.
611. 54:06We all know that at the moment there are not available these renewables to the extent.
612. 54:11Right now I think it's 42%.
613. 54:13about less than half of the electricity share is just renewables.
614. 54:17Even if we expand that to 80% in the next few years, that just doesn't run in parallel with the targets that the legislature is setting
615. 54:24to the ICT industry and also to the data center operators, namely to be climate neutral by 2030 according to the Green Deal.
616. 54:31completely.
617. 54:32The federal government says in its coalition agreement even new data centers must be operated climate neutral from 2027/28 already
618. 54:41will be.
619. 54:43And so I'm wondering how that's going to work then, if we're going to be using renewables particularly in the short
620. 54:49period until 2027/2028, we can't even expand as much as we need to for the required consumption of such renewables.
621. 54:59Energies by the data center industry will require.
622. 55:02So, we have a tradeoff there somewhere, and I'm curious how we're going to resolve that first of all, we need to
623. 55:08solve it in such a way that at the same time we ensure the competitiveness of the data center industry, of the digital infrastructure
624. 55:15overall, the business models, the innovations that we need to continue to develop in order to remain competitive, in the eye
625. 55:21and not reduce ourselves too one-sidedly only to the achievement of such goals and completely disregard what is otherwise
626. 55:29is still happening in competition in this industry, so that's a big challenge.
627. 55:34My response to that would be, we have climate targets, and I hope we have a consensus there, because if
628. 55:42we don't submit those climate targets,
629. 55:43That's what they've shown,
630. 55:44After all, I come from the social and sociological sciences.
631. 55:47Climate science, if we don't meet these climate targets, we're going to get into really rough waters.
632. 55:51That's why we have these challenging target systems that we're working hard on now and that's where we have to take responsibility now from the
633. 55:58perspective of different actors and that a responsibility lies with whom the framework conditions make
634. 56:03I have just said some of the things that would be necessary, in which direction we have to go there, and while that is happening
635. 56:08But the companies but also public institutions have to,
636. 56:11I'll say something about my own institution in a minute, we also have to get going, while the framework conditions are
637. 56:15are still emerging, so to speak.
638. 56:16That has to run in parallel.
639. 56:18You should not wait as a company.
640. 56:20I think we have discussed Mr. Rohleder a time or two.
641. 56:24They should not wait as a company until the regulations are there that push them in the certain direction
642. 56:29but they can, after all, approach things on their own initiative and in the Industry 4.0 platform we have this
643. 56:35discussion now from the corporate perspective, they are taking that on board, they are trying to make that part of their business models
644. 56:40to make.
645. 56:41When I just said in my presentation only a small percentage of large companies are bringing AI digitization
646. 56:47with climate neutrality in their own companies so far at all together, then there's still a lot of
647. 56:52Scope to go ahead and run things.
648. 56:54So, we haven't reached an optimal standard in the companies already and we have to, that's a little bit more
649. 56:59Finetune through incentive structures, but we have to develop incentive structures and in parallel, the companies must
650. 57:04part of their business model, and all the data shows us there's still a lot of room to go up. so I don't
651. 57:09just finger pointing.
652. 57:10We want to build a climate neutral, public administration, build climate neutral research and that's where we work in all
653. 57:16their institutions at the universities and also at the Federal Environment Agency on strategies on how we do that.
654. 57:21And it's a helpful process here.
655. 57:22I don't know how that works for you at the university here.
656. 57:25At UBA, we have identified the 5-6 areas where our resource consumptions and energy consumptions are
657. 57:30in particular arise and have now developed roadmaps,
658. 57:33How we're going to get down there.
659. 57:34If you don't look at it there, if you don't create the transparency, nothing happens, then you are in the routines
660. 57:39of procurement, so to speak, we have to take our different responsibilities on these different
661. 57:45levels. Gentlemen with an eye on the clock, although there is still a lot to talk about, I would like to conclude this first part of our
662. 57:53Conference, problem description would like to conclude.
663. 57:58Mr. Messner, Mr. Süme, Mr. Thomas, Mr. Rohleder, thank you very much for your inputs and the discussion.

To enable the transcript, please select a language in the video player settings menu.