Essentially we have three asset classes in digital infrastructure. So when people talk about digital infrastructure, they essentially mean one of three things, either the mobile towers or the fiber networks, domestic and international, or data centers. Mobile towers are very interesting from an investment point of view, and they carry a lot of value, and that's why they have been a focus of many new companies, ventures, M and E and acquisitions and licenses. Fiber networks are, of course, very critical. We covered this in the politics piece of the discussion and somebody in cable networks in particular, carry a big weight in delivering the value created within the digital ecosystem to their consumption. We all use facilities like zoom. For example, without submarine cables, you wouldn't be able to listen to me or hear me or see me, because wherever you are in the world, there is a cable system that carries the traffic from the Zoom platform to where you are, and from my where I'm sitting, to the Zoom platform, and then the critical piece on, or the center piece, or the epicenter of the digital infrastructure space, is data centers. This is where the bulk of the economic value creation happens. This is where all the digital value creation is centered around. And today we're going to talk a lot about data centers and their power consumption, given that today is a discussion about power. So these are the three access classes in digital infrastructure. Today's focus would be data centers. I'm going to cover the model we're adopting for this discussion follows an industry model published by infrastructure masons, though I have adjusted the model a little bit to fit the purpose of this discussion. So essentially, we have four pillars of discussion. If we want to adequately cover digital infrastructure, you have the politics power, the planet and people. We already covered the politics piece in our second episode back in January. So today we're going to cover the third piece, which is power. And the next session will be around planet, which is the environmental and the sustainability aspects of digital infrastructure and the facilities. And the last part would be the people, which is the impact of digital infrastructure, the skills, skill gaps, the training of people. And of course, inevitably, the impact of these investments, especially in AI, you know, what people do with their careers and so on. So that's the game plan. As I said, this is number three. So we need to jump into discussion about power, which is the focus of today's discussion. So please join me in this journey. Today we are talking about power. I'm an electric engineer by trade, so I love nothing more than talk about power. But I understand that we have a mixed audience, so not everyone is comfortable about these technical aspects, so I need to invest just a couple of slides to establish some basic definitions. What is it that we're talking about, so that we can all continue on the journey the same understanding? So this is the simplest way I can explain this discussion. Power is the rate of flow of energy. So this example is a straight forward one. You have a who's filling a bucket, so the rate of flow of the water is the power. So that's measured in watts or kilowatts. So power is a flow state. So it's a rate of flow of energy over time. Energy itself is the water in the bucket. So when, when you're sort of thinking about power and energy, they are sort of two sides to the same coin, but they are not the same thing. Energy is what you start with, and then the rate of flow of that energy over a transmission line or over a particular appliance or a utility. That's the power, that's the power. This is going to make it a little bit more, hopefully clearer. So again, this is another example. You have a bottle of wine you're pouring, you set up some wine, so the amount of wine in the bottom, that's the energy and the rate at which you're filling the glass, that's power. So power is measured in one of two things you can if you go back to your high school physics classes, if you remember those, there was a unit called Jool for energy and essentially a wat joule per second. Again, you see a watt is a flow of energy per second. What or power is a flow statement. However, the utility grid operators that the companies to which we pay every month for our households, at our office energy consumption, chose to measure energy by something called a watt hour, rather than messing around with Jool and so on. So what So one hot hour is 360 sorry, 3600 joules, you have kilowatt hours, megawatt hours, gigawatt hours and terawatt hours. Now it's not what per hour, it's what multiplied by hour. So do not say it's what per hour, it's what our simply because of how the units are set up. So now we know that power is the rate of flow of energy. Power is measured in for the most part of this presentation, in kilowatt hours, megawatt hours and so on. Interesting thing that you will notice about all the information I'll share is that electric engineers do not know anything about terawatt hours. We as telecom engineers are comfortable after Terra comes Peter and then exa and then zeta and Zen Yota. Electric engineer, for some reason, stuck at a terawatt. So you will see 1000s of terawatt rather than using petawatt. That's it. That's all you need to know to join us today. So and you're gonna hear a lot of about terawatt hours and megawatt hours and so on. As an example, these are typical household appliances that you and I use in our households. And whenever you buy a box of these, there is on the box, they tell you the consumption, the power consumption of this unit. So you can tell when you are buying if this is going to be a heavy consumer of energy. Is going to add dollars and cents to your energy bill or not. So this is an example the skin wood air fryer here, which consumes the it has a power rating of two kilowatt. If you let this running for three hours, that means that you have consumed six kilowatt hours of energy. That means that you will pay your electric and utility operator at the end of the month equivalent of whatever rate they charge you per kilowatt hours. But you multiply that by six, so that's what you need to know. And instead, in this case, we're going to be talking about servers, about GPUs, CPUs, you know, electrical components used in data centers. We're not going to be talking about air fryers, but at the same energy, the same principle of energy consumption, that let's jump into the meat of the presentation, which is the discussion about power. And hopefully we'll be able to ease ourselves into it in a manner that makes everyone interested. This is an interesting statistic. This is, again, we're talking about digital infrastructure and data centers in particular. So this infographic is an interesting one. There are 25 countries on this infographic, the largest one being the USA in terms of number of data centers, and the smallest one being UK. So that this does not covered every country in the world. The cut off is at 25 the ranking of between one and 25 and then every thing else is bonded here. So the rest of the world, this includes UAE, Saudi Arabia, Egypt, Nigeria, all the other countries in the world are this part. And the top 25 countries, in terms of number of data centers, are included by name in this on this infographic, I just want to extract some information for you so that you understand, you know the lay of the land, and some basic information about the data center board. So the number one ranked in the in the world is USA. USA roughly has half of the data centers in the world. That essentially means that it is the prime target for all the research studies, all the intelligence, because USA is such a big market that we need to actually learn from it and copy as we develop infrastructure across the world. Number two is a bit of a surprise. Number two is German. Globally, a second highest number of data centers exists in Germany. Number three is the UK, but four is China, and number five is Canada. So what does that tell us? It tell us that the US is way ahead of everyone else terms of the number. I told you that the German is number two, but look at the numbers. USA is 5300 and Germany number two is 500 so it's not even close. It's not like Germany has like 3000 or 4000 data centers. Us is such a huge market that every bit of consulting or intelligence you see online has to focus on the UAE. So USA as a priority. So that's the number. This is, by the way, March 2024, and the numbers, of course, keep changing, but the USA lead globally Make, make no mistake about even China, by the way, the distant fourth player in this. So that's why the USA is such a prominent figure, and that's why you see President Trump talking about data centers and AI. That's why you see everyone is concerned about this, because it's such, such a competitive advantage for the US in terms of their control over the digital landscape. People who have joined us in the past two sessions already saw this. I am getting a treat for our new joiners. I upgraded this, sorry. Updated this with the figures for January. So what this says is that this is the total megawatt for all the data centers in each market. For example, in London, you see that the total power in megawatts. Now you know what a megawatt is. The total power in megawatts for all the data centers in London is 996 in Beijing is 1864 can clearly see that Virginia, not the US, just Virginia, is the highest market or concentration in the world with 5536
megawatts. That's 5.15 gigawatts. And immediately you can see how the US is such a massive force in this, because on top of Virginia, you have New York, New Jersey, you have Charlotte, you have Atlanta, Texas and so on. So the combination of all the markets in us is what gives the US such a formidable force in the digital landscape. And you can immediately also see that the global South, if you will, Middle East, Africa, for the most part, we do not have enough to be ranked in this top 40 list global so the key point here is that everything in data centers is measured in megawatts, and our session today is about power, which is measured bigger, so now we understand what when I said power is the most important pillar, because everything is measured in power, everything all the other stuff is just ancillary, is just enabling, just supporting, everything is empowered. I'm going to give you one more statistic about the global landscape of power. This is from Cushman, wig field. I want you to ignore the vertical axis. It doesn't concern us right now. What they try to do is to actually, practically represent the amount of power in each market by the diameter of the circuit London. You see the assessment in London there. This is Frankfurt, Dublin, Mr. Dam, Paris Miller, because these guys are the largest. They're considered primary markets or powerhouse markets. And then you have all the other markets in EMEA ranked here, and you see familiar names, you see Johannesburg, you see Dubai, you see Abu Dhabi, you see Tel Aviv, Cape Town, and so on. Riyadh here as well in Saudi Arabia and so on. The smallest market in this list is Istanbul. And you can tell from the horizontal axis that Istanbul, total megawatt in terms of data center capacity, is around 160 170 megawatt. Anything smaller than that does not show on this map. So again, you see how the world is in a race. Everybody's trying to actually put together as much power as possible to support the digital infrastructure development plans, especially when we talk about AI. One thing I need you to make a mental note of is this flap D thing. So flap D stands for Frankfurt, London, Amsterdam, Paris and Dublin. These are considered the five biggest market in EMEA or Europe. And I also need you to make a mental note about the Nordics. If you see Helsinki here, Oslo, we're gonna touch on the Nordics and a couple of slides for now. This is how the EMEA region of of which Egypt is part of, stacks up in terms of the total mega megawatts for power. So just exemplary to to understand Virginia, of course, from the last slide, we know Virginia is not 23 gigawatts. It's 5.5 but here they included all the future projects in the pipeline under development, so that when Virginia completely develops, they will have 23 gigawatts capacity. That let's jump into some nice comparisons. This is interesting. The year 2020, if you remember, was the year of the COVID or the Coronavirus pandemic. And this was a milestone year for data centers, because for the first time ever, we started using more and more of these platforms that are housed in data centers. We started using Zoom calls more. We started using teams. We started using online education and online entertainment like Netflix and streaming service and so on. So it makes sense to measure everything. To look at 2020 and the interesting statistic here is that this is the total consumption of a country like Egypt in that year in terawatt hours. That's the total energy consumed in Egypt. And this is data centers. So you can see data centers globally, and there is a range there, depending on the margin of error and the scope of data data centers globally in 2020 consume more energy than Indonesia or even the UK, or comparative tool to South Africa or Egypt, I should say the other way around. So the point here is that this statistic tell you that data centers are becoming more and more energy hungry, and they're becoming more and more integral to what we do, whether business, education, entertainment, whatever it is. So that statistic tells you that more and more focus is focused on the energy consumption, which is our focus today, of data centers. This one is also nice, same thing, but this one is somewhat, somewhat more sort of controversial, because people do not really like the fact that a company like Google, the aggregate or consumption of all the facilities of Google globally is more than a country of jewel. So here you have the total megawatts, sorry, terawatt hours consumed country or per company. And here that the population. And, for example, Nigeria. And this is 2022, figures consumed. 32 terawatt hours in a year. Google consumes 24 Microsoft comparable meta, which is the owner of Facebook, Instagram and WhatsApp, consumes 12 and happen at the very bottom? What business tells us that if this was a country, it would consume more energy than Jordan or Iceland or Ghana or courage? Again, emphasizing the same point, data center consumption now is becoming a topic of international focus, and we need to be aware of that as we plan our networks, and also as we plan our investments in network expansion and so on. I
hope you're still with me. So
what's happening right now? What's happening is a transition from an old regime of data centers into a new regime. So we travel transition from from the typical retail Co Location to sort of data center around you know 1990 not 19 2016 2017 that's a typical data center environment into the high performance computing and AI right now the transition translates into more power consumption per rack. So you can see the rack here, a typical rack which consumes six kilowatts with national 12 kilowatts. Now we're moving into the AI heavy infrastructure, which uses anywhere from 50 kilowatts to 150 kilowatts. So as we migrate from the traditional retail Co Location centric data center environment into these AI work horses that you hear CEOs of open AI and Nvidia talk about, the inevitable result is get we're gonna need more power. We're gonna need more power, and also more infrastructure, as I'm gonna explain in a second within the center specialist infrastructure. So that's what's happening. And this is the transition that all countries have to go through. We cannot escape this. So what does that mean? What? What? What does that translate into you? You're gonna face, actually, three design challenges as you migrate from your run of the mill or clean vanilla data center into specialist AI data centers. First one is going to need more power density. That means that granite racks that have 120 kilowatts per rack, as opposed to the traditional racks running enterprise applications, also going to take into account some redesign requirements, mainly cooling. These things run extremely hot, so you need specialist cooling systems to actually cover the cooling or the heat from these AI data centers. Also going to need to adopt a different location strategy, which in a sense is an opportunity for emerging markets, because right now all the primary markets are saturated. So right now, data centers are moving to remote markets seeking energy. So anybody who has energy is a prime target to accommodate these data centers, because the primary markets like London or Paris, Frankfurt and all the major cities are almost saturated. We cannot find energy to run these things. So this transition is ongoing, and this provides some opportunity. If you are in an emerging markets like Egypt, Ethiopia, Nigeria, Kenya, you can actually play this location strategy thing and try to attract investments into your own market. I have to spend a couple of slides on this whole GPU thing, because I know we keep hearing about what Nvidia is doing, what the MD is doing. But you know, we need to put some context to this. So why is it that these processors consume so much power? What can sort of, what can happen next, if you would. So I need you to understand this slide is the architectural difference between a tradition and CPU or laptop, right right now, if you're watching this on a laptop or a desktop, has a CPU, probably an Intel Core, I seven, I five, whatever it is, or an AMD Ryzen and so on. The typical c2 architecture has multiple cores. I'm simplifying here this the ALU stands for arithmetic logic unit. So these are different cores. This is what the number crunching and the animation and the Excel calculation, everything happens here. For that, you need some control logic and some cache to copy the information from memory. So that's a typical CPU design. Number, of course, is usually low. So when you go to buy a laptop, they tell you, we have like, eight cores, 16 cores, whatever it is. Now the GPU architecture, by comparison, is designed for parallel processing, so the logic there is that we need to have smaller cores, but running parallel. The G comes from graphics. These things were originally designed to handle graphic rendering in games. And in a typical game, when you have a scene where you have five light sources and you have materials like glass or gold or metal or rock, you need on the fly calculation so that you can see how light reflects over these materials, and you can make decisions in the game based on if you're walking on water, if you have glass in front of you, and so on. So the G comes from graphics. However, the logic is the same if you want to actually compute large language models for an AI application. So what you need to see immediately is that, look how many cores are there. These are green ones. These are the cores. And here you have a limited number of cores, so GPUs operate in parallel, and they're smaller in size, but you had a massive amount of cores, 1000s of cores, within a GPU. And this is what Nvidia builds, or other companies, and I'm going to show a couple of examples. And because of this, you're stacking so many cores in such a space, the consumption of power ramps up considerably compared to these CPUs. This is a typical example of an offering from a company called AMD. And again, they are fairly known in the processor business, and AMD manufactures both CPUs and GPUs. This particular chip is called an APU. This is their marketing term for a chip that combines the CPU and GPU doesn't concern us as much. What concerns is this one 750 watt. I needed to imagine that this single chip consumes 750 watt. That's as much as a small water heater or a wood or a water capital, if you would. So the power consumption becomes an issue with these things because I had a massive number of transistors and a massive number of storage and memory on this chip. Now we're going to get into the nitty gritty stuff of the favorite of everyone in video. So Nvidia, last year's model is the Blackwell chip. And the Blackwell chip, if you can imagine this, it consumes 1.2 kilowatts of power. This is insane. Nobody could imagine this five years ago, that you would have a single chip that consumes so much power. And for, for, for you to understand the complexity of this black wheel chip, it has 208 billion transistors there. The black wheel chip is made up of 100 plus layers, a layer of like a floor in a building. So if you can imagine a skyscraper of 100 floors squashed into nano, nano meters of dimensions. So this is the black will chip. Of course, back in March, they launched the black will Ultra doesn't really matter concern as much. Here I want to show you something that's even more scary, which is this chip. This is the maybe 200 so this is one black Well, plus one black well. And they have a grace CPU, so similar to AMD, having a CPU and a GPU on a on a card, Nvidia is doing the same. What concerns us here is that this single building block, this is like the four by two Lego if you play by Legos, this is like the standard building block of all the infrastructure I'm going to show you. Single thing consumes 2.7 kilowatts. So just so without integrating anything else, just a single card consumes 270 2.7 kilowatts. Next one is a very nice video. It's going to give us time to give us time to time to sort of slow down the presentation a little bit. This video comes from wi Nvidia, and it's going to show for you the entire Nvidia AI architecture. Um, try not to gasp at the end, because typically when I saw people this the last five seconds, is people gasp because they cannot believe so much power in so much infrastructure. So this is the video that I'm going to show you. And it actually starts from building the Blackwell GPU, which is the workforce of the modern data center, and then we're going to build up from there to build massive Data Centers. You
so now I get an idea about how massive these things Are. This is the NVIDIA architecture of AI data centers. And as you saw, we're starting from a single processor and then ramping up into this one. The final one is 32 GPUs. This is the stuff that Microsoft and x ai and Google are buying. And you can imagine that this consumes a huge amount of energy. The numbers there are just really scary. Let me show you how scared you are. Look at this. This is a post on x by Sam Altman, the CEO of open AI, saying that the first this is the MVL 72 that the one you saw this track consumes 120 kilowatts, and he is. He's posting on january 31 only months ago, at the first full eight track, GB, 200 ml, 37 now running at AZ. Now, the funny part is that as massive as this assumption is in video, in March, just released something that consumes even more power. So please understand that this is very recent. The Open AI just installed these back in January, and already they're out of date. Already there is a new one. This number is scary. This number is really scary. 120 kilogram per rack. Very few data centers in the world right now can deliver 100 to 120 kilowatts to Iraq, terms of electrical power. So that already tells you that we're already in trouble as far as power is concerned. The next video is a very nice one, and it comes from the CEO of Nvidia back in March 18, on their GTC conference in San Jose, California, and it's a really nice, interesting way of looking at things.
One big idea is that every single data center in the future will be power limited. Your revenues are power limited. You could figure out what your revenues are going to be based on the power you have to work with. This is no different than you know, like many other industries, and so we are now a power limited industry, our revenues will associate with that? Well, based on that, you want to make sure you have the most energy efficient compute architecture you can possibly get.
Again, this highlights why we have this discussion. We're talking about power because power happens to be the limiting factor of the development of all this AI infrastructure and data centers and digital economy and so on. Everyone is limited by power. Power is definitely the most important thing if you want to be serious about developing infrastructure within the digital economy space, and nobody has enough power. That's the sad reality of our rule today, nobody has enough power. Let me show you a proof of that. Look at this. This is posted March 27
10 days ago, days ago. So
look at this when, when OpenAI released their newest image generation engine, the one that everybody was posting cartoons online about, and Alton was complaining that our GPUs are melting. That means that the GPUs were burning so much power, he had to actually put a limit to how many generations you get per day. So the free customers would only get three per day, because GPUs are melting. When he says GPUs are melting, he's saying that I'm paying too much for electricity, because essentially the electricity is that he needs to actually run the GPUs has to come from a provider, and that means that he's gonna his cost is gonna ramp up. So the point here is that he's managing his cost by saying we're limiting uses, he's managing his course because the GPUs are running over time and burning too much power, and that means that his costs are running up. That's the reality of the business. The business is based on GPUs. This is a very bad joke, but unfortunately, it's real. Back in March 18, Nvidia released the next generation, which is going to be available in the second half of 2027 a single rack will consume 600 kilowatt now it's really unclear how anybody can come up with so much energy. Only thing we can have to think about is renewable energy sources. And towards the end of the presentation, I'm going to talk to you a little bit about nuclear, because these are the only sources that can give us these insane figures. With that, you understand we have a problem. I established the basic baseline. Everybody's trying to figure out how to actually satisfy the appetite of the GPUs. Nobody has a really solution except to develop new energy sources. So let's look at some figures. The next section is full of charts. I'm gonna jump between them. I don't want to bore you with too many charts. However, when you get a couple of the presentation, you'll get all the charts, as well as all the sources I use in my research to build this stack. So fact that I'm gonna jump a slide of two doesn't mean that you will not get access to them or the supporting research. I'm just cognizant of the time and we need to cover as much material as possible. So let's size the problem now. So this comes from KKR. So those of us who are old in the internet business remember the internet one, 1.0 era. This is the beginning of the internet. This is the dawn of the internet. And the period was from 99 to 2006 investment firm KKR estimate that the total power infrastructure requirements for that period was around two gigawatts. So all the energy requirements globally were around two gigawatts. And then we go into Internet 2.0 this is the period where we saw the cloud platforms, the streaming platforms, everything that we become accustomed to. This is in the period Oh, six to 22 and this was an eight gigawatt opportunity. Probably can can guess where I'm going with this. This is what the ninth method we're looking at for AI and hetero computing. This is, this is going to be a 30 gigawatt opportunity. But look at the time frame. This is like 1617, years. This is like, what, 1213, years, just only five, six years. So we're gonna need four times more the energy that we needed to develop all the cloud platforms and all the streaming platforms and so on. But we're gonna need it in six times. So that's half six years. That's half the time. So that tells you that we're facing a very unique challenge globally. We need to come up with so much energy to satisfy the requirement of these massive AI systems. This statistic, this is from the US. And if anybody's asking, what the US? Because the US is the largest country that has data centers. As I show you, I show you, showed you in the first infographic, and the projection is the data center power consumption is going to double or triple by 2020, here's a question. For you to think about. Do you know any other industry that's going to double or triple in the next four years? Probably not. So this is a very unique industry globally, airlines, tourism, petrochemicals, agriculture. Nobody has has this projection in terms of public this comes from the US. This this is about energy in the US. And 40 years of energy as captured in this slide. In this period, you had growth because of you had more PCs, more computers, more equipment. So digitalization drove up the consumption of power this in the US. This is measured in 1000s of terawatt, terawatt hours. So between 1990 and 2000 in the US, it was a good point to be an investor. In energy, because you you had increasing demand. Here. The demand started to flatten a little bit because of the energy efficiency wave swept the industry. So the growth was dampened a little bit by that energy efficiency. And here's comes the dark period. This period is flat, by the way. Here you can see a 2008 economic crash, the crisis, as you did. You see I did there in this period, actually all the way between 2020, 2020, 3023, the curve is almost flat. Essentially, you're ending the decade where you started. Means that investors were not really keen on putting any more dollars in power generation infrastructure because there was no demand for it, boom. Look at what happened here. 2022, 2023, and then everybody's racing to generate more power. Okay, if you if you want some triggers. Growth Factor here was 2% annually. It was a 1% because of the energy efficiency here is practically zero. It's a flat curve, more or less. But here we're looking at 4% this is the compound compound annual growth rate for energy. So now money is flowing back into energy. My people are investing in energy because there is demand for it. Of course, in the US, especially under the current president, everybody's concerned about the carbon footprint objectives, because people are going to say, You know what? I need power. I don't care about sustainability and so on. That's a concern that many people in the US have right now, but that's the US energy sector in one slide over 40 years. Again, we're focusing on data centers, and this is the consumption of data centers in terawatt hours between started 2014 up until here, things were sort of nice and easy maximum 4.4 so that means that 4.4% of the total energy in the US were consumed by data centers. And then boom, the projections indicate that by 2028 you're going to be consuming in terms of data centers in the US, between 6.7% and 12% so you can immediately see that the jump came around here, which is that chat, GPT launch date and so on. So you can see AI has a massive impact on the energy and the power consumption in the US. I study again, same logic. These people are trying to estimate. This is the Electric Power Research Institute. They're trying to estimate the growth factor for data centers. In this case, they had four scenarios, low growth, moderate growth, high growth and high growth. And they're trying to estimate how much percent of the 2030, electricity consumption will be consumed by data centers. And you see the figures between 4.6% and 9.1% the US is an extreme case. Globally, most countries will not be near these figures, so you have no reason to be concerned about consumption of data centers out of your National Grid, because you do not have the same level of massive infrastructure of the US. Still, it is an interesting point to study. Going to show you this side is scary if you don't see it scary. Let me show you how scary it is. So this is done by McKinsey. This was done September last year, and this is again, the energy consumption in terawatt hours of data centers across the period from 2023, 2020, search looks nice. Then you consider growth rate. McKinsey estimates that the growth rate is 22.3% what does that mean? It means that data center power consumption doubles the 40 months. So you went from 147 292 that's the one, the first doubling. And then you went from 292 to over 600 there's not a single industry on the planet that doubles every 40 months. Nothing doubles every 40 months. Of course, this is not sustainable. I mean, if we extend this towards 2040 nobody expects it to be double every 40 months. But for now, this is, seem to be the consensus of the industry information. I'm going to jump it. This is a nice slide. These are the four scenarios I showed you in a second. The high growth, sorry, the higher growth, high growth, moderate and low growth. And this is the McKinsey study to the gold Goldman Sachs study. So what this bar graph shows us is that this is the baseline starting from the 2023, year consumption. And you're trying to estimate by 2020, 30, how much growth will happen. The figure here tells you how much percentage of the total US risk consumption in 2030 is going to be consumed by data centers. So again, for the study I showed you, it's between 4.6 and 9.1 McKinsey is very bullish. It's going to sit between 11 and 12% and this is a 660 hours I showed for the moral of the story here is that we expect exclusive growth in data center consumption in the US, as well as globally. I'm going to show you a global slide. This is a global slide by McKinsey focus on AI. So look at this. So McKinsey estimates that related AI workload, the stuff that we use, chat, GPT and all that the stuff is growing at an annual growth rate of 39%
39% annually. These are insane figures. The other workloads, which is the cloud and all the other infrastructure, is growing at 16% annually. So on average, you have a total growth rate of blended 22% this is a 22 figure I showed you about so understand now the US is facing a situation that no other country on the planet is facing. They need to figure out where to come up with that energy. I need to come up with that energy fast. The US is to maintain leadership of the global AI race against China and Europe. I'm going to jump to a global perspective now we're meeting the closure of the charts. So bear with me, please. I know this is not the nicest thing to look at, but it's unfortunately very necessary for you to understand the crisis that we globally look at. This is global. So this is the International Energy Agency estimate, and this is IDC study, and you can see that globally by 2030 situation is not as extreme as the US. I mean, we expect here, in this case, to be consuming 3% of the total electricity consumption to data centers. This won't be estimates 13% and so on. So you get the idea that there's going to be tremendous growth, but it's not as extreme as the US. And for comparison, I'm going to show you something that actually Nick puts these things in better context. This is the same study. So this is 2030, we're looking at a pie chart of data of the different uses of energy. As you can see, data centers is a tiny sliver, tiny slice of the pie. Compare it with space cooling, which is the AC or space heating, which is the heating systems used in cold countries like Europe and so on, heavy industry electric vehicles, and all of a sudden you get the notion that data centers are not that beast. You read online, people tell you the data centers are going to swallow electricity. They can actually put people out of business. That's not the case in terms of energy because if you run the analysis properly, you see that globally, it's not going to be that significant compared to cooling or heating or other energy uses. So that tells you that data centers will have massive growth, but they will not be a dominant factor, as opposed to other energy uses. First thing we're going to show you is the venue chain, or the supply chain, of of energy. We keep talking about energy in terms of generation, but energy needs to be delivered into a data center via a combination of infrastructure assets. You need to have the media voltage which connects with the grid, low voltage for distribution within the building, backup power, central power supply, or UPS and so on. But what happens to the supply chain if no energy is available? Guess that no constraints showed you that it used to be around between six and 10 months lead time of if you can order a generator at 2019, now we have to wait six, between 10 and 16. All the blue bars there is the extended delays or lead times for ordering all these different kinds of equipment. Why? Because manufacturers of digital, of electric and equipment, cannot manufacture them quickly enough. So everybody's compressed. Not only power generation utilities are compressed, but even companies who manufacture switch gear, transformers, distribution panels, everybody squeezed. Good news is that we're done. Now we're gonna get into the nice part of the presentation. Hopefully, I'm just gonna check my chat and it before if anybody has any questions to make sure I do not leave anything behind. No.
Okay, fantastic. Okay,
we did not miss any questions on the chat. Fine. So now let's go into the part of the presentation that's going to actually tie it all up for you in abundance, so that we can link what we discussed the last session about politics and this one, and sort of try to actually put some some intelligence at your hands. The next slide comes from Mr. Macron, who happens to be visiting Egypt these days, the French president. And this is a very nice clip. I need you to listen to that clip carefully, because it shows what the head of state in this time of intense international competition is talking about for his country and how he's planning to put his country
first. This is how to as much more data centers, and we are back in the race now. Why? Because, in order to do so, you need chips. We start to accelerate the programs. France is one, is part of the European countries manufacturing chips. We will build new partnership, partnerships with American and Asian players. In order to manufacture more ships and secure our capacities, you need available spaces and energy. And here, I want to showcase what is a great advantage of France, we have available land. We worked very hard with the local authorities, and I want to thank the regions present here. We identified 35 sites available and and the most important for me, we have low carbon, pilotable, available energy. And this is unique, unique if you compare with the rest of the world, because we have a model of production based on nuclear energy, which means that more than 75% of our electricity is produced by this nuclear energy. It's pilotable, it's low carbon, and we produce more than what we use. Last year, we exported 90 terawatt hour 90 which means we can localize a lot of data centers on top of what we need in terms of electricity for our own companies, our own our own households. So it's very important in this world, where I have a good friend in the other part of the ocean, saying, drill, baby, drill. Here, there is no need to drill. It's just plug. Baby plug. Electricity is available. You can plug. It's it's ready.
I don't care how cool you think you are, you will never be as cool as Mr. Macron saying plug. Baby plug. You know, the point here is that energy is a presidential discussion point. Residents of countries are now discussing how to get energy to satisfy the requirements for their own digital infrastructure. And France is not an exception. You see it around the new E in Saudi Arabia, of course, in the US and President Trump and the UK and so on. Energy is critical, critical, critical point that it requires presidential engagement to facilitate and try to develop strategies for this highlights the energy dilemma in Europe. This is this report just came out. By the way, it's March 17. This is in the middle of 2024 the three colors you see are the different prices for energy across Europe. So the blue ones, which is the Nordics and France and Spain and Portugal is the lowest cost of energy, more or less in Europe. And then you have the bulk of Europe is yellow, and then Eastern Europe is extremely expensive. What does that mean that there is a room for energy arbitrage in Europe. And although Europe acts as one, of course, the UK exited, it's not really one in terms of energy there is huge disparity in energy prices that leads to all sorts of interesting opportunities. This is the electricity prices in Europe per non household for industrial use and data centers. And I need you to actually consider this. This is why the Nordics is so successful in data center development because the energy prices is so low compared to other parts of Europe. And of course, the other part is the environment. It's very cold in the Nordics for in winter, so practically, you get between 90 and 100 days a year, a day a year, that you hardly use your cooling system. Second interesting point about this slide is Ireland. Ireland is a very interesting point. Remember, I told you about the flavi at the beginning of the presentation, Dublin is one of the key markets in Europe, but you you wouldn't conclude this from looking at this, because Ireland is expensive. The Irish managed to do is that although they have expensive energy, they have a very favorable tax regime, and they have a very business friendly environment for attracting these kinds of infrastructure. They have a very well educated workforce, so Ireland managed to actually attract investment in data center, although the energy prices are some of the highest in Europe. Does that tell us every player can be messy, but you can actually figure out a way to play football and make a name for yourself without being this perfect player missing same thing Ireland, the Nordics, everybody's trying to play with advantages they have, and there is room for all sorts of strategies here. You don't have to be the cheapest to compete. This is interesting. I don't think many people realize how many projects in power are going internationally. But this one is really interesting. UK doesn't have any meaningful solar generation. UK is the, usually, the energy in the sorry, the weather and the climate in the UA UK is, is really bad in terms of sun. So the UK is purchasing power from Morocco. So they're building this massive cable, 3800 kilometer cable from Morocco, where Morocco has a lot of solar energy and wind energy, and they're going to build this cable to connect it to the UK to power up data centers and other industrial users. Theory, this is at almost a 30 billion US dollar investment. The amount of energy is massive. The length of the cable is massive. And this is something that looks like science fiction at the beginning. You can't imagine that the company a country like UK, is purchased apart from Morocco. This is how extreme the power requirements are, and people need energy no matter what this is. Another example. This is Singapore and Australia. Singapore is very, very tiny country. There is not a lot of space there. There is not a lot of resources to generate energy. So they're buying energy through this cable from Australia, and they're paying $20 billion investment to have this cable, which is 4300 kilometers, to get power from Australia to power up data centers there. This is called the sun cable project. These are insane projects, very expensive, but this is what, what, what the crisis we're facing is forcing people to do this is interesting. Microsoft just signed last year a $1 billion investment for a data center in Kenya. Here's the game plan. This is why Microsoft is such a dangerous company, and they're the smartest people on that planet. I don't care if anybody says Microsoft are the most dangerous company compared to Google and Amazon and so on. So what Microsoft is doing is that they have this fantastic tool called the Global schedule. And what does that, does that to allow them to do is that they have data centers in different parts of the world, and the cost of energy in these places are different depending on therm, geothermal or solar or wind and so on. So what this global schedule allows them to do is that they send AI workloads to the cheapest energy source first. Let me give an example. If you are BMW and you're orienting for Microsoft, 8000 servers to process the next design for the next seven series. Microsoft takes the business from you and send you the server at a certain let's say $10 per hour rate is fixed. But then in terms of cost of delivery, they can shift the workload to where the power is cheapest, so Kenya, Middle East or east, whatever it is, so they're able to optimize, in real time, the utilization of the global infrastructure to minimize the cost and maximize the profit. This is what what the power crisis allows people to play with. I'm not going to talk about renewable energy today, because this the presentation is too heavy already. I'm going to land it under the sustainability or the planet aspect next, next episode. But I do need to give you some information about the Nuclear Power Solutions, because there is a lot of misconceptions about it. This is a very interesting quote, cannot win an AI without nuclear. So nuclear is not something that is exotic rare. Nuclear is going to be the power horse of the industry, whether you like it or not, just that some countries are far ahead of others on the curve. Nuclear is going to happen. I'm going to tell you how it's going to happen. These are interesting news from October and September last year about major players like Google and Microsoft signing nuclear power purchase agreements with generators. So you can and already clearly understand this is going to be very, very important for these companies who are committing to purchase these amounts of power. They need you to understand that there are three categories of nuclear reactors. There is the traditional one in Asia. By the way, we're building one in bah so this is traditional nuclear reactors that you have in mind, but then you have something called SMRs micro reactors. And almost always make sure that you understand, whenever you read an article, listen to a video, if talking about micro reactors or SMRs or traditional reactor. This is a typical example from Rolls Royce. They have an SMR that can output 470 megawatts. Is the size of only two football fields. This is a very interesting development. This is also an interesting development. Look at this. This tube, which is 23 meters by 4.5 meters, can generate 77 megawatts. You can only you need to refuel with different fuel, be 21 months. So companies like new scale and rolls, Royce and so on, are really innovative in terms of their designs of new, very space efficient nuclear reactors. I'm going to talk to you a little bit about micro reactors, because this is the hottest part of the industry. So micro reactors provide a number of benefits. They are provide clean energy. They are versatile. They can deploy it very rapidly. They have mobility aspects, and they're modular. Micro reactors provide energy between five and 20 mega watt of energy, which satisfies many data center requirements. Very interesting developments there. I just want to highlight this twisting house has the model of micro reactor that can just five megawatt and it has over eight years of review refueling cycle. That means that you only have to refuel this micro reactor once every eight years, and then it runs. Gives you energy based on that very few energy sources can compete with this option. It's very versatile, very modular, very robust. Of course, the way forward, there is a lot of certification, a lot of inspection, security inspection and safety inspection, that has to happen before these things are deployed. But here's the good news, we already know how to make these things. The US Navy has been running nuclear power submarines since 1954 we already know how to make very small micro reactors. We just need to take this is 70 years ago. By the way, for 70 years, we have companies and contractors and engineers and project managers who are very proficient in generation after generation of developing these small size reactors. All you have to do now is just get it out of the submarine, put it in a different enclosure, and figure out a way to deploy it in a basement of a building or in a campus to power up data centers. This is what a modern nuclear micro reactor looks like. Just fits on the container on the back of a truck. You can hold it to wherever it's needed, and it's self contained, it's safe and so on, and the amount of energy is huge now, Mr. Elon Musk about
modernizing space, and here's another reason why nuclear energy is important, because you need these things on Mars, because Mars you do not have an atmosphere to get you wind. You cannot run a wind turbine on Mars. And solar is also questionable on Mars. So another reason why people are investing in nuclear, micro reactors and these kinds of infrastructure is that besides meeting the requirements of data centers, you will need them when you go to colonize had a base on the moon, or maybe colonized Mars. This is a demo last two years ago by Rolls Royce. This thing, this small unit, outputs 50 megawatt of capacity. So you already can feel that this thing is moving forward very quickly. This is a two bits of news that I wanted to highlight to you. This is March 28 or 10 days ago. $100 million of investment is being made by the Department of Energy in the US to deploy SMRs in Canada, our films in Canada. This is last Friday. They issued an SMR construction license. So when people talk about nuclear, they talk about it as if it's something that is like science fiction or something that is just imaginary. No nuclear energy is having a renaissance now, and these packages and new form factors will allow everyone to use nuclear energy in a safe way. Of course, we have to pass through a safety inspection and certification process, and practically speaking, in terms of timeline, this is not going to be available before 2023 2035 because you need to go through all this massive amount of certification and security inspection for deployment in a situation where you have a nuclear micro reactor in the basement of an office building or a data center in downtown Cairo, for example. I'm finished with the presentation. I'll get just going to leave you with a couple of closing remarks to wrap the whole things together for you, and then we can open for questions. I do apologize the material is a little bit heavy, but power is such an integral part of the digital infrastructure development cycle that it needed to be adequately covered. When you get the presentation, you have all the material that you need to do on your search, but I just need to give you my parting thoughts before I open the floor for discussion. First one is that there is a war ongoing, whether you realize it or not, there is a war ongoing. It's not only USA versus China, it's every country in the world trying to make sure that it does not get taken advantage of, and they have need to have their own infrastructure to run the applications and the systems they need for economic value creation. Second point is that energy is not really about energy. Not only about power, it's about sovereignty. Could the examples of Singapore and Australia, or the UK and Morocco. If you cannot figure out enough energy to what you need, whether whether your own resources or you buy it from somewhere else, then your sovereignty is going to be at risk. Countries who do not have enough energy to run digital infrastructure within their own borders will have an issue maintaining the sovereignty, because they're going to be, fortunately, sort of slaved to someone else also providing them the infrastructure they need. So talking about energy is talking about the digital economy and also talking about the sovereignty and economic growth prospects of countries.
Now this is the summary. Do you have power and
you have power? What does that mean? Well, if you have energy, if you have electric and energy, that gives you leverage. Australia got a lot of concessions from Singapore in order to agree to give it energy. Morocco got a lot of cooperation from the UK on multiple tracks, industrial, agricultural, tourism, in order to actually cooperate in energy, if you have power, it means that you have more energy resources, like France, for example, bragging about having exported 90 terawatt hours. If you have energy resources, then that gives you some political maneuvering and cars to play with. The countries in the Middle East, including Egypt and some countries in Africa as well, do have energy resources to play the game internationally. But this is what I want you to understand. Energy is about sovereignty. It's about playing a political game. It's about protecting your economic growth. Process for your people cannot be complete without a comment on Egypt. The puzzle in Egypt is really interesting. There is a lot to gain, a lot to stake. The point is that the puzzle pieces are really aligned, yet. We need a lot of effort from all stakeholders so that the puzzle is sold like this. But I'm really enthusiastic about the future of Egypt and digital infrastructure. We just need to actually have a plan, because this doesn't seem to be a consistent about the direction we're moving into as far as the development of these things. This is the scary scene I like the most about. Remember the movie, these are not mountains. They're waves. This is what you should feel about the digital infrastructure. Width coming towards us. It's like a mountain, or like a huge, a huge, moving towards you very quickly. And it requires infrastructure. It requires energy. I need to plan for it. Otherwise, it's going to be swept by this wheel. For a small country, you're going to functionally be swept by this. And you can find yourself sort of buying infrastructure services from someone else because they developed it and you did it. Another skill statistic for you. This whole thing only materialized because of chat GPT. Remember? Chat GPT was launched number 30th, 2022, it's less than 1000 days ago. All this talk about digital infrastructure, energy and politics and planet and sustainability and data center and AI and power and processors and GPUs, all this fully started less than 1000 days ago. This is a very scary thought, because things are moving at an insanely quick rate. We need to have a plan. Every country on the planet needs to have a plan. It doesn't have to be a perfect plan, but you do need to have a plan, because that is what you require. We're all running out of time because this thing is moving too fast. So there is nothing more I can say, except that whether you have a company to run, whether you advise a government official, whether you are a part of an expert group somewhere, or whether you are a consultant, please make sure that people understand the urgency, because these things are moving at a rate that we have never seen before. And with that, I close my presentation and thank you for listening, and we open it up for Joly,
I see a couple of the I see someone from Nigeria.
Can you read it?
Yeah, sure. I'm reading the question now from Nigeria. I just want to ask this question. I see you display Niger data consumption is 29 kilowatt per kilowatt hour. That mean the RAI servers have their spec and business. Well, if I understand question, the question properly is, how can such a high consumption or high energy consumption rate for GPUs is sustained in Nigeria? Well, you're not alone, my friend. I mean in Egypt as well, as well as most parts of Africa. For that matter, Kenya as well, and even South Africa, nobody has that amount of energy lying around waiting for a data center to be built and supply it. So you're correct in identifying that this is a challenge. We have the same challenge in Egypt, and that's why renewable energy and nuclear energy is very interesting prospect to explore. So if you are serious about developing infrastructure, the first place to start, where is it that we actually generate the power I need to power those data centers, and regardless of whether you have a solar or wind or geothermal or nuclear, you have to actually put together a plan, because that's the first block, because once you have power, then you can actually move into the development of the data centers and talk about AI and so on. There are other challenges, by the way, in AI space, including getting buying processors. In America, mid processors is really difficult right now because there is a limitations on international customers buying Nvidia and AMD chips. So that's another challenge. But if you sort power Chinese at least you can buy Chinese chips or Korean chips, or India chips and so on. Energy in Nigeria and in Egypt and in Kenya and almost all the Morocco and all the other African countries remains a challenge to be resolved, and it requires the cooperation, not only of the cooperation, not only of the energy ministry or the electricity ministry, the Ministry of digital infrastructure, the Ministry of Economic Development, the Ministry of international cooperation. Everyone needs to cooperate to develop this I think I addressed the question, anybody else has the question? Another comment about the content.
Okay? So I like the Macron part. So I like when he said that France has identified 35 locations or sites, and they and they have a plan, as you were saying, and they have their own competencies to do this. And also, when you said the Morocco, UK or Singapore Australia project, the undersea place was transporting the energy, or a collaborative effort in Africa? Do we have something of the sort, something similar, that we have identified? We have an strategy for that, maybe as African nations, or whether for the electricity part or for hosting data centers on our land, because I think Africa, maybe North Africa, has a lot of competencies already in terms of space, energy. Waiting for you. Would like to hear your thoughts on that.
Professor is correct. North Africa does have a lot of potential, because we have a very well educated workforce, and we have a lot of land that we can actually build these massive infrastructure projects on problem, which is the first part of your question is that energy, we still all challenge in energy, because the traditional energy sources that we have in our countries were not designed for these kinds of massive loads. They're designed for industrial or residential uses. So essentially, what we have to do right now is to go back to the drawing board and figure out, how can we accelerate the development of new renewable energy projects to meet the demand for these projects. The good news is that energy, sorry, Egypt, Morocco, Tunisia, Algeria and most of Africa actually has a lot of renewable energy potential to play too. That's the good news. The bad news is that we need investment, we need technical expertise, and that requires a lot of coordination. So that's the first part. There are the ingredients of a fantastic meal, but we don't have a chef in the kitchen yet to cook the meal for us. The second part is, are we cooperating regionally for power generation projects? Not to my not. Morocco is doing its own thing. Kenya is doing its own thing with geothermal, because Kenya is lucky to have geothermal energy. Egypt is doing its own thing. There was talk about two years ago about the cable, power cable between Egypt and Greece to be developed. I'm not up to speed on that project. My understanding it didn't, didn't happen. But the point is that we're not really developing these Pan African power networks, yet might be able to do it. We can show that, similar to France situation, being able to export so much energy because we have nuclear, a country like Egypt can be in a position to explore so much energy using solar or wind, for example. And then we can go to countries like Tunisia or Sudan or Mali or Chad and tell them, hey, if you want to develop data centers, take the car from me. We'll develop not a submarine cable, we'll develop a terrestrial cable, underground or other kinds of cables to give you energy. So there is potential to play this role regionally. But again, it's not clear that people are paying attention to it yet, and not all countries are prioritizing these kinds of investments because they do not see the full potential about the digital infrastructure and its effect on their economies.
Thank you. So if I can follow up again, Macron, when he was saying in his reporting, said, our own companies and people. So he was speaking in the global AI summit, but he was still about France and French companies and serving the French people. Because change, although, because of the nature of the business you might have when also need to have it to host on your lands and make money of international organizations and companies. The second thing also, I wanted to get your insights on when the immediate CEO Johnson was saying, it's impossible to go without nuclear. Why is he so sure? All this is about innovation, and things are shifting and changing all the time
because, because the energy potential of nuclear is unlimited compared to any other energy source. Was, was a slide. In one of the slides, we said that one kilometer, sorry, one one kilogram of uranium can generate 24 gigawatt of power. You don't have the same level of efficiency. Solar, for example, burns so much land because you need to lay down these solar panels. And solar panels, even the best, top of the line ones have a 30% efficiency. That means that 30% only of the energy that comes from the sun and hits the surface of the panel gets transferred to electricity 70% and this is cutting edge stuff. This is the latest 2024 to do, 2025 panels from China. So compared to that, nuclear is a very efficient, very modular, very space conservative, and you can get a lot of it. And so in terms of maintenance as well, you refute these things every two, three years, in some cases eight years. So that's what he said. There is no other energy source that can compete with nuclear those aspects. And of course, it still maintains the clean energy mandate, because there is not a lot of there is not CO two emissions from it, although in the case of nuclear, there is spent fuel that requires careful disposition of because they have nuclear waste. But again, it shares that zero CO two emission with solar and with wind and with geothermal. So that's where Jensen Hong is coming from. He He looks around and there is no other energy source that can compete at the same level of efficiency with nuclear, and that's why the US and Canada and the UK, you saw rolls, Royce and so on, are betting so big on it, because they understand that long term, while solar and wind and geothermal can play a role. The ultimate solution has to be widespread use of nuclear. I'm looking at the chat again. I hope I answered your question. Mariam, if not, let me know. Mr. Abdul Fattah, saying, How can I access the previous session recording, including this one? Well, my understanding, everyone who registered for this session will get emailed a link to this recording, and then if you reply to that email, we can send you the recordings for the previous sessions. And also you can find them, I think, online on the YouTube channel for ISOC Miriam, correct me if I wrong before I move to ESPN question, please Yes,
yes, and please communicate with us on internet, master page on LinkedIn. We will put the link for that as well. Might be easier as well. Okay, so now I have a question, and they have a like live channel, so we'll find it okay. Somebody
posted already. Thank you for that. Jon posted part one and part two. Okay, so Abdul Fatah, you can find the links in the chat for part one and part two. The address yes means questionnaire means from Vodafone Egypt, and her question is that believe that in Egypt we have the land and the energy sources, but unfortunately, we are still behind the DDT center race, correct? Do we think? Do you think we are out of the race or we still have an opportunity to catch up? Also, what would be the best strategy to start? Okay, that's a two part question. Let's address the first one. Are we out of the race? No, we're not out of the race, but we're really, really behind. We are not out of the race until the race is over. So the race is not over yet. There is no clear number one and number two, although we can say that there are front runners and there are laggers like like Egypt, the reason for that is that we still have not as a society, as a community. This is not a technical discussion. This is about economy that touches every one of us as a society. We have not really given this. It's rightful place in our priorities. We still view data centers as something that the Minister of telecommunication or the ntra, the regulator, or a couple of investors should worry about. It's not something that we really should not concern ourselves about as a community, as a whole, or as a country, because it's not really that important. Once that view changes, similar to what happened in UE in Saudi Arabia and Kenya and Morocco, more people said, Wait a minute, this is too big of an opportunity to miss. Let's put our heads together. And you have a cabinet, many councils, and you have COVID Asian mechanisms, and you have portability, and you have KPIs. Once that happens, then all the resources you refer to in terms of energy, energy and land in Egypt would be put to good use. So far, this is not happening because it's not really a priority, and it's not something that is being tracked as a key KPI for the government in terms of digital economy development. And so it's not higher yet. It's not visible yet. Also, what would be a best strategy to start well again, the best strategy is something that's of an open ended question. Every country plays the game differently. Kenya plays with geothermal energy. Morocco plays with solar. And Morocco decided to export energy. We may decide to actually not export energy, but to actually figure out a way to build these massive data centers, but then we have to figure out what kind of land to give them at what price, what's the price of energy we're going to charge these data centers, what kind of investment incentives we need to offer them, and also what kind of mandates that we can relax? For example, security in Egypt is always a concern for investment because there is a lot to a lot of requirements if you are running a data center that you have to comply with, and if you're running in a race, you don't want these kinds of extreme requirements hampering your movements. So also, we need to be very proactive and try to actually offer something that makes sense for investors, and not actually burden them with so much requirements for security and so on a middle ground solution is needed by all the stakeholders. This is not something that concerns only the technology team. This concerns all the teams in the government, and they need to actually work together to support investment and support a still competitive advantage of Egypt, which I personally think Egypt is really entitled to, but we Have a lot of work to do.
Welcome shark. Nevis.
Back to you. I think we covered. Oh, there's a big question on the chat again, how effective are nuclear regulatory frameworks in African countries, and do they provide the stable, transparent environment to attract a retain investor in the refuge? The short answer, according to my limited knowledge, is no. I'm not aware of a mature regulatory framework in any African country for that kind of investment. Again, we're talking about these micro reactors and small modular reactors. We're not talking about the larger reactors, because in a country, for example, you have been building one on the north coast for the last four years, online in, I think, three years, if I'm not mistake. But the point is that this is the traditional nuclear reactor, the one that, for example, was in Fukushima in 2011 that had this incident at Fukushima. Daiichi, to your question, I don't think African countries have developed nuclear regulators, similar to telecom regulators. When that happens, they will be able to actually issue licenses, and they will actually be investment opportunities. They're going to be investment sorry, inspection bodies and so on. For now, I don't think it's a priority for most countries. And again, this is for good reason. If you are Kenya, for example, you have geothermal so you feel that you have enough energy to satisfy the demand that you need. So you don't really need to go all the way into nuclear. Other countries in Europe or America or Asia Pacific feel the need for nuclear running out of development potential for traditional renewable sources. So in Africa, it's a long way. I do not expect African countries to develop these kinds of projects anytime soon. But that doesn't mean that they cannot play the game. They can play the game using wind or solar or geothermal, and definitely, we have shown examples like Morocco, that you can have a credible strategy using what you have. You don't have to play the game like the Americans or the Europeans played. You can play to your own strength. But ultimately, big, big, big, big picture and long term, nuclear is going to be the workhorse of the global industry. Companies like Google or Microsoft or Amazon or open AI will definitely have to rely on nuclear for our countries in the Middle East and Africa region. We have other options that probably can be more accessible and more cheaper for us to
develop. Anyone have any other contributions or insights to share? I
Well, we thank you for joining us today, and we hope, we hope to see you on the upcoming session, which I think is going to be next month in May. Or would you like to have it earlier?
Well, again, this one should have happened three weeks ago, but we have to postpone it because of a sort of a very important in Egypt. We're not going to get into that right now. But the point is that definitely, let's, let's do the follow up one, which is going to be the planet, because I need to cover the renewable energy aspect, which I did not get a chance today to cover it. I already covered nuclear. So let's do something in May. Let's say the first or 10 days of May should be a target date. This allows everyone to sort of digest the information we can gather again in May, and hopefully you should be able to close the session with the people aspect sometime in June. And then based on all the five sessions, we can determine what do we do next in terms of maybe a tiny discussion with multiple speakers, Q and A sessions and so on. But my intention is to do the next one in the first 10 days of May, and I hope that's okay with everyone on the go.
I'm sure everybody will be interested to participate in the new one. As you said, I think we will conclude. It will be very nice if we after the we finish the people, we can do an open discussion, maybe we organize it with speakers, or we have like an open mic for an hour with everybody to contribute, and there will be a back and forth discussion. I think this would be also a good option when you investigate that and come back to everybody. Thank you everyone for participating and thank you for your questions. Very much for your insights and all the valuable information you've been sharing. I have, I took a lot of notes. I have multiple follow up questions, but I think we have to leave now 730 so
Okay, understood, the only action item is, yes, is the presentation or I'll make the presentation available via a link. I'll share it with Maryam and the Internet master team, and then I think that will be a mail to everyone, so you will get a link to download the presentation, and then you have all the material. That's the only action item
on the call. Yes, thank you very much. But it was amazing, and I think we have a lot of takeaways from today's Thank you very much everyone. Yes, yes, to everyone for being here.
Okay, thank you enjoy the rest of your evening chow chow Thank
