So we have excellent universities where we study and do research, the number of scientific publications is truly massive and of value, we absolutely cannot transform it into industry. Today A2A is the largest venture capital body in climate in Italy. Brescia is a truly unique case at a European level because the entire city is heated by waste heat. Let's say that Milan is a city that could see a third of the city heated by data centers, right? If you look at our country, Italy only on high speed. It's an infrastructure that we could define as a strategic infrastructure built in these 20 years.
Let's start. I'll immediately ask a bomb question. So, who is Renato Mazzoncini?
Who is Renato Mazzoncini? Renato Mazzoncini is an engineer, since we're talking with Engineering Italy, I'm an electrical engineer, so I'm a person who started working in design. I designed the new Italian high-speed lines when we had to switch from 3000 volts direct current of the current historical network, 25,000 in alternating current necessary to run trains that draw 12 MW of power, so it was the French system and I worked at Ansaldo Trasporti, at Storic Trasporti today Tacci Rail and then I did, let's say, I did my career in transport, becoming managing director of Ferrovie dello Stato, always doing, I'd say, only the engineer which is not a trivial thing. In the context of large companies, especially state-owned ones, where issues that have nothing to do with management or technical capacity influence the choices of being appointed to those positions. Since 2020 I manage A2A. I have a lot of fun, the company has grown a lot on the two mainstays of this planet to safeguard, so the energy transition and the circular economy and then for the rest I teach at the Polytechnic of Milan since 2017, Mobility Engineering, a course for the last year master's students, it's a demanding 10-credit course. I do it because I think the relationship with the kids, with young people is very important and I believe that in a perspective of giving back, of generational passing, this works well. Having said that, I have three children, a wife, we have a house here in Brescia, I also have a house in Milan because we split between the two cities necessarily with work. Last weekend, Sunday I was doing the Pisgana on Inadamello with seal skins and ski mountaineering and tonight, today is Wednesday, I have the fortune that I'm here in Brescia and so I go, as I often do when I'm here to play basketball with a group of friends from university times. On Wednesday we don't get bored, we play, we meet. The Wednesday five-a-side football has become basketball. Yes, for us it has always been basketball. Exactly. And so I try to keep a balance between professional life and private life. I think the ancients said mens sana in corpore sano, they were perfectly right. So I believe we need to keep a bit of this balance.
Coming a bit to A2A, when a person thinks of A2A, more simply, in my opinion, they think of energy production, energy sales, as you said before, the two mainstays and waste management and in reality if you look at the update of the strategic plan you see that there are also 1.6 billion euros destined for data centers. So what is a bit of the strategy behind this investment?
A2A is a multiutility and multiutilities are born in Italy and Germany and that's it. Probably we are two countries with a similar municipal socio-economic fabric, that is, similar types of cities in terms of size and also probably in terms of the propensity of cities to also do industrial activities. In other countries, think of France, England, Spain or others, it's the large state-owned companies that are vertical on clusters. So take France, EDF energy rather than Suez or Veolia do water, waste, that is, they are vertical clusters. There are no companies that transversally do from water to public lighting to waste, to energy, to gas networks and everything. This model is a truly typical Italian and German model. Only that Italians don't realize how much in reality our country is more liberal than one might imagine. And so in this perspective, let's say, of the market, at the end of the 90s utilities in Italy are listed on the stock exchange, while in Germany no, for example. And so for example also the SM in Brescia rather than the M in Milan are listed, let's say, at the end of the 90s. When you go to the stock exchange, the stock exchange requires an equity story, that is, it says "Ok, so at this point how does it grow?" Because the investor on the stock exchange has an interest in having a company that grows, wants to see a path, let's say, a path and so the path that came most naturally in those years, I'd say also easiest, anyway rational, was that of mergers. It says "Okay, I grow by merger". And so in those years, meaning in the 2000s, between 2000 and 2010, the great Italian listed multiutilities are born, so A2A from the merger of SMM and then it merged inside the other multiutilities of the Lombard territory, LGH5 AIB, and ERA is born, Iren is born, Acea, in short, so those that today are considered the great listed multiutilities are born in those years. Exactly. A2A becomes the biggest, let's say, the biggest of all, also because it comes from two truly very solid realities, that is, mom and dad were and Milan to compensate Brescia contributes not only M but also Amsa because otherwise the SM of Brescia, despite being a city seven times smaller than Milan, was bigger than the M of Milan and so they put together basically the SM of Brescia, the M of Milan and the AMSA of Milan and the A2A group is born. After that, let's say, what is the real problem? We are on the stock exchange, so we are the largest listed multiutility on the planet. Only that this is not good news in itself because they made a panda and the investors on the stock exchange are subjects who study the markets and once they have studied the market, that is, the type of company, then they invest in all companies of that type. But if you are alone it's inefficient, right? Rather it's inefficient, that is, they have to study you and then they don't have many peers to invest in, right? While I know I invest in an automotive and then it's validated, that's it.
So, aware of this issue, we tried when I arrived, we did the first ten-year industrial plan to simplify the storytelling to make it more understandable and we divided our world into two mainstays which are those you find in the strategic plan, energy transition circular economy, at least these two words everyone understands. Energy transition substantially is inside energy generation, networks, market plus a bit of end uses, public lighting, electric cars and whatnot. In the circular economy part you have inside the world of the waste cycle, the world of district heating, the world of water and now we have inserted the world of data centers. We are trying to address, among other things, the two big problems of the planet because at COP they systematically talk about the CO2 issue which is certainly a huge issue you face today with decarbonization. So the energy transition pillar is the key pillar on this. However, if we imagine a world in 2050 miraculously decarbonized, something that is not exactly happening, anyway let's imagine it happens if in the meantime we don't solve the problem of the earth overshoot day, that is, the day when we run out of natural resources of the planet which this year, if I'm not mistaken, was July 25th, but it was July 25th.
Every year, by the way, it almost moves up.
Yes, yes. And it's the average of the chicken, meaning that in Italy it was May 15th, in the United States in March, in Qatar in February, that is, and then there's Africa which goes instead the year after and compensates.
However, if we continue to take out of the pantry more than we put in, we can also decarbonize everything, but then we crash anyway. So we obviously deal with a part of circular economy, but think about water, the water cycle how important it is. The waste cycle is the downstream part, then someone upstream needs to build products with recyclable products and imagining to recycle them after, but anyway we do our part. District heating is an essential tool because it's a tool where I can recover waste heat. For example, in Brescia we connected the two steelworks and we connected the thermal plant and with this we heat the entire city. Brescia is a truly unique case at a European level because the entire city is heated by waste heat and data centers enter the circular economy pillar because data centers are certainly a frontier that in Europe we must face but they are also an opportunity from an energy point of view because data centers have substantially the energy that enters a data center for 65% goes to the server room, but the rest goes fundamentally to the cooling system, so to cool it down, because they heat up a lot, just put your knees on them for a quarter of an hour and you realize what the situation is. Sure. So if the data center is designed like the one we have for example launched as a pilot here in Brescia with Quarneote and to recover heat, at this point you can put it in the district heating network. The next plant we are doing in Milan in a few months. The data center issue is a truly essential issue. In Italy we have 168 data centers for a total installed power of 570 MW, so do the division super peanut 3 mega something. Today we talk about data centers of 1000 MW, 1 GB, so 3 MW. The biggest we have in Italy is the one in Rozano, of the Team which is 30 MW to give an idea. What did we imagine would happen? What we thought would happen until a month ago, that is, until before the war in Iran, was that in Europe AI learning, that is, artificial intelligence training, was not done in Europe because it requires great computational power, very large data centers and so the projects were US and Dubai, Emirates, Saudi Arabia, the buying, acquiring models already somehow trained. So what we expected from us was a development of data centers, from us I'm talking about Europe, of data centers that would substantially allow having cloud, having the terminals of the final applications that we use from Microsoft to all the rest. So a part also from us, but more the whole theme of data recovery, like public administration data, sensitive data that we prefer to keep here rather than in the United States or in China. Now the situation is changing a bit because two data centers were bombed in this war and so subjects like Amazon are reconsidering projects in Europe for reasons obviously of stability. We had estimated, so, in the report we presented last year at Cernobio we had hypothesized a base case around 2.5 and a best case around 4 GB and 3, so it means anyway compared to the half gig of today a nice growth. Very concentrated in Lombardy because there's a theme of being close to demand, there's a theme of latency, so there's still interest not to move too far from the core business by 2035. I don't exclude that these estimates might grow a bit. And at this point what is needed at a global level for data centers? Sites that are connected in high voltage, because above 10 MW the connection is in high voltage, not in medium and the high voltage connections at this moment are made by Terna and are all congested. Think that only on renewables there are over 300 GB of requests, they manage to do 10 a year and they are all high voltage connections, so there's also an infrastructure problem, that's right, there's an infrastructure problem, so in reality if you have existing sites you can have a totally different approach and we especially in Lombardy have many generation plants. Think of our plants in Piacenza, Mantova, Cassano d'Adda, on the Mincio, rather than the thermal plants of Silla, Milan, of Brescia or other cogeneration plants that fundamentally already have a high voltage connection, some Monster also from 1 gig and others smaller that can be somehow used to provide energy to data centers, if the data center can be at that point powered directly from one of our generation sources or anyway there's a site, let's say, connected to high voltage that can be partly dedicated to this. So we have a, let's say, a positioning in terms of connection capacity, generation capacity that can strongly accelerate development and so the reasoning we made and I close on the industrial plan, okay? But there's a party because it seems that the data center theme is truly from an economic point of view a great opportunity of the coming years. Going to be the DJ doesn't amuse me much, I prefer to participate in the party. So, we said, let's bring out some money. We allocated 1.6 billion in the plan, in order to also participate in the investment and so we want to invest directly. With 1.6 billion, according to our estimates, we could reach an 8-10% market share of all data centers built in Italy. Obviously we don't put the servers, the racks, we do the infrastructure which however is a complex technological infrastructure because I repeat, you have an energy management theme, cooling system, water management, all the activities that we do and on which the fact that we are very crossing between the two mainstays is the umpteenth demonstration that all in all this panda makes sense, in the sense that the more we go forward and the more to address the important issues of the planet, the two mainstays together are needed, so for example a data center needs a network and we work on that, electrical generation, but on the other hand if you connect the data center to a district heating network, you recover heat, you do district heating, you reduce water management because you have anyway a water management situation that benefits a lot from a district heating that lowers the PUI of the plant a lot and so I mean it can work very well.
So I see there's also a strong technical component that excites you. So as an engineer, what aspect working in A2A stimulates you the most, excites you the most, has excited you the most or will excite you the most?
I'd say building infrastructures. Draghi, you know the famous Draghi report, right? What does Draghi say in that report? Basically he says: "Guys, in 2000 Europe had the GDP of the United States and we were double that of China". 20 years pass, the United States gave us the beer, they are 40-50% ahead of our GDP, China has caught up with us. Fundamentally Europe has remained still. Why Draghi's analysis? Because we didn't take advantage in these 20 years to build strategic infrastructures, something that Europe had done from the post-war period, let's say, from the 50s to 2000. And indeed it's true, if you look at our country, Italy, only high speed is an infrastructure that we could define as a strategic infrastructure built in these 20 years. And indeed A2A is also a bit of a mirror of this, in the sense that A2A was born in 2008, so this year it becomes of legal age. However, if we go back simply from mom and dad and to place us in 2000, the last major investment made had been the thermal plant of Brescia. Then substantially for 20 years from 2000 to 2020 the existing assets were well managed. The investment costs, the so-called capex in the company, were about half a billion a year, which was what was needed to do good maintenance of all the plants that goes from the Valtellina dam, to the district heating of Brescia, all the plants that are maintenance work more than anything, done very well, so it was a company that maintained excellent rigor from a technological point of view, but it didn't build anything more, so it's a bit of a confirmation of what Draghi says. When I arrived I found a company, I must say, very orderly, very well managed and also with many engineering and technical capabilities inside. I called my CFO and I said: "Listen, but what is the maximum speed we can go without going out of the curve at the first curve?" He does all his calculations and says "Well, we can invest 1.8 billion a year?" I say, "Sorry, then why are we investing half a billion?" He says, "I don't know. And this is our riscapetite". The fact is that in January of '21 we come out with the ten-year plan of 18 billion euros, imagining to realize 70% of investments in infrastructures per year, because the gaps that need to be filled to solve the problem of energy transition circular economy are massive. So the last of the problems we have is what to do, that is, it's really, how to say, it's easy, there's the embarrassment of choice. There's the embarrassment of choice, and so we get to work and I must say we have invested on average 2 billion a year. The company, moreover, in the meantime, has doubled in size. All the economics, the profits, have gone from 300 million to 700 million, which means that investing in the Green Deal, because then we kept that North Star, once again it confirms a right choice and so we started investing in renewables, so solar, wind, obviously keeping good account of our excellent hydroelectric, investing in electrical networks. Last year we also acquired an important piece of the network from Enel, investing in water. We did a lot of work on reducing losses, on new treatment plants, investing in new plants for waste treatment, also new, we are the only company that has built new thermal plants in these years. In short, we invested a lot of money in infrastructures, so, if you ask me what is the thing that amuses me the most, it's seeing physically the fact that at this exact moment we are talking we have 200 active construction sites and so I like to see a company that is building, that makes the infrastructures needed for the future. After that, since I am, as I said, a very technical engineer, electrical engineers are among the most technical of all and obviously then many details interest me, so it's not simple, obviously to keep the two levels because then I have to deal with all the troops. However, the typical curiosity of engineers in these cases helps a lot.
It helps a lot, certainly.
Speaking of heat recovery data centers, I wanted to reconnect a bit to that, that is, technically how does the data center work, it heats up and we're fine with that. So how does the heat transfer, heat reuse happen?
The chip heats up, it's really the microchip that heats up. So, for example, what we built here in Brescia happens in an extremely simple way, that is, by conduction, that is, physically the racks are designed so that the board where the chip is protrudes, imagine from this rack and the pipe of the district heating water circuit passes right in contact. There's a small heat exchanger, but really metal to metal, very simple. And the water comes down at 45°, 45-50° and returns up at 65°, so it gains about 15° of temperature and then it's connected to the district heating and it's connected to the district heating. So if you go inside that data center you can see the red and blue pipes. The blue ones are the cold water pipes logically that pass and the red one heats up. Typically then, since our district heating is an old generation district heating, that is, it's a district heating that heats houses that tendentially have radiators inside, they don't have underfloor or ceiling heating that work at low temperatures. So the radiator for us requires an operating temperature around 60-65°, so it means that the temperature of the heat exchanger pipe that runs in the city must travel around 100° and so the data center is not sufficient, but I use heat pumps to raise the temperature and put it in the district heating. Sure, but they are all megawatts that we recover. The heat jump is reduced, let's say, the thermal delta and so the amount of energy I need to bring it to temperature is reduced. Depending on the...
Technologies, now in Milan we are experimenting with another one, but substantially, fundamentally the conduction mechanism, that is, the heat is... there is a small heat exchanger that exchanges directly from the chip. And fundamentally we could define it as a water cooling system. The difference is interesting because if you enter in July into the data center we have here in Brescia, let's say if you stay inside for more than a quarter of an hour, you faint, they come to get you with a stretcher.
But this is interesting because since nobody tends to work inside and the data center is off, lights off and air conditioning off, because the cooling I care about on the chip is heated in this way. It's so true that it's one of the rarest data centers on the planet that has a PUE, which is the main sustainability KPI of data centers, and it's the ratio between the electrical energy I emit and the energy that goes into the server room. So it tells me how much energy doesn't go into the server room. Well, a place where the lights are off, the air conditioning is off, all the energy goes into the server room, PUE equals 1, which is the dream of all data center operators.
It's a bit like the concept of dark factories that are developing around the planet. So if someone needs to go in for maintenance, they enter, turn on the light, turn on the air conditioner, but they consume that little bit of energy for the time they are inside. While when the data center cooling setting is typically air-based, you have to keep it constantly air-conditioned, which is the classic situation you find in any type of data center. If you go now to any, even corporate ones, you enter the little room with the air conditioner, the air conditioner at full blast because it has to keep the temperature.
So your logic, the intuition is to cool everything when what we care about is the chip.
Exactly. That's the point, right? Someone once wanted to invent a suit for workers that was cooled or heated to avoid having to air-condition the entire warehouse. Then at the moment they haven't managed to do it yet, but there are a series of situations where it's truly inefficient.
Because then there's a paradox, you have usable heat, instead you use other electricity to cool.
Exactly. Exactly. It's like that. It's a paradox, but it's... so no, perhaps, let's say, the path that almost everyone still follows.
Yes, yes, it is. We estimated that in Italy we could manage to heat almost 800,000 apartments with recovered heat from data centers. Besides, district heating now, perhaps more in a city like Milan than Brescia, is also the only tool to manage to decarbonize, because you know where district heating was born around the world?
In Manhattan, in New York, right? The steam from the manholes in Robert De Niro movies is district heating heat.
Why did they do district heating in Manhattan in the 1800s? Because you already had these skyscrapers. So the idea of having a coal boiler inside each of these large buildings, because at the time it would have been coal that heated the building, would have generated a monster pollution inside Manhattan and also an occupation of very precious space because even then real estate in Manhattan was worth some money.
Spent on central plants. So they invent district heating, they build the plant outside, so they move the pollution and space occupation outside because the heat exchanger is a small thing. Small thing.
Move to Milan. In Milan today we have district heating only for a piece of the city, 15-20% because it's connected to the Silla waste-to-energy plant, the one in Milan, managed by us. The rest of the city is essentially heated with gas boilers. If I have to replace a gas boiler with a heat pump, I have two problems. The first is that it's typically in a basement room and instead I need air to do it. The second is that the heat pump takes up much more space. And the third is that even in Milan now, real estate per square meter is worth a fortune. So you understand, mutatis mutandis, that developing district heating, moving it outside, I manage to more easily have heat exchangers with, for example, data centers and so on, and so I have the benefit that I manage to decarbonize, effectively close the gas boilers, recover very precious real estate square meters inside the city of Milan, and I manage to recover data center heat precisely because 40% of data centers will at least develop around Milan. Let's say that Milan is a city that could see a third of the city heated by data centers, which would be an amazing thing.
So this heat recovery, let's say, is certainly one of the levers to make data centers sustainable, which are anyway very energy-intensive.
It is. Is there any other lever that comes to mind to make these infrastructures sustainable, considering also the development you were talking about and that is expected in the coming years? For example, I must say that national and regional laws are also giving good indications. For example, using brownfield areas, not greenfield areas. So now the regional law says that if you want to use a greenfield area, you pay double or triple fees. So please use already industrialized areas and not take away land. To use water management efficiently because the other big issue is water use and therefore water recycling mechanisms, and these are other topics. Then for the rest, you should talk to the Nvidia people, because then we have to see if in the coming years...