Power Shifts: Data Centres, Diplomatic Tensions and the Energy in Between

Transcript

Power Shifts: Data Centres, Diplomatic Tensions and the Energy in Between

I think that's where we stand today. We're raising Superman. And sadly, we're being very bad parents.

First of all, I'm not so obsessed with the idea of leadership. Why should we be leading?Sometimes there is a second mover advantage in the economy. We don't need to be always the leader.

it's not just about the availability of power for the data centres, it's how do you keep power prices also affordable for people?

if you ask me about one thing that the world needs to be aware of and work on is what I call the mad map spectrum, the mutually assured destruction versus mutually assured prosperity

clearly AI can impact the industry in a positive sense, for sure. in the utility space and in the energy space, it can have improvements. But let's face it, it is a vampire at the moment. I mean, it is clearly sucking up a lot of power. So to my earlier point, if it's not managed properly, it will create a lot of pressure on society, on energy prices, on the consumers.

So the best thing is to diversify, diversify the sources of energy, diversify the countries of origin, diversify the supply chains, and diversify also the transit routes. And of course invest also in new sources for the future.

The opportunity is absolute abundance. That's the opportunity. And if we all agree on the opportunity and chase that, that would be amazing. Genuinely we would redefine longevity, we would redefine productivity, we would redefine climate change. Everything is possible with more intelligence.

I'm daring the whole world that when I can borrow 400 IQ points from AI, I'll probably be able to have a garden outside where you walk to one tree and pick an apple and walk to another tree and pick an iPhone.

We are living in an era defined by a new kind of power shift. It's not only measured in borders and markets, but in gigawatts and gigaflops. Every line of code generated in every industrial process optimised by artificial intelligence is tethered to a physical reality, the data centre. These silent humming monoliths have become the new cathedrals of the 21st century and their appetite is unprecedented. In a statement in November last year, OpenAI said they needed 30 gigawatts of compute capacity. This is a third of the peak demand on the US grid. They said that they would be adding about one gigawatt a week. A gigawatt is about the power created by a small to medium sized nuclear power station. But by 2033, Altman predicts that they will require 250 gigawatts. This is as much as India's yearly usage, the most populated planet in the world.

So let's get our heads around that for a second. How on earth are we going to be able to create and sustain the power demand of AI? But this isn't just a technical challenge. It's a diplomatic powder keg. With the strait of Hormuz slowly recovering from a blockade and the global energy market reeling from price hikes, we've entered a world of energy realism. Furthermore, is the machine a monster that will break our grids and stall the transition? Or is it the only navigator capable of steering us through the complexities of a changing climate and the demands of a green transition? Nations are being forced to choose do they prioritise rapid AI dominance, fossil fuel security, or their net zero promises? Welcome to the next five podcast. I'm Tom Parker and today we are launching a new six-part miniseries called Power Shifts. We'll be discussing what is powering our modern and fast evolving world today, as well as the blueprints, the tech, and the strategies to transition to a net zero future.

For this first episode, we're going into the diplomatic boardrooms inside the data centres and deep into the energy powering everything in between. We are joined today by three Titans to tackle this volatile topic. First up, we have Jose Manuel Doral Barasho, Prime Minister of Portugal from 2002 to 2004 and the man who led Europe through a decade of crises as European Commission President from 2004 to 2014.

My pleasure. Thank you.

Also with us is Mo Gordat, a software engineer, entrepreneur, author, podcaster, public speaker, and former chief business officer at Google X. He is the man that saw the scary, smart future of AI before it was on every smartphone.

Honoured to be here. Thank you.

And finally, Miguel Stilwell Dandrada, CEO of EDP, a global energy company.

Thank you. It's great to be here. Well,

It's great to have you all. Jose, let's dive straight into the fire. We are now nearly five months out of the initial February strikes that ignited the war in Iran. While the temporary spring ceasefire set to expire on August 21st technically holding by a thread, the strait of Homus remains an economic minefield. Temporary memorandums have allowed traffic to resume bringing oil prices down to pre-war levels at around $72 a barrel. But we are coming off a massive shock where Brent Crude briefly peaked at $126 and Qatar Energy is still declaring force majeure with some utility companies that could see disruption to gas markets continue for months to come. Jose, you've navigated European crises for over a decade, but we are looking at a fundamentally altered landscape where energy security and maritime openness are being systematically weaponized. Europe is still reeling from the immediate macro shocks to wholesale gas, fertiliser supplies and transport costs.

From your now backstage perspective in global diplomacy, how deeply have the scars of this 2026 crisis fractured our global economy? And are we looking at a permanent structural shift in how the West has to secure its energy roots?

I believe, yes, I believe we are not going to come to a status quo Ande. I think this crisis is very serious. Of course, we would like to have some kind of ceasefire, some kind of truth. But the fact is, and by the way, I'm on record saying that for some time, the conflict will continue. The issue in the Middle East and Gulf will continue, whatever the leaders say, because the internal causes of the conflict are there. I was in office already 40 years ago. I remember that time some optimists, including some friends of mine saying now there is a wind of opportunity to solve the Middle East crisis. The reality is that we are worse now than you were 10, 20, 30, or 40 years ago

In terms of casualties. The terrorist attacks of Hamas against Israel, many people died, but also the very robust reaction of Israel and now the full scale attacks of the US and Israel against Iran. So I don't see in the foreseeable future a permanent, settled, stabilised solution. So we cannot predict the future, but we can prepare for different futures. And what we should is prepare for the worst possibilities, hoping of course that some kind of arrangement can be found. But for me and also the best opinion I have from people who are now in action that I've been touched with, we should be prepared for a difficult scenario in terms of supply of fossil fuels from that region. That's why some alternatives are being explored, including new pipelines. And it's the more realistic scenario, I believe.

Well, yeah, Miguel, the war in Iran has essentially vaporised cheap gas prices that many thought would carry us to 2030. In this high price, high volatility environment, how has the energy landscape changed for a developer like EDP?

Well, I think it's certainly reinforced one of the key things, which is the volatility that you have around fossil fuels. So more than the specific price as of today or tomorrow or what it was yesterday, it's just the sheer fact that you have all of these geopolitics, whether it's now, whether it was back, for example, with the Ukraine war, where you suddenly have these spikes in the gas prices or the oil prices, which severely impacts industry. It impacts domestic consumers. For developers like ourselves, what we've believed for a long time that it's worth developing your own endogenous resources, which is renewables. In many cases, we bet very much on hydro, on wind, on solar, which are produced locally and which are not volatile. You can essentially lock in that investment. You can lock in long-term energy prices which are immune to this volatility. And I think that's been one of the big, for example, positive points of countries like Portugal and in Spain, which have very high penetration of renewables.

They've been able to navigate through these energy crises without having impact on the domestic consumers or even on the industries because they have this high penetration with prices that don't change because they're not so dependent on fossil fuels. So more than the specific prices, which is obviously volatile, it's the fact that there is so much volatility in the fossil fuels and what that means for the economy.

Absolutely. Mo, let's look at AI in this equation for a little bit. You've been an advocate for AI safety. In your book, Scary Smart, you warned about the intelligence of AI, but did we underestimate the appetite? Microsoft's electricity consumption has nearly tripled since 2020. So have competitors like Google, I should say there are other companies. We are hearing that hyperscale data centres can use as much power as 100,000 homes. By some counts, AI is now the world's fifth largest energy consumer between that of Japan and Russia. How can we deal with that?

So first of all, the idea of Scary Smart when I wrote it in 2020 was to warn against a rogue AI that might harm humanity's wellbeing in general. I've moved on from that. Of course, that remains to be the case. Some experts will say 10, 15% possibility that one day those things may cause us harm. But I'm much more in a mindset now with my next book Alive that the biggest urgent immediate risk is how humans will react to the rise of AI. And this is centred around an illusion of a Cold War, an arms race between the US on one side and China on the other side. And the need, the conviction that if one wins, the other loses and everyone else loses as a result. And I think that this is a very limited view of what is possible. On one side, it makes everyone else, including Europe, which has quite a bit of potential with AI, believe that it's a two-horse race.

So accordingly, they don't really even let their horses into the race at all, which is quite shocking for me. Sort of almost tilting Europe to behave more like Africa than Europe if you think about it, just leaving the race altogether. But at the same time, those two very different models that are the two horses, if you ask me, are actually advocating two very different things. If you go to China, the grid is mostly very, very developed. It's mostly renewable. It's redundant in many ways over the capacity that is required. And the big struggle for America is energy. And some of the reasons why, if I may say this without being too political, why someone like OpenAI would ask for so much energy is from one side, it's the only way American businesses know to win the war. It's bigger, bigger, bigger, bigger, bigger. Borrow, borrow, borrow, borrow, borrow, build infrastructure that is ahead of the demand.

But at the same time, I think the challenge is because the infrastructure wasn't there. So basically the idea is China was ahead of the curve while sadly America was spending more on military and defence and so on. China was building infrastructure and grids and so on. Having said that, I will have to say that thinking that AI is going to completely consume energy assumes a very interesting conviction that we are still using mainframes today. So when tech started, when we started with computing in general, we built massive infrastructures that were as big as this room with air conditioning and false raised floor and false ceiling and wiring and so on. Now most of that compute is available in your phone. And the idea of specialised AIs becoming more and more away from frontier and towards the edge, smaller devices capable of doing quite a few things.

It's not unlikely that you'll have an AI on your phone next year that is as good as the last version of cloud. And how many of us need something that's much more powerful than that is questionable. So is the need for massive data centres a tech challenge or is it a political position around trying to control something that's not really in the benefit of humanity? Maybe an approach to distributed energy and everyone having their own specialised AI might be an interesting model.

Yeah, Miguel, Mo just said is data centres about the tech challenge or this political discussion. You recently mentioned that data centre requests in Portugal alone could double the country's entire consumption. So how does a utility company, I suppose, navigate this tech discussion or political one and even plan for that?

Well, the first thing I'd say is that we're seeing a huge amount of demand, not just in Portugal, but for example, also in the US. And to Mo's point, I mean the build out is absolutely unbelievable. It's something completely unprecedented over the last couple of decades and you're seeing that today. So it's real demand that's on the ground and you're signing real contracts for the next 10, 15 years for the data centres that are being built. To deal with this amount, you're having huge investments on the network side for sure, because you need to connect all these data centres, but you also need to connect the new generation that's being built. And you're having to build a lot more generation as well. Now in the US, that's happening at an extremely high pace, which is why it's one of the reasons we're having some of the highest growth we've seen over the last couple of decades in the US in terms of build out of solar and batteries because it's being sucked up by the hyperscalers.

They're sort of rushing to buy that power to supply the data centres. In Europe, you mentioned Portugal. It's true. We have a data centre being built to south of Lisbon, which will represent 20% of Portugal's consumption, one data centre. So I think the big challenge there is it's on the network, it's on the generation side about how you supply that. And it's also how do you find the balance so it doesn't

Put pressure on the domestic consumers so it doesn't become a political problem for the people that live in the country. So you don't socialise that cost among everyone. And that's a regulatory and a political issue. And I think that's one of the challenges that many countries still have to face is how do you deal with all of this demand that's coming from data centres without then having to backlash from the consumers, which are also voters at the end of the day?

Yeah, actually we're going to get to that point in a little bit, but again, talking about the build out required for this, not just building of the data centres, but then the power demand that is needed it. We said at the top of the show that one gigawatt equates to the small to medium-sized nuclear power station. Now it takes 12 to 15 years to build a new nuclear power station. If you're in China, it takes an average of seven. But over here in the West is 12 to 15 years. The demand is immediate and increasing rapidly in the next five years. So how on earth are we going to be able to do it? Because we can't build a new small nuclear power station for every gigawatt.

Think that is a real, real challenge. I don't think that people quite realise how much of a physical challenge that is. To your point about the nuclear power stations, I would just ad, not only does it take that long, but normally there's huge cost overruns and it's delayed beyond what are the original plans.

So if it's already difficult to build a solar farm or a wind farm or a hydroplant, building nuclear plant is extremely difficult to licence and to actually then build out. So it's not going to solve your short, medium term issues. So really it's about either gas or wind solar batteries. But even those require a huge amount of investment, a huge amount of new projects that need to be licenced, that need to be permitted, where you need to find the physical space. You need to make sure the communities don't have a backlash and they don't push back against it. And so we are seeing real challenges to actually get that generation built. And so what the consequence of that, if you don't get that done, if governments don't react and have the sense of urgency is you will see power prices going up and you'll start having a backlash against data centres and you're already beginning to see that partly in the US and you'll start seeing that again here in Europe.

So this is really an urgent issue and I'm not sure people are sufficiently focused on it yet.

Yeah, absolutely. Jose, let's look at the political side of it as well. How does the EU maintain its global leadership when it's currently scrambling to keep the lights on? The cost of living down and let alone the servers running.

First of all, I'm not so obsessed with the idea of leadership. Why should we be leading? I mean what we should is to, I already mentioned this several times. Sometimes there is a second mover advantage in the economy. We don't need to be always the leader. Some of the more successful countries in the world are in fact second movers. A good example is Switzerland because of its size. So probably in some areas we are not going to be able to compete to the United States in some of LLM and the AGI and the, let's say frontier models. In others like the open source, the Chinese are much better. And by the way, I think they are leading globally. But we can in some areas do it well. We have a huge, let's say, engineering capabilities, great engineering in Europe, great science. We have also industry that is now going to be renewed.

There is a new industrial policy in Europe. In terms of energy, of course we have to adapt. We are starting point in a more vulnerable position because we are not except for renewables that as Miguel said, of course, these are indigenous. So it's also I would say basically today an issue of resilience and security, even more than the climate transition. That's why it's coming so up in the priority. But in fact, fossil fuels, we have not the same amounts that other parts of the world have. So we are going to adapt. I believe what is reasonable to do is to diversify. Diversify is always the best way to deal with a problem of unpredictability or volatility. By the way, the Chinese have done it. Chinese, they are now leading the climate transition in terms of green industry by far, not only in solar panels and now increasingly in wind turbines, but now in electric vehicles in batteries.

But they kept fossil fuels. They invested in renewables and they are also promoting nuclear, not speaking about coal.

I mean, that's really amazing. And now in fact, contrary to what I see very often in media that say the gulf is a problem for China. In fact, they are benefiting because they had huge reserves and also in fact they're now more appetite for renewables. And who can give that technology for renewables? China. So in fact, they are winning on all because they were able to diversify. So the best thing is to diversify, diversify the sources of energy, diversify the countries of origin, diversify the supply chains, and diversify also the transit routes. And of course invest also in new sources for the future. So what is important now from a political policy point of view, but I would say also from a corporate point of view, for the different organisations to be nimble enough to be adapting to new scenarios.

And I know it's not very sexy, speak about prudence, but prudence is necessary because we don't know what's going to happen. So the best ways to diversify at the same time, renewables is certainly important for the points I already mentioned about resilience.

Well, Mo, Jose said there, you've got to be nimble to adapt to different scenarios. The same must be true of trying to adapt to this fast-paced AI build out. The tech is changing all the time. So we've got this tech, but what about the risks? There are biases, hallucinations in safety critical industrial code and the sheer footprint of data storage. How do we protect ourselves against some of these risks?

An I just first double click on the second mover advantage? That's incredible when you really think about it. I have to say any mover at all is good because Europe is not moving. But second mover is quite interesting. If you look at the dot-com bubble and how there was so much hype and a redundant infrastructure built for the internet that was followed by a cleanup and a shakeup and a lot of equipment that was left over and then the real players came in and built the Googles and the likes. So there is an advantage there. I think the idea is to keep an eye on how to build that based on the current competitiveness of the other sides. And where do you position Europe within this is a very interesting one. The idea of where we stand with security, with privacy, with alignment, which is my big call for the last eight years trying to tell the world that we need to build AIs that have our best interest in mind, is that the decision makers in our world today around technology advancement are not being held accountable.

And they're not being held accountable mainly because of the arms race. So it's not once or twice that some head of product or head of policy or whatever at an OpenAI would go out to the American government and say, "Do you want us to lose to China? You have to make things easier for us.

And it is quite interesting that when you don't hold them accountable by definition, they will build what they think is needed in the way they know how to build it. And unfortunately geeks are very comfortable, all of us with AB testing. We'd put something out there and see if it works. And if it doesn't, we'll just pull it out and put another thing. And that's the typical way of quickly developing technology. And I have to say if there is, I mean Jose and I, we both agree on the chaos of the geopolitics and on the economics of our world today. But if you ask me about one thing that the world needs to be aware of and work on is what I call the mad map spectrum, the mutually assured destruction versus mutually assured prosperity spectrum of what we're up ahead here, up against here.

And on one side, this crazy mad arms race might actually cause a lot of harm to humanity, a lot of harm to jobs, a lot of harm to the energy infrastructure. We're moving faster than we need to for the single reason of competitiveness. I mean, it's not like our world is falling apart and a better LLM is going to save us. It's basically because Anthropic doesn't want to lose to Google and Google doesn't want to lose to OpenAI. That's the only reason why it's so mad. Having said that, that mutually assured destruction can be even more clearly seen in the current wars in the world. If you actually look deeper behind the scenes and count how many of the deaths are attributed to AI, whether through targeting or through drones that are AI driven, you'd start to realise that we might be heading in a direction where drones will rain on us.

Everyone will make many more of them. And that's mutually assured destruction for sure. The other side of this is mutually assured prosperity. And I have to tell you openly, and I say that publicly in front of everyone, what we've built with artificial intelligence is sort of a utility almost like energy. Yo plug in and you borrow IQ. It's like a socket in the wall where you borrow energy. And I think that today I assess that sometimes if you really know what you're doing on math skills and on scientific research and on intelligence itself, you could probably borrow 80 IQ points, maybe a hundred sometimes.\ IQ points from AI, I'll probably be able to have a garden outside where you walk to one tree and pick an apple and walk to another tree and pick an iPhone.

From a nanophysics point of view, it's doable. It's possible. It's just the limit is our intelligence. And that basically is a world of abundance for everyone, for every human. The difference between mad and map is a mindset. And when you really ask humanity, how are we going to handle that? Unfortunately, humanity waits for an event before they start working on map. Unfortunately, a catastrophic event. And I think the world today can say, can we stop competing on this? I mean, literally five years away, we have prosperity for everyone. Can we just work together on this like we did on CERN or the Space Station or whatever and make it more reasonable, pace it better for the world and build it in a way that ensures human wellbeing? I think that's really where we need to be.

Yeah, absolutely. Jose, Mo said there, it's a mindset choice between the mad scenario and the map scenario. When future generations look back at this period, how would you like Europe to be remembered in how they played into these two scenarios potentially? Do they want to be known for their technological innovation? Is it sustainability leadership or perhaps second mover advantage? Is it democratic values or a combination of all three? And how can Europe strengthen its autonomy without turning inward?

We cannot sustain some form of autonomy, relative autonomy, because to be completely autonomy, autonomous is not doable. By the way, it's not desirable. We see in history that all countries that wanted to be completely autonomous, they lost. We have to be open. And I continue to believe that Europe, because trade has been, by the way, extremely resilient. Europe should continue to be open and avoid the idea of full-scale decoupling. I believe it's crazy, this idea. So we need to invest on open as much as possible. Of course, using intelligence, using prudence, trying to be more resilient, but to keep as much as possible the world open, that's the source of prosperity. The concern I have now following what the very instant remarks of Mo, and why I'm concerned with the current developments in some areas that AI and some of these new technologies in fact are dual use technologies.

They are used for war. That's the reason. By the way, part of the big investment is coming from war or defence or war departments, however you want to call them. And this is why I do not believe that there will be any kind of possible global regulation of AI.

It's impossible. Whatever the governments may say, they will not do it because I don't see the possibilities of, let's say, United States, China, not speak about Israel or Iran or others sincerely cooperating on this matter. So AI will happen, whatever they may say, and we have to be prepared. Knowing the risks, the risks are there, some risks that we have to mitigate. I think the potential of AI, I'm not a technologist, I may be saying some wrong things, but I have the proof of being in touch with some of the leaders of this revolution, not only in US. Whatever the Governments say, in fact, AI has huge potential for mankind in science, in medicine. It's amazing. The cognitive capabilities of humans are going to increase. I'm linked with some projects of education in the global south. In Brazil, for instance, where kids from the poor backgrounds in Sao Paulo, a state of Parana, they learn faster, 33 to 35% faster with AI than without it.

So there is a huge potential of AI. So we should embrace it and not to be negative or defensive about it. Of course, we have to mitigate risk, but we have to. And what I hope that Europe answering directly your question, that Europe invests more in innovation. We have the talent, we have the science, we have the tradition. I mean, why can't we do it? We need scale at European level for obvious reasons, for deeper markets, for energy markets. We need scale because as I always say in Europe, all countries are small. The problem is that some that they have not noticed it. Even the biggest country in Europe is smaller compared to United States or China or even India demographically or Russia geographically. So we need scale, but we need to make at national level the reforms using our potential to compete, to compete in our own way because there is a European way of doing things that is not worse than others.

I don't believe on what the Americans are doing better we will be able to compete. The train already left the station, by the way, with China as well, but there are areas where Europe has the great potential and I hope that it links the idea not to be always the leader. I always say arrogance is a form of stupidity. Why should we be the leaders everywhere? I mean, in part of the problems we have today in the world is that everyone wants to be the leader. Now what we have to do is how to serve better our people, our citizens. And that Europe, I mean, it's not bad. Europe is not bad. I mean, here we are in Portugal. Portugal has a higher life expectancy than the United States. And by the way, also China and lower child mortality. But to keep this model, we need more growth and that's where the problem of Europe is.

We need a growth mindset and we don't have it because also of demographic reasons. We are getting old and we are not feeling the same sense of urgency. You go to Asia or you go to the US and you see young people trying to think how to monetize the innovations. In Europe, we don't see the same level of urgency.

Yeah, thank you. Miguel, we've just heard obviously that AI can prosper so many different industries there. Do you see it as becoming something that can help the renewable transition? Is it something that can help build capital? It can help utility companies find ways of producing energy better? Or do you think that it is an energy vampire that is essentially distracting resources away from decarbonizing the public grid and transitioning to renewables?

Well, I think clearly AI can impact the industry in a positive sense, for sure. The same way that Manel spoke about what it can do in medicine or what it can do in data or all of these different industries. So for sure also in the utility space and in the energy space, it can have improvements. But let's face it, it is a vampire at the moment. I mean, it is clearly sucking up a lot of power. So to my earlier point, if it's not managed properly, it will create a lot of pressure on society, on energy prices, on the consumers. And so to that extent, I think there is the need or this urgency to really go out and sort of figure out what is the right regulation, legislation, et cetera, to get stuff done so that you can keep the supply demand balance in line.

Because it's not just about the availability of power for the data centres, it's how do you keep power prices also affordable for people? And I think just one of the points or two points that I think that I would just like to reinforce this issue around Europe. I mean, Europe has many, many, many qualities, but it is not known for taking speedy decisions. And we don't need to be first movers and we can be second movers or even third movers, but we need to move. I think that was one of the points that Mo also made. And I think we are still taking too long to take certain decisions that are critical. And a lot of times it is at the national government level. I actually think the European Commission has done a fantastic job on many levels in terms of the recommendations, in terms of defining what needs to be done to speed up permitting and licencing.

So we have the diagnostics, plenty of reports, but actually getting that translated into national member state level and actually getting that done, whether it's the local agencies or the national agencies actually having that sense of urgency at that level, that is what is critical. The other point I'd like to pick up just on what Mo said about the AB testing, which I think there's a certain amount of that that you can do within this industry. This is typically a very conservative industry. Obviously if the AB testing leads to a blackout, okay, you don't want to test that

Scenario. So you want to make sure that you are staying within certain parameters. And we did have a blackout last year, so we know the economic impact that that has. So you want to be careful about it, but we can be more ambitious I think about within certain parameters being more ambitious about the technology we put on the network, about the technologies we deployed to produce generation, about the licencing and permitting processes. Let's bring AI to that. Why do we need to fill up a room full of paper so it can be analysed when that can be done by an AI machine? And let's build the systems to get that done quickly. So there's a lot of things that can be done I think. But again, we need that sense of urgency. We need that flexibility on the regulation, on the legislation to actually take some of those decisions and be able to move quickly.

Mo, is this energy hungry monster actually the key to our survival though? Is it the only tool fast enough to solve complex climate problems, balance our grids, optimise our renewable assets and save our cities?

No. Life was fine before AI. Absolutely. We were okay. The pace of human development is suffering from systemic error. You take a car engine and you tune it a little and you get a hundred horsepower more and then you add a twin turbo and you get a hundred horsepower more and then you put, I don't know, nitro in it and the engine burns down, melts down. And I think what is happening is not a result of need as much as it is a result of opportunity, but the opportunity is immense. The opportunity is absolute abundance. That's the opportunity. And if we all agree on the opportunity and chase that, that would be amazing. Genuinely we would redefine longevity, we would redefine productivity, we would redefine climate change. Everything is possible with more intelligence. It's just that the way... Think of it this way, you know the movie Superman?

An alien being comes to planet earth with a superpower. That doesn't make the child grow to be Superman. What makes the child grow to be Superman is that the family that adopts him tells him to protect and serve. Not to kill others with drones, not to compete in a cold war, not to compete for resources and places of intelligent supremacy and power over others. If you took that same child and told him to rob every bank and kill every enemy, he'd grow up to be a supervillain. And I think that's where we stand today. We're raising Superman. And sadly, we're being very bad parents. And when you really, really think about it this way, it is a choice. Humanity is faced with a choice. And that choice unfortunately not being made on your behalf, is now being made on your behalf. And so somehow, somewhere we need to stand up and say, look, we love this.

We love the idea of infinite longevity, total abundance. We love those. Can we please make sure? And by the way, governments have a role to play in that. Can we please make sure that what you're developing is aligned with that objective? Can we please have like FDA, when the FDA says, "Hey, by the way, we love that you're trying to help people with medicines, but we're going to review your work a little bit. Can we do that? Can we say, Welcome Superman. Let's do this right. Let's not rush with that child because that child can destroy everything.

Yeah. Jose, when you look at the EU's regulatory approach to AI and energy today, do you see it functioning more as a constraint that's slowing Europe's competitiveness or as a catalyst that's forcing more durable and investible projects? And practically speaking, as geopolitical priorities shift, the US-China race increases, energy security post-Ukraine and Iran, the EU's own strategic autonomy push. Where does real execution actually happen? Is it still in Brussels or has the centre of gravity moved towards national capitals or private capital and industry consortia making decisions faster than any regulatory framework can keep up with?

The issue of regulation is very complex and we should avoid being simplistic about regulation in Europe. And also it's very easy to be cynical about Europe, but we have to understand why there is a regulation at European level. Some regulation makes sense because to have a common market, to have internal market, you should have common regulation. So when the commission comes with proposals for regulation, it was my commission in fact, that put forward the climate change package 2007, 2008. And it makes sense because if not, the European Union at that time we were coming to Europe became commission present. It was when we went from 15 to 25 countries and afterwards 28. But afterwards I was no longer there. UK left European Union in 2016. And anyhow, to have a common regulation for 27 or 28 countries makes sense. Another point I want to make very honestly with you, and I'm no longer in European, I'm not speaking on behalf of any institution, is that worse than European bureaucracy is 27 national bureaucracies.

That's my experience, honestly. Because typically, and by the way, in this country as well, typically the idea is that they always want to ad a new layer of bureaucracy, so - called gold plating. The directives are made in Europe. By the way, there is no European directive, not a single one that is done only by the commission. They are approved by the government. The commission takes initiative, but afterwards they have to be approved by the governments and by the parliament. What happens afterwards that the directives have to be transposed. And it's very interesting to compare directives how they are transposed, let's say from Germany to Latvia or from Sweden to Greece or Portugal to France, very different way. And almost all cases, they only add the so - called Christmas tree fact. Everyone wants to put there some kind of specific issue. I could give you many examples.

So this is the problem. So answering your question, where is the centre? We need both. European Centre, I mean European Commission, European Parliament, so Brussels, Strasburg is unavoidable because it's there that the initiative of legislation comes, but without strong ownership by the member states will not be able to do it. So this is the point. I personally believe that in some areas, including on climate and energy, the European Commission has been going too far away. And now it's more a problem. To be great if in some areas European Commission declares a regulatory holiday, no more regulation for some time. But I know it's very difficult. Why?

Because in principle, everybody agrees with less regulation. I don't know any politicians say, "I want more bureaucracy." I've never heard a politician say, "But what happens when it comes to concrete discussion for its energy or climate or consumer protection or issues of diversity, gender, whatever, you go to the European parliament because that's where power is. People have not that idea, but the commission is independent. I was 10 years leading the European commission. I've never received instructions from the governments. Sometimes one government or two, they try to put. And I said to, I'm not going to say which government it was, but my conversation was in French.

And I said. I don't accept instruction from my country. Why should I receive from you? " And the commission is independent. DCP is independent. But the question is that of course the commission operates in a political setting. So before coming with a proposal, it cannot ignore what France or Germany or some of the more important countries think. But afterwards where the issue is in European parliament. Because if you try to reduce regulation in consumer protection immediate, oh, the lobby in European, more consumer protection, climate, no more environmental rules, more regulations to protect this and this and this. And there the commission is very defensive because the governments cannot put down the commission, but the parliament can, and by the way, it happened already once. So this is extremely important to understand. That's why I believe the industry today and the economic sectors, if they want the European regulation to be more sensible, they should not only work in Brussels, they should work in the national capitals.

That's critically important. But also work with European parliament because now there is a moment because the sense of urgency is coming, the drag report and there is now more awareness. Europeans, they are feeling that they are more squeezed between this global power competition. I think there is an opportunity to come with better or at least less regulation than before.

Miguel, with the huge power demand of AI and the investment required in the infrastructure, where is the cost going? Increased energy demand will increase energy prices at a time when energy is already high. Is it the consumers that will end up paying for it? Well,

At the end of the day, the consumer always ends up paying for it. And I think that's the reality and that's why we need to be very mindful about how we build out the infrastructure and how we think about these different regulatory models. And I think one of the key issues has been precisely that because of all of this demand that's happening, let's say in the US, which is one of the biggest markets we operate in, we do see this pressure on prices and we do see a higher, let's say, sensitivity to all the affordability and sort of issues around the concerns of domestic consumers and the backlash, the resulting backlash against the data centres. So we always need to keep in mind at the end of the day, some consumer is going to pay. Now you can decide out of policy that you don't want a domestic consumer to pay or you don't want that cost to be socialised.

And therefore you start having models where you tell the data centres or the open eyes of this world, et cetera, you need to bring your own generation. That's sort of one of the models which is being used. And so they have to procure their own energy and make sure that when they are building that data centre, when they are procuring that compute, that they are also providing the amount of power that is necessary for that and that it's got to be new power. That's why it's this energy additionality concept and they can't just be sort of sucking it out of the rest of the economy. And so this idea of trying to keep this balance between supply and demand is incredibly important because otherwise you do get this effect of just translating some or socialising those costs among all of the other consumers. And that inevitably will be counterproductive and will come back to bite.

That's very, very clear in the US. In Europe, we're still not seeing that in detail. I don't think a lot of the governments have woken up to that yet, but it is definitely going to be an issue. And if we see, you mentioned earlier that in the case of Portugal, but it's true also of other countries, if you are going to get a significant step up in the amount of consumption in the country, as I say in Portugal, just 20% over the next couple of years and potentially could even double in other European countries, the same thing. We need to start thinking about that beforehand so we can actually take the decisions and put in the measures to make sure that that doesn't become a political problem or a problem for the consumers.

Well, as we're coming towards the end of the show, we've got three last quick fire questions, which will require some quick fire answers, even though they are big questions. First, Sam Altman of OpenAI back in March prophesized a future where intelligence is a utility like electricity or water and people buy it from them on a metre. Do you agree? Is this where access to information is heading? And what are the implications of this for society? I'm going to ask each of you. Mo, you first.

Yeah, I think intelligence will be a utility. It will just come from a microgrid. So he's not going to be selling it. It's going to be on every phone and in every computer and we will all benefit from it. It's not going to be centralised as he thinks.

And what implications does that have on society?

You said fire, rapid fire.

Sorry, that was so rapid. I was like, I want some more.

What are the implications for that on society?

I think society will have to decide for itself. If we continue to centralise the power of everything that we need within a larger and larger platforms, I think you'll probably struggle with that. That includes individuals, it includes cities, it includes nations, it includes continent. At the same time, you can easily start to see a trend where when the challenge of the world becomes a little above your capability, a lot of us resign and say, "You know what? I don't want to be part of this. I'm going to build my own little community, my own little microgrid, my own little forms of intelligence, my own little successes." And I think you're going to start to see more and more of that. The diversions of massive, massive power concentrations along with a lot of power democracy if you want, because AI empowers all of us to do things is going to make it not centralised.

I tend to see a world where each and every one of us will become an entrepreneur. Each and every one of us will become a village if you want with others. And hopefully that will be a better world, I think.

Thank you, Jose.

I agree that first of all, AI solves problems that we humans cannot solve as simple as that. That's very important. Human species is now coming to the conclusion that we are not the most intelligent entity in the world. Not only AI solves problems that we cannot solve, that sometimes we don't understand how AI solves. That's the reality. So this is going to develop. So this is of course a very important asset. Intelligence is extremely important asset. So now maybe AI is going also, and I repeat, I'm not a technologist, but having the privilege of being in contact with some of the best on this matter, maybe it's also going to solve some of these problems because why? And then I echo what Mo said. Why should all be huge platforms and some of concentration? I don't see precisely because AI can help us solve the issues on another scale, be it regional, be it local, be it national, be it individual.

So I really believe the potential of AI is huge also for us human beings, each one of us having creative solutions. So we don't know yet, at least I don't know yet, how it's going to develop, but I see there great opportunities. So knowing the risks, but it's better if we embrace opportunities Because if not, we are going to lose the competition. That's for sure. If you just are in a position just of reacting and just opposing, we will not succeed. So let's see the potential of AI also for new models, decentralised models, closer to citizens that help us solve some of these problems.

Miguel.

I think sort of the engineer economist view of this definitely should be metered because it's not for free. And I think if it's not free, it makes you then really focus on finding the best use cases, the most cost-effective use cases, making sure you use that intelligence, which is always a scarce resource because it has a huge cost for what makes sense, whether it's the latest medicine, whether it's longevity, whether it's other stuff rather than producing perhaps not very value-added information or charts or photographs or stuff like that. So I think that making sure it's not for free and that you actually have to metre it and you have to pay for it will also end up focusing us much more on what really ads value to society.

I feel like we're in a chicken and the egg scenario at the moment because to train AI, we need data centres. To build data and centres, you need capital investment and supportive policy. And then you need to upgrade a grid so it can deal with the power demand for all of AI to really take off. So do we think at the moment we're in this hype stage? Do we think that AI is all hype or are these just teething issues when we look back? They're big issues, but teething issues when we look back on it and the benefits are soon absolutely going to be realised. I'll go back to Moh.

So the era of bigger data centres to consume more at the training stage I think is decaying. There is still benefit coming to AI models from bigger and bigger and bigger models, but I think a lot more of the advancement is coming from recursive learning. It's coming from basically AI building itself, reviewing its own models from knowledge of mathematics and physics and material sciences and so on that will actually improve the way we build things. It's coming from energy efficiencies, it's coming from better software, better algorithms and so on. A big chunk beyond the initial 2023, early 24 bigger is better is now behind us I think. So would that mean that we would break that cycle from an investment point of view? I think unfortunately the early movers may be surprised on the number of GPUs that they bought that may not be as necessary, especially as NVIDIA continues and China comes into the game to develop better and better GPUs and other forms of compute.

So the trick here is to understand this. A lot of the advancement of AI going forward is not just big boxes and big hardware. A lot of it is coming from the minds of better scientists as well as the minds of AI itself. So there is a lot of magic in coding. Jose?

Not everyone will win. So from a market point of view, it's almost sure there will be some moment of correction. That's obviously. So all those investments that are really mind-blowing that are going on, there will be some moment of correction.

We are not yet sure how we'll translate all these investments in, in fact, opportunities for growth and productivity and competitiveness. So there will be winners and losers. Having said that, I continue to think the transformative power of AI and other technologies, and I'll speak about AI, but other from quantum computing. What's going on in the field of biotechnology, of genetics? We are at the beginning of a huge scientific revolution that is disruptive really. So I continue to think the opportunities are there, but how it's going to operate? We have here people that know much better than me concretely about the engineering of this, of data, the dimension of data. I honestly don't know. So I think we should follow attentively the developments with prudence, but being sure that this is a transformative, this is not just a hype. No, this is not just a fashion.

We are going to have to live with AI for the foreseeable future so better to understand it than to resist to it.

Miguel.

Well, I think definitely teething issues as always happens when you have a boom, when you're not living your business as usual, cruising speed. This is obviously placing a huge amount of pressure on the system, certainly on the energy system, but sort of deriving from all the growth around the AI. So it is putting a lot of pressure. That's why I think the idea of urgency and people really stepping up to get stuff done is really, really important. You could have crunches, you could have situations of severe stress. As I say, the energy sector is particularly relevant because if it collapses, I mean then you don't have AI, but you don't have anything else either. The economy shuts down. So we need to manage this carefully and as fast as possible. So that sense of urgency, testing, thinking about how we can do things faster. And I'm sure we'll get there, but building things, building large infrastructure that will last for the next 35 years is very different from testing software, which can be updated overnight.

So we need to be very intentional about how we do this and make sure we don't go through a boom and a bust cycle or that we actually make stuff that is important and valuable for the next decades.

Well, very quickly for the last question, what is your one big hope of where we will be in five years' time? Miguel?

My big hope would be that we would have sped up this whole process, that the sector would continue to grow both the AI and the power sector with it, that we would've simplified a lot of the licencing and permitting that the governments would've stepped up or we have smarter regulation and that this would be growing even faster than we are today.

First of all, my hope is that we keep the world avoiding worse situation in terms of war. The geopolitical risk is increasing day by day by global competition. So let's work for peace. That's more very, very important and not forget that. In the economic developments, I continue to believe I'm overall positive about technology and science. Not only AI, but also the best minds in the planet are working on those issues. And I prefer intelligence than stupidity. And so I continue to be positive and I hope that this will help solve some of the problems that we are now getting in touch with.

I wanted to say the same. I was going to say, let's hope we're here in five years. But for the context of being in Europe, I would hope that Europe participates enough so that this turns into a multipolar contribution. And as you rightly said, we don't really have to build the frontier models in Europe, but there are so many ways Europe has always excelled at using basic technology to build amazing things. And so let's hope that Europe starts to participate, reduces the regulation and the bureaucracy in ways that allows them to lead in certain segments of this, whether it's science, whether it's applications and so on. And hopefully because of that, it doesn't become a Cold War anymore. It becomes more multipolar if you want.

Well, this has brought us to the end of the episode. I want to thank each of the three guests. Miguel, thank you so much. Thank you. To Jose.

Thank you as well.

And to Mo. Thank you very much. Very much for having us. Well, today's conversation has made one thing abundantly clear. We are no longer just managing an energy transition. We are managing a global power rebalance. From Jose's musings on the future of Europe as a global political power to Miguel's look at the sheer physical load of AI that it's placing on our utilities and Mo's challenge to have the mutually assured prosperity, the stakes have never been higher. There is a lot of power that can shift today, power for control of energy, control of markets and control of data. Who will win and what our world will look like in five years' time is anyone's guess. And if today's episode was about the demand for this new era, we need to look back at the supply side. In episode two of PowerShift's series, grid-bound modernising grid infrastructure, we take a deep dive into the ageing nervous system of our planet.

We explore why building a data centre in 18 months is a miracle, but waiting 10 years for a grid connection is a tragedy. I'm Tom Parker and this is The Next Five podcast. Thanks for listening.