The IT Breakroom Episode 1: The Green Data Center Debate in the Age of GenAI

Episode Summary:
In this episode of IT Breakroom, we we chat with Stephen Bowes-Phipps, VP of Data Centre & Cloud at State Street. We discuss what the modern green data center looks like and whether truly sustainable data centers are possible in the age of GenAI. The conversation covers hyperscalers vs traditional data centers, energy constraints in the US compared to Europe, nuclear power, photonic chips, and whether sustainable infrastructure can keep up with AI growth.

About The Guest:
Stephen Bowes-Phipps MBCS CDCDP AUH, VP of EMEA Data Centres and Cloud at State Street. He has led infrastructure strategy across regulated financial and public-sector environments and has played a role in shaping EU-level standards around efficiency, sustainability, and risk management.

Key Topics Discussed in This Episode:

00:00 Intro
07:01 What Is A Modern Green Data Center?
11:44 Can Data Center Heat Power Cities?
16:37 Hyperscalers vs Traditional Data Centers
20:31 NVIDIA Dominance vs Emerging Photonics Chips
29:48 Who Really Benefits from GenAI?
32:32 Europe vs US Data Center Growth and the Nuclear Energy Question
39:45 Will AI Outpace Sustainable Energy Solutions?
44:34 Predictions for AI Data Center Growth & Battery Technology

Watch on Youtube:

Listen on Spotify: https://open.spotify.com/episode/1QwfvSghzNI4ghmgs1X4jl?si=ec4d0fedd7f44d7e

Episode Transcript:

(Transcript generated automatically from the original recording. Because the video has been edited, some sections of the transcript may not align perfectly with the final episode):

Meredith Lawrence:
Good morning. How are you doing?

Bowes-Phipps, Stephen:
Hello. Good. Thank you.

Meredith Lawrence:
Wonderful. So nice to meet you. Thank you for joining us today on the IT Breakroom. You’re actually our first guest, so we’re super excited to have you.

Bowes-Phipps, Stephen:
Oh, I’m very excited to be your first guest. Let’s hope we can start it with a bang.

Meredith Lawrence:
Yes, start it with a positive bang. I love that. Tell me what time it is over there.

Bowes-Phipps, Stephen:
So it’s just gone 1:00 in the afternoon.

Meredith Lawrence:
OK, so you’re almost halfway wrapping up your day.

Bowes-Phipps, Stephen:
Yeah, it’s a Friday, so who knows? Who knows what might come out of the woodwork?

Meredith Lawrence:
Yeah, you can’t complain. Perfect. Well, let’s get started. I want to kind of just jump right in to get a little background on who you are, where you are today, and then we’ll just move right into our conversation on data centers.

Bowes-Phipps, Stephen:
All right, great. So yeah, I’m Steve Bowes-Phipps. I’ve had a long and varied career in data centres. In fact, I started off by pure accident managing a data centre when I was at university. I went out and did a placement at a computer company and they got me to manage their data centre. It was called a computer room at that time. You had mainframes in it, small systems and stuff. Little did I know that that would be the kind of place that I would inhabit for the rest of my career. I tried to get away from it several times, but I always seem to come back.

I think I’ve worked for various different types of companies. I’ve worked for financial institutions, I’ve worked for the public sector. The enterprises and the challenges are all slightly different, but also all the same. At the end of the day, people have IT and they need that IT to be stored somewhere and looked after and supported, and that’s what the data center is there for.

I think since around 2007 or 2008, there’s been a huge focus on greening data centers. It’s about trying to ensure that the energy we use in the data center is used as effectively as possible. That’s really been a great passion of mine. Since then, I’ve worked a lot with organizations, both as a consultant and as a direct employee, to try and bring better utilisation, effectiveness, and efficiency within the data centre space. For me, that’s as interesting as what the data centre does from a purely business perspective.

Meredith Lawrence:
I gotcha. Now, have you primarily worked in Europe, or are there any other areas? Or is that kind of your main focus, London and Europe?

Bowes-Phipps, Stephen:
Well, when I was a consultant, I went abroad wherever I was needed. I’ve been all around the world pretty much with data centres. There are certainly some differences, and I’m sure we’ll touch on that a little bit later, but there certainly are some differences in the way that data centers are approached from an infrastructure perspective both in the US and in Europe, and in the Middle East and APAC.

It’s interesting because data centres get a lot of bad press. People think they’re energy hoggers. Public opinion these days is that it’s a hot topic. People say they take up lots of space, hardly anybody works there, they use all this energy, they drive up costs, and what are we getting back for that? There are different approaches in different parts of the world, and it really depends on what the constraint is.

Energy now is one of the hot topics. How do we get enough energy for these data centres? Before, it was planning. How do you get planning permission to keep building these large warehouses full of servers? It’s definitely changing the landscape of where people perceive data centers should go.

There are some huge data center hubs. In the US, you’ve got Virginia, which is probably the largest conglomeration of data centres in the world. You’ve got London, which because of its international connectivity has also been one of those established places. Then you’ve got Amsterdam, Paris, and traditionally they’ve been called the FLAP-D countries. FLAP for Frankfurt, London, Amsterdam, and Paris, and D for Dublin.

There’s been huge growth and a network effect, because as you build more data centres, you attract more resources, and then you build more data centres because all the resources are there. But that’s changed so much over the last couple of years with GenAI coming into the picture. Now people are building data centers in other areas because you just can’t get that much power in such a small location.

Meredith Lawrence:
Right. I mean, we’ve seen a huge increase even in Georgia. That’s where we’re based. We were talking to the Chamber of Commerce of Georgia and they said we’re becoming the second largest investment area for data centers in the US as of last year. So it’s super interesting to see how much we’re expanding and growing.

But let’s get right into it and talk about some of the buzzwords going on these days with modern green data centers. What does that actually mean? When we hear the phrase modern green data center, what does that really mean? Is that just a buzzword or are there actually data centers being built and framed as these modern green data centers?

Bowes-Phipps, Stephen:
Well, it’s a great question, and let’s start with the good news. The good news is that since 2007 or 2008, when we started looking at this, the industry has made huge strides to improve the efficiency of the plant, the electrical equipment that essentially makes up a data center. We’ve seen huge advances. We’ve seen great innovation in that space. If that was the only thing we were concerned with, we could probably give ourselves a nice big pat on the back and say well done.

Unfortunately, it’s not the only thing. There are two things going on. We’ll get to the GenAI thing, which is the second thing, but the first thing is the IT itself. Ever since we started increasing compute power, people have taken advantage of that and built inefficient systems. If you didn’t have much memory or disk space, you’d write a very concise piece of code to run an application because you had to make the most of every byte and every CPU cycle. But today, that’s not the case. You have these massive systems, massive compute, huge amounts of storage, huge amounts of memory, and so people get lazy.

When you look at green computing, you can look at the data centres and say they are a lot greener than they used to be. But then you look at the IT and say we’ve still got a long way to go. There is actually a principle which is fascinating to think about, and it’s called Landauer’s Principle.

Meredith Lawrence:
I’m ready to hear this. I haven’t heard of this before.

Bowes-Phipps, Stephen:
He was a physics professor, and he basically came up with the notion, backed up by evidence, that around 99.97 percent of the energy that goes into a data centre comes out as waste heat.

Meredith Lawrence:
That’s about as high as you can get. Yeah, that’s crazy.

Bowes-Phipps, Stephen:
99.97 percent. So when you look at a data centre, does that make us think differently about what a data centre is? Is a data centre really a hog of energy, or could it be a supplier of energy to other industries or to residential buildings? Once you start to think down that path, you start to think maybe green data centers are not just about the way we take energy in and use it to produce digital services. It might also be about how it comes out the other side and adds value within our communities and businesses.

Mostly today, data center providers haven’t really thought that much about that. They’ve kind of ignored it and pushed it aside and said, we’ve just got to build data centers. But now they get pushback from governments, local districts, and residents saying, hold on a second, we need that energy. We need that energy to light the football pitches for our kids. We need that energy to heat our homes. We need that energy to run our businesses. So you can’t just take everything that’s there just because it’s there.

Meredith Lawrence:
So how does that actually look? I did see a case study that Finland was one of the countries taking heat from the data center and then reusing it in residential areas. It got me thinking, do these residential communities need to be located close to the data center for that to even be effective? Is that happening just on a small scale with one-off case studies, or is that actually something people are considering when they’re building the data centers? Or as you said, is it still more about just building and the output, and we’re not really at the stage yet where we’re trying to use that heat to potentially warm homes or commercial buildings? What does that look like?

Bowes-Phipps, Stephen:
Yeah, that’s a really good point. If you look at GenAI, which is where most of the demand for building data centres comes from these days, there are three stages to GenAI. It’s the latter stage, the inferencing stage, that most people focus on because it uses the most energy and the latency needs to be very low so that people get instant responses.

Inferencing is one of those things that could potentially be built in very urban areas to provide very quick response times, but they also use a lot of energy. So there is that challenge around how do we build something in an urban area where there is limited energy capacity, but that can also provide a benefit to the local community.

There is a very interesting small company from the UK, and they’ve got several case studies in the UK and one they’ve just done in the US, where they build data centres for free next to leisure centres to heat swimming pools.

Meredith Lawrence:
Oh wow. OK.

Bowes-Phipps, Stephen:
They basically say, look, if you allow us to chuck our heat into your swimming pool, then we’ll build this data centre for free. It’s not a huge data centre. It’s quite a small one, but it can save hundreds of thousands of pounds on energy for that leisure centre company. Imagine the cost of trying to keep a swimming pool heated all year round.

Meredith Lawrence:
That’s creative out-of-the-box thinking. That’s a cool concept.

Bowes-Phipps, Stephen:
In the UK, most leisure centres are publicly run, so it’s funded by ratepayers and taxpayers. Therefore, there’s a huge benefit there because we then have to pay less rates to sustain what is a community good. They’ve basically taken this concept of reuse of waste heat and put it into a commercial perspective where they can add value and make a profit at the same time. It’s a great development.

But what we need to do is convince the large organisations, the Equinixes, the Digital Realities, the huge colocation players, and also the large hyperscalers like Microsoft, AWS, Google, and Meta, that they should really be doing the same. If we want to shake off the reputation that all we do is take and don’t give anything back, we need to change that.

It’s a bit of a shame because I think we’re held up to very high standards because we’re a new industry. How many people do you see campaigning outside of a steel smelting plant? They use a lot more energy than a data center, I can tell you.

Meredith Lawrence:
Exactly. And I was going to say, I feel like over the past year or two this has become such a hot topic. It’s interesting you mentioned back to 2006 and 2007 when some initiatives were being started to be more energy efficient. It’s 2026 now and this is such a hot topic, especially for communities worried about the power grid and their energy bills rising and things like that.

I do want to put in perspective that we’re just talking about a regular data center. How does that compare to these hyperscalers? I think people have a hard time understanding the difference between just a business having their data center to function compared to these massive hyperscaler companies. What’s an analogy or how would you describe the difference between those two?

Bowes-Phipps, Stephen:
If we take the technology out of it and look at a typical enterprise data center, it’s going to use very small amounts of energy. People have been complaining ever since I can remember building and running data centers back in 2000 that these chips are going to need so much energy. But in reality, most enterprise computing is actually very low energy use. It doesn’t really require that much.

That’s why traditionally when we’ve looked at waste heat reuse, the heat you got out of the traditional data centre has been very low grade. It’s very low temperature. You can’t do much with it. You send it along a pipe and by the time it gets to its destination, it’s pretty much lost most of the heat it had to start with. So trying to do anything with that has never been a focus because you haven’t been able to do much with it. That really hasn’t changed in 25 years.

But what’s changed completely is GenAI. This is a very different situation. Because now, instead of data centers providing the solutions for how you cool systems and run compute, NVIDIA is driving the conversation around how data centres get architected, how you provide power, and how you provide cooling. Their latest GPU platform, called Rubin, comes in its own cooling cabinets that provide on-chip cooling.

They’re not leaving it up to the data centre to decide. They’re saying, here’s a system. If you want to support this system, you’ve got to plug it into the chilled water circuit and you’ve got to cool it. Before, it was the data centre that decided how that system got cooled and powered. There’s a lot of concern in the data industry that we’re losing control of the roadmap for how data centers get built and supported, and it’s because NVIDIA is so dominant in GenAI.

Meredith Lawrence:
I was just going to ask that. Is that just kind of the go-to for everybody to buy NVIDIA product and just follow suit and trust that this is what we need and who we need to purchase from? Are there any emerging competitors, or how do you view that? Even from a marketing background, I’m aware of everything going on with NVIDIA. It’s such a powerhouse.

Bowes-Phipps, Stephen:
This is my favorite subject now. I have an IT background. I’m not an engineer. I’m an IT guy who happens to manage data centers, and there are very few of us in this industry. It’s mostly engineers.

What engineers are very good at is solving problems. If you have that mindset where all you want to do is solve problems that are thrown at you, become an engineer. So when people say, here’s a compute platform, I need you to power it and cool it for me, an engineer goes, great. They really get their teeth into that.

Whether that’s an electrical engineer, mechanical engineer, or civil engineer, if they’re devising bridges or aqueducts, it’s the problem of how do you get from A to B and how do you do it safely and successfully.

However, what engineers are not always that good at, which IT people are good at, is looking at the technology and remembering that data centers are there for the IT. They’re only there for the IT. If the IT keeps going along the same roadmap that NVIDIA envisages, starting off at very high kilowatts per rack and then potentially 300 to 400 kilowatts per cabinet based on the GPUs they’re bringing out, those kinds of challenges are not sustainable forever.

Eventually we’ll be creating data centers that can only support a specific type of chipset, and that’s where we’re going. We’re building data centers that only support NVIDIA chips or AMD chips or whoever is building out those large GenAI environments.

But you only need to look at what’s going on from a technology perspective to realise that something could happen that would change this overnight. For me, the most exciting area is photonics. Photonics is about light. There is a company that was just announced last week, I remember reading it in Data Centre Dynamics, and they have brought out a photonics chip that uses light and therefore doesn’t require these huge energy-hogging chips. You can do inferencing using light.

What’s even more fascinating is that scientists are currently looking at glass.

Meredith Lawrence:
OK, I have not heard this before.

Bowes-Phipps, Stephen:
Did you know that the mathematical models for light waves passing through glass are exactly the same as the mathematical models for inferencing?

Meredith Lawrence:
OK, interesting. That is crazy.

Bowes-Phipps, Stephen:
They’re the same. So scientists reckon that they can replicate inferencing calculations using just glass, etching circuits into glass and shining light through them.

Meredith Lawrence:
Have they tested this before or is this just theory?

Bowes-Phipps, Stephen:
No, this is in the lab. They have proven it. It’s just not commercial yet. These things take a little bit of time. But what is going to drive commerciality? The impossible energy and cooling requirements of the current path.

NVIDIA has basically staked much of its future on normal silicon. I think recently they have invested in some photonics, so they see the writing on the wall, but they’re ignoring it. You only have to look at some of the major players like Microsoft and Google. They are looking at photonics in a big way because they see why they’re spending billions with NVIDIA when they could invest in alternative chip technologies that could replace NVIDIA completely.

Meredith Lawrence:
I was interested in that because I was wondering if that would be more of an option for small to medium-size enterprises who aren’t able to access the latest and greatest NVIDIA equipment. But it’s interesting that even the huge hyperscalers like Microsoft are looking into that as well. How soon do you think that could even be an option? Are people looking at this right now to test out in smaller areas, or is this something that would take five to ten years from now?

Bowes-Phipps, Stephen:
This British company has just gone through a funding round and they’ve raised billions to invest in and sell these photonic chips. So I don’t think it’s that far away.

I think once it becomes established, people who are building these data centres for NVIDIA chips are going to find themselves with large empty data halls that nobody can use. It’s really difficult to swing around because once you build for a very high-density data center, if you then try to focus that on enterprise workloads, the infrastructure’s not right. It’s too big, the cost is too high.

The problem’s not going to be for the hyperscalers themselves. The colocation people are the ones who are going to get hit hardest because hyperscalers are funding data center growth using the cash piles they have. Microsoft has a huge amount of cash it’s sitting on, same as Google, Oracle, and Meta.

Meredith Lawrence:
It’s too much cash to even really visualize.

Bowes-Phipps, Stephen:
Exactly. They blow a few billion, a few hundred billion, and Jensen Huang’s probably given them a few hundred as well so they can buy his chips. They don’t care if that money goes up in smoke. But the colocation providers, who are investment backed, don’t have those kinds of cash piles. They’re the ones who are going to be caught in any kind of crash.

There are still no peer-reviewed case studies that I know of, or anyone I know knows of, that demonstrate any real return on investment with GenAI. I think it’s going to end up being very specific areas where it’s going to add huge amounts of value, maybe pharmaceuticals and health.

Meredith Lawrence:
That’s exactly what I was going to ask. What type of industries or companies do you think are really going to benefit from this? We’ve seen amazing applications in pharmaceuticals and research and development. But then there’s also a ton of marketing buzzwords about how great GenAI is and how it will make everybody more efficient. It’s like, what do we actually need this for? Let’s take a step back and think who can actually benefit from this and what industries can benefit from this. Do you foresee every industry trying to adopt this, or is the focus more on the latest and greatest technology for the medical field? What’s your opinion?

Bowes-Phipps, Stephen:
I think for limited language models where you’ve got a walled-off set of data, GenAI is going to be hugely advantageous. It’s going to open people’s eyes up to other uses for that data or expand the use of that data.

But trying to put GenAI into the whole internet is doomed from the very start. I’ve read articles saying that within the next couple of years, 99 percent of the information on the internet will be generated by GenAI. So basically it’s a case of, have you heard of mad cow disease?

Meredith Lawrence:
Yep.

Bowes-Phipps, Stephen:
You know how mad cow disease happened. Basically cows were eating themselves. Food waste ended up in their feeding cycle, and they developed this degenerative disease that caused neurological issues. Some people have likened ChatGPT-5 to exhibiting signs of mad cow disease. It’s basically eating itself or eating the output of other GenAI systems.

Meredith Lawrence:
That makes sense. Even from the marketing world, when we’re creating content, you create thought leadership content and that’s what shows up when people search on Google. But now people are using AI to create content, so the AI is just reading the other AI that people have created and teaching itself. It may or may not be accurate, and we’ve seen that a lot. Even in our own industry, I’ve searched some things and thought this is not super accurate or it’s a little misleading. You need a little background knowledge to make a discernment between what it’s saying and what’s actually true.

I don’t think that’s far off, and it really depends on who’s using it as well. I did want to segue into the growth happening in the UK compared to the US, or even Europe compared to the US. How is that going? Are there more regulations and restrictions when people see the headlines that hyperscalers are trying to move to Europe? Do you all have more regulations over there where it takes more time to get approved, or is it kind of loopholes? How does that look?

Bowes-Phipps, Stephen:
The biggest thing that’s holding back data centre development in particular over here, but also in the US, is the constraint on energy. How do you get access to a grid that is at full capacity, or maybe not necessarily full capacity, but where there aren’t connections to get that capacity?

I think pretty much within Europe and America, we’re looking down the same avenue of, is nuclear the answer? I’m a great proponent for that, because when we talk about nuclear, a lot of people get scared because of Three Mile Island or Chernobyl or Fukushima. But those are huge nuclear power plants. If you have an accident in a huge nuclear power plant, there’s obviously a lot of fear around that. But nuclear doesn’t have to be huge.

We’ve got small modular reactors, and over here Rolls-Royce produces a small modular reactor that is being encouraged by the government to develop nuclear power at a much cheaper rate than these huge plants. Part of the cost of putting in a new nuclear power plant is that everything has to be built on site. If you have to build everything on site, including the walls and the dome and everything like that, you’ve got delays around materials coming to site, delays around construction due to weather, and it makes everything really difficult. The timescales go out and the costs go up. Everything that was supposed to cost maybe ten billion then suddenly costs thirty or forty billion, and people say, well, how do we get our money out of that? We have to put the energy price up.

Small modular reactors can be built in factories and shipped out because they’re small. What’s even better is the very small modular reactors, sometimes called nano reactors, and they come in an ISO container, like a shipping container.

Meredith Lawrence:
OK.

Bowes-Phipps, Stephen:
You can basically put one in your parking lot. It produces 30 megawatts of power. You could have two of those so you’ve got dual feeds to give you resilience. It’s taking up a few parking spaces.

Meredith Lawrence:
And it’s not taking up that much space. Wow. OK.

Bowes-Phipps, Stephen:
But the problem is that regulations in both the US and the UK and in Europe were designed around these huge power plants. One of the oddities around having any kind of nuclear power plant is that you have to have an armed guard. Most police in the UK, except at airports, do not walk around routinely with weapons, particularly loaded ones. So how do we overcome these outdated regulations and the permits required to move quickly in providing our own energy?

That would solve several problems. It solves the problem of getting energy, which is what people need. It also solves the community problem because the community is saying, if that’s how they’re going to do it, we don’t want them investing in gas power plants because that’s increasing our climate issues. We don’t want them digging up more coal.

Nuclear is a zero-emission energy source and it has one of the safest records in terms of human harm historically. The one with the worst safety record is coal. Not only is it dirty and produces pollution and climate change, it also kills far more people than nuclear ever has. So for me, nuclear regulation is the thing that we really need to look at and focus on. Several governments, both the UK and others, are focused on trying to do that for SMRs. I hope they do it for the very small ones as well and allow people to move forward.

They’re still not cheap, but as they develop commercially and sell more of these, they can bring down that cost of power. The first one’s going to be super expensive because it’s got to overcome all the hurdles of regulation and permitting. After that, it will get cheaper and cheaper. Then you sell the waste heat into the community or give the waste heat away, and I think there’s a really positive story there that we should really be focusing on. We’re making moves there, but I’d like it to speed up.

Meredith Lawrence:
Making progress. I wanted to ask as well, in an ideal world, we have better infrastructure, we’re returning to nuclear energy, we’re reusing heat and things like that. Do you think the pace and scale of AI is going to outpace the positive changes we’re making? Are we able to find balance there, or are we still kind of in unknown territory to make that call?

Bowes-Phipps, Stephen:
Isn’t it interesting that the head of AI at Microsoft this week stated that all white-collar jobs will disappear to AI in the next 18 months to two years? Now, he would say that, wouldn’t he, because his company is investing billions in AI. I read one of the comments on LinkedIn that said then you’ll need an army of white-collar people to clear up the mess afterwards.

Unfortunately, I think there’s a lot of people spouting a lot of nonsense. A colleague of mine in the industry, now retired but never seems to stay retired, once said if you want to work out what somebody is saying and what their intention is, you need to understand why they want to say it. Who are they? What interests do they represent?

I don’t think we can believe pretty much anything that Sam Altman says, or Jensen Huang says, or that Microsoft or any of these people say, because they have a stake in all of this. We need to turn to other sources to find out what the reality is.

As I said, in pockets there are going to be some huge advances and huge benefits to what GenAI can give us. For everything else, the rest of us, people who work in normal jobs, are going to learn to live alongside AI. It’s going to be a tool for us, just like a calculator is.

Excel doesn’t stop you from knowing maths. You still need an education in how to perform calculations because the only way you can check that the answer you’ve got is right is by knowing roughly what it should be. Accountants use various ways of doing that. They arrive at an answer from several directions and then use a reasonableness test. Has it got the right number of zeros? AI is going to be exactly the same. It’s a tool. It allows you to do research very quickly, but if you don’t want to be made a fool of, you’ve got to go and check that stuff and make sure that it’s telling you the truth.

I’ve found a lot of times that it’s great at extracting something in a way that I want to use it, but I’ve had to put a lot of work into the content because I don’t like the way it says it, or I don’t like the information it’s drawn from, or sometimes I doubt its accuracy at all.

Meredith Lawrence:
I think prompting plays a huge role, and that’s what a lot of people are saying, learn how to prompt AI to use it more effectively. But overall, that’s kind of a good message for consumers. Learn to use it as a tool, don’t listen to all the fearmongering, but also be very aware of the pros and cons of AI and how to use it to make your life or your job more efficient. You don’t need to use it for everything. Use it as a tool.

Everything is happening so fast, which is why people get nervous, and rightfully so. But it definitely can benefit in some ways as well. To wrap up today, I want to know your opinion on what you think is next. In the next three to five years, are we going to see the grid just filled with data centers and hyperscalers, or what do you think is going to happen?

Bowes-Phipps, Stephen:
I think politically it’s too sensitive for governments to allow the grids to be totally consumed by data centres. We’ve got a situation in Ireland at the moment where 18 percent of the grid is supporting data centres. It’s been a real hot topic over there, and they’ve had to adapt the way they plan out support for building data centres.

I think the administration in the US is saying you’ve got to go and get your own energy. Overcoming those challenges is going to be the real key. But I also think it’s not just the chipsets. It’s going to be around other coexisting technologies as well as technologies that support data centres and the grid.

One of the things I feel really passionate about is that the next key technological change that will transform humanity is battery technology.

Meredith Lawrence:
OK.

Bowes-Phipps, Stephen:
Can you imagine a world in which we have the ability to recharge a battery very quickly and then have it last for a very long time? People think of its application in phones, cars, homes, businesses. It will be a game changer.

The key to that is not just the technology itself. I actually think these batteries are coming, by the way, because I had a conversation a couple of weeks ago with somebody who’s produced a battery that’s going to change the EV market overnight.

Meredith Lawrence:
Oh, awesome. OK.

Bowes-Phipps, Stephen:
So I think it’s coming. But we also need to worry about the geological factors. Where are we getting these minerals and materials from? We don’t want to be getting them from Russia or North Korea or other places where human rights are an issue or where supply could be used as a weapon against us. We’ve seen that. We’ve seen people say we’re not going to send you those rare earths anymore because we don’t like what you’re doing with tariffs.

Having this everyday material that’s available everywhere around the world is going to be key. Some people talk about hydrogen. Hydrogen’s fine, but there’s very little green hydrogen. It’s all blue hydrogen. Blue hydrogen takes a lot of energy to create, whereas green hydrogen is the stuff you can produce from renewables.

So I’m optimistic that the path forward is one where we get to a point where we’re using clean technologies from places around the world that aren’t going to hold a gun to our heads, and that benefit mankind. That would enable us to make the best use of these huge innovations in language processing and things like that, which can transform our lives. Data centers are a huge part of that because that’s where the internet lives and where the sophisticated models live. So I’m hopeful.

Meredith Lawrence:
That’s refreshing to hear because everything we use as a consumer these days demands a lot, and the data centers are powering that. It’s refreshing to hear that you’re optimistic on that, just because there’s so much buzz and conversation about it these days and it’s like, what’s true and what’s not true? It was really great to hear your opinion coming from someone who’s actually in the space and has worked in the space for a long time.

Well, thank you so much for being on the podcast. I forgot to ask my intro question, but I do want to ask, what’s in your coffee mug today? What are you sipping on?

Bowes-Phipps, Stephen:
Well, I like the way you’ve got that small cup there. You’ve got this massive one that you’ve been getting up before.

Meredith Lawrence:
I know. I have my water and then I have my 8:00 AM coffee.

Bowes-Phipps, Stephen:
Well, I am English, so this is tea.

Meredith Lawrence:
OK, I wasn’t going to make that assumption. That was going to be my guess, but I was like, let me not stereotype immediately.

Bowes-Phipps, Stephen:
Unfortunately, I sit straight in that stereotype. It’s English breakfast tea as well. It’s not even Earl Grey or chamomile or anything like that.

Meredith Lawrence:
Wonderful. I was like, we’re either going to be drinking tea or we might be starting happy hour a few hours early. So one of the two.

Bowes-Phipps, Stephen:
It’s a bit early for me, and I’ve got to take my son to football practice later.

Meredith Lawrence:
Oh, fun. So no happy hour. Alrighty. Well, thank you so much for joining us today.

Bowes-Phipps, Stephen:
Thanks so much for having me. It’s been a pleasure.